The Composite H-Quad

Having trashed my Tricopter to no end, I thought I might like to take a stab at building something of my own.  I wanted it to be light, very strong (I beat the crap out of things), make it backpack transportable and also minimize vibration. I knew that dissimilar materials do a lot to kill vibration so I began to consider using a laminate of G-10 fiberglass sheet epoxied to a lightweight Poplar wood core. The G-10 provides huge tensile strength and the wood provides compressive strength. I wanted an H layout instead of an X Quad as I intended it for photography and with H-quads it’s a lot easier to keep the rotors out of the picture.

To further reduce vibration I sandwiched the motor arms with neoprene rubber where they attach to the body. But wait, there’s more! The attach screws themselves float with a combination of rubber washers and something called “Neobond” washers from a local real hardware store (not a big box). These things have a thick soft rubber washer bonded to a steel washer. The resultant setup has the arms firmly held to the body, all via rubber-to-rubber connections. I went one step further and floated the camera mount platform between two layers of a Sorbothane-like material called Kyoshoto Zeal. It has completely killed all Jello effects in the video.

Making it transportable was a challenge, and I was only partially successful.  I began by making the arms removable with just a couple of screws holding them in place. I then set it up so the four  ESCs are on the body and the three motor wires from each connect to the motors via bullet connectors which have color coded heat shrink on them. So I just match the colors and I know each motor will be turning in the right direction. What I didn’t take into account was the additional extension of the propeller blades, 5″ at each arm end. This makes for a somewhat long package to transport, made worse by the fact that each end, the props, are bendy. This has proven to be a pain.

Although the paint job may seem frivolous, I assure you it’s not. When the quad is flying some distance away and I’m not running FPV, it’s often times very hard to discern its orientation. You might think it’s pointed toward you, so you push the stick forward to bring it back, when in fact it’s pointed the other way and is flying merrily away from you. The white nose end provides a contrast which tremendously improves orientation recognition.

All the goodies in the current incarnation. This weighs in, without battery, at 1,454 grams.

I floated the camera mount off the body by using rubber spacers. This meant the camera mount, constructed of 3/32″ modeling plywood, was a separate piece. After having to rebuild a first iteration of a mount that didn’t work well, it occurred to me I could make multiple mounts for different purposes and simply swap them. This allows me to quickly change between a forward looking mount (about a 15 degree downward slant), a 45 degree oblique view mount, and a pure vertical mount to look straight down. My still camera, a 14 megapixel Canon A2200 is running  CHDK software which I’ve setup to shoot an image every 5 seconds at a shutter of 1/1000 second. This has provided excellent pictures. I actually find the forward looking mount does the best “scenic” shots, and I want to use the vertical mount for mapping or photogrammetry projects. I think the 45 degree mount might be useful when doing shots on hillsides or rapidly rising terrain.

Detail of Swann HD camera used for flight video. It also provides a mediocre video output which is fed to the transmitter for FPV use. Its  1080 recorded video is quite good though (and it was cheap!)

Detail of forward looking camera mount. Camera platform “floats” between two layers of green Kyosho Zeal vibration absorption sheet so there’s no hard connection to the body of the aircraft. Works great for killing Jello.

Canon A2200 with FPV camera (Sony 600TVL Super HAD) peering over its shoulder. This allows for a video feed while taking still shots. This configuration weighs 1,556 grams without battery. It seems heavy.

Canon A2200 on 45 degree mount for oblique aerial views. 

Canon A2200 on vertical mount. Use of this requires caution as it’s just asking to damage the camera during a hard landing. Also it’s prone to dusting from the rotors. Note the use of the Zeal gel to the left of the camera to kill vibration.

Detail of the vertical camera mount and vibration isolation setup. The arms are insulated from the main body by neoprene strips top and bottom. The camera mount is further insulated from the body with Neobond washers. Oh yeah, the landing legs are sliced sections of PVC rain gutter downspout. Cheap and VERY springy!

Vibration isolation stack for the camera mount using “Neobond” washers. They are rubber washers bonded to steel washers (I’ve flipped one over on the lower right). The center washer floats the camera mount away from the main body. There are no hard attach points to it.

The disassembled Quad and accompanying equipment after packing it into a remote desert location. Tarp was to provide a dust free landing/takeoff area to protect the cameras. I failed to consider the propeller extension in my design and discovered they stick out in transport, leading to bent props. Lesson learned.  

Construction notes

The following are some shots taken during the construction which show the inner workings.

Arm components with G-10 fiberglass upper and lower plates and Poplar wood core. One of the assembled arms weighs 90 grams. The G-10 is 0.046″ thick.

Dimensional drawing of arm design. For those of you that know of such things, these dimensions result in a configuration roughly equivalent to a 560 mm class quadcopter.

Detail of arm end showing wood core. I extended the G-10 plates out to act as a protective skirt for the motors. When you crash as much as I do you think of these things. The wood is .500″ thick.

Body components during epoxy process. The bottom plate has the larger holes. I made the wood web here Balsa instead of Poplar and it has proven to be plenty strong. The interior depth is 0.800″ and allows for running the wiring.

The assembled arms and partially assembled body. The holes are for lightening and to allow wiring to pass through. The black strips on the arms are the neoprene spacer/isolaters.

An earlier version with a camera mount that didn’t work. Too much vibration resulting in Jello video. With the camera mass out on the “diving board”, it oscillated too easily. So I stiffened the mount, brought the camera inward and added Zeal vibration isolation gel.

All in all I’m pretty happy with the result. It flies well, for around 9 minutes and feels like it has a lot of power. The APM flight controller provides amazing abilities, some of which, like fully autonomous missions, are further up the learning curve from me. But I have benefited greatly from its GPS and the ability to simply throw a switch and have it come back to me from wherever it is, under its own control, and land. That’s just friggin’ amazing.

This design has been fantastically strong, due to the composite construction. During one crash incident I flew through a tree at 20′, then impacted a concrete V ditch. I figured it was toast, but the result was only a couple of broken props and minor scrapes to the arm ends. In a second crash (see a pattern) I managed to fly it into desert sand at high speed, sideways (the upside was I learned how windy is too windy). Again, replace a couple of props and it’s ready to go. Its robustness has greatly exceeded my expectations.

Things I found that worked:

• The composite vibration isolation concept.
• Floating components using rubber washers.
• Kyosho  Zeal rocks!
• Color coding ESC wires for mindless connections.
• Multiple camera mounts.
• CHDK is REALLY useful.

Things that sorta suck:

• Doesn’t transport well enough. Maybe next version will have folding arms.
• Prop blast on dirt surfaces and delicate cameras don’t play well.
• I could have used thinner G-10 and perhaps Balsa instead of Poplar for the core.
• Need better wire management in the body. It’s a rat’s nest.

Note from Tom: This search trip and writeup is brought to you courtesy of Adam Marsland. Blame any typos on him. But if you think it’s written well I’ll somehow take credit.

Date: 4-14-2013

Participants: Adam Marsland

General Search Area: Mountains on southeasterly side of the mouth of Smith Water Canyon

Rationale for Searching This Area:

Still the most likely place for Bill to have wound up, and wanted to add some redundancy to previous searches and also further explore an area from previous trip.

Impressions of Area And Findings:

The approaches up Smith Water Canyon are, as has been previously stated, hella steep. The terrain levels out a bit once you get high up. Many jumbled rock formations.

During the search, I was able to make a call from a spot that I think may have a significant probability of being the source of the mysterious Sunday cell ping.

Coverage Level:
Good in the higher elevations. I kept my eyes open bushwhacking back down, but once darkness fell I was focused on saving my own skin. Coverage after dark, with a headlamp, was surprisingly good (albeit in a very low probability area). The headlamp did a good job illuminating the brush…probably more so than daylight.

Comments:

Owing to an unexpected trip to Las Vegas and an equally unexpected detour on the return trip south, I found myself in the Joshua Tree area with a little time to kill. I was tired and my feet were aching from the previous night’s wedding gig playing songs I didn’t know for three hours in ill-fitting hushpuppies, but I wanted another shot at the Bill Ewasko search before the summer heat set in and touring commitments would take me physically out of the area for several months. After debating which end of Smith Water I wanted to tackle, I decided to go back over the area on my first trip, hopefully fill in a few spots, and if I was lucky and had enough stamina get back to the open area on top of the mountains where I’d felt the strongest sense that Bill might be nearby on my previous trip.

I parked the car on Park Road to set out just before the ungodly late hour of 4 p.m. I had noted at the park entrance that sundown was at 7:15. I was dubious how much searching I could do before daylight faded (it’s three miles from that point to the mouth of Smith Water), but not being an early riser by nature and living 2 1/2 hours from Joshua Tree, I figured it at least would tell me if I could get any real searching done with that late a start. The answer turned out to be yes, but with a significant increase in the risk level attached to the hike. Don’t try this at home, folks.

I made good time down the various washes toward Smith Water. I made a call out from Quail Wash and it occurred to me that one of the only places you can actually ping the Serin Tower AT 10.6 miles in the area is from Quail Wash. I had been shocked to notice for the first time a large barrel hidden inside a bush that I remembered from my last trip through this area, and reflected that it really is a lot easier to miss something hidden inside a bush then something laying in a rocky area. I mused that, despite all of the traffic through Quail Wash, it wasn’t inconceivable that Bill could still be there, having expired at some point heading north towards Joshua Tree. If he had taken shelter under a large bush, people could have passed within feet of him and never noticed. And there are lots and lots of bushes in Quail Wash. (Not to mention quail, which weren’t present the last time I was here)

I pushed these speculations aside as I headed back into the bowl above the mouth of Smith Water. After taking a quick look at one of the hills I had searched on my previous trip, I decided to angle towards the left (southeast) side of the bowl and try to angle upward towards the high ridge above. I was dubious I had enough time to make it, particularly since, so as to not duplicate previous searches, I tried to take a fairly idiotic path upward. This notwithstanding, I was surprised to note footprints in one particularly sketchy spot. Looking at Tom’s search grid, this appears to have been one of his crew. Shortly after that point I diverged upward and started making my way towards the top of the ridge.

I had gotten most of the way up when I made probably the most significant discovery of the day. In climbing up I had circled around a rocky area from the back and, completing the maneuver, stepped over a low ridge so that after circling south I was again facing north towards the canyon. Upon doing this suddenly I could see Joshua Tree village in the distance. Based on my experience the previous week on the north side of the canyon, I felt pretty certain I could get a call out at that point. I tried it…and succeeded.

This was significant since judging by my landmarks, I was closer to the 10.6 mile radius and also lower in elevation than the points where I had previously been able to get a call out from this ridge, further to the west. Moreover, I was only a few feet from that low ridge I had just crossed, behind which was ALL lower lying land all out of sight of town. This fit all the criteria for the mysterious Sunday morning ping, since someone could conceivably have traveled from the Quail Mountain area all the way to this point in the dark of the cell tower, cleared the low ridge, and immediately have cell reception. Once I got home and reviewed the location and GPS tracks, this impression was confirmed. If someone traveled all the way north from Quail in the cell phone dead zone, this is the spot they would wind up. And from what I could tell, once there the cell reception would be sudden, and strong. If someone had been trying for three days to get a call out, and they hit this spot…well, I could imagine the battery going pretty fast.

My personal feeling is that this spot has a high likelihood of being the source of the ping. To me, it all fits.

Just to my right at that spot was a tower of rock. Remembering Tom’s theory that Bill’s final cell ping could have come just prior to a fatal fall, I searched the area thoroughly. No Bill.

I continued to head east, aware that the sun was beginning to set, but still wanting to find the level terrain area at the first summit of the ridge I had come upon in my first trip to the area. I had felt, more than anywhere else, that Bill was in that general vicinity; it offered the most places to hang out and take shelter. I could not find it, however, even though I know I was in the area. My sense was that I was just above and behind it My desire had been to search just to the west of that spot, so I struck out for the area that I thought was in about the right place (from what I could tell reviewing the tracks later, I was in roughly the right area). There was a high rocky cliff and an area below where a fatal fall could easily occur, and two other similar spots just yards from me. I scanned the landscape that rapidly dropped away below carefully. Nothing.

At this point, with the sun beginning to set, I had to think about getting back. I decided to bushwhack straight back toward the car, which would take me through the Quail Springs area, an area which I remembered Tom had expressed a desire to search more thoroughly. I knew a dusk bushwhack through unfamiliar terrain was risky (and remembered a comment that the terrain above Quail Springs was particularly steep), but I figured it couldn’t be any worse than Smith Water and I might gain some unexpected insight into Bill’s frame of mind racing the clock.

At first, the terrain was fairly even and I did a good bit of looking around as I went down. I could see from foot and pole prints that previous searches had covered this area, but I did poke around a few rocky spots, including one just south of my suspected cell phone ping spot. As I progressed north towards Quail Springs though, it became apparent that the way down would get steeper and steeper the further I got, and I was still a long way from my destination. As darkness began to fall and the terrain got less forgiving, the hike started to get harrowing. I did start thinking about Bill, conceivably finding himself in a similar situation in a bushwhack gone wrong that fateful Thursday, with that unhappy understanding dawning that the further down you got, the more dangerous it got, and the longer you took, the darker it got. A bad tradeoff, but I kept reminding myself that you’re not REALLY in any danger until you injure yourself, and took proper care as I negotiated the ever steeper descent into the area around Quail Springs. I had a hiking stick; he didn’t. Looking around, it wasn’t hard to imagine how Bill could have gotten himself into an odd spot. After all, the only reason I was where I was was follwoing a line of sight from where I had just been to the arm of a hill that was in line with my car. Otherwise, there was no reason for anybody to be there.

Once I got down (EXTREMELY relieved) into the wash that snaked up into the hills towards the spot I descended, the wash was rocky enough that rather than following it out of my way I kept moving in a straight line over a low hill that separated me from Quail Wash. Right at the point I finally put the last of the mountainous rocks behind me, I came across a half of a jawbone in my path. It had no fangs and human-sized teeth, so I took a picture and some film. Tom later indicated it was probably from a deer or a mountain goat. Not human.

The slog back to the car was, just like last time, brutal. I hit the road nearly a mile from my car and had I not had the GPS track to refer to I would have never found it. Interestingly, once I donned a headlamp to make my way back, I had much better visibility into the brush than when it was daylight. Thinking back on my idle Quail Springs theory earlier, I fancifully envisioned a group of night hikers, dressed in head lamps, walking up Quail Springs doing a lo-tech X-ray of the brush therein. And trudged on.

GPS mileage for this trip: 8.9 miles

Cumulative GPS mileage to date: 685.6 miles

GPS tracks for this trip in Google Earth kml format

GPS tracks for this trip in Garmin gdb format

GPS tracks for this trip in gpx format

JT51 tracks

JT51 tracks in blue, original search tracks in black, tracks since in red and Serin cell tower 10.6 mile radius in dashed green.

A detailed view of the JT51 tracks, in blue, in the SE Smith Water Canyon mouth area.

Date: 4-15-2013

Participants: Tom Mahood

General Search Area: Southeasterly slopes of Smith Water Canyon mouth and slopes above Quail Spring.

Rationale for Searching This Area:

Explained in way too much detail here: Current Bill Ewasko thoughts, January 2013.

Impressions of Area And Findings:

The usual treacherous terrain, although I must be getting used to it as it’s no big deal anymore.

Coverage Level:
Very thorough. I moved rather slowly and did a lot of scanning with binoculars. I’d say my coverage level was 80%, out to about 100′ on each side of my travel path. I also spent considerable time doing binocular scanning of the northerly slopes of Smith Water Canyon, from its mouth to about the eastern grove of Acacias. There had been some thought that Bill might have been headed down Smith Water and possibly ascended the unsearched north slope in an attempt to get a cell signal.

Comments:

A lot of miles have been spent searching the area southeast of the mouth of Smith Water Canyon to no avail. On many levels it makes sense, that Bill may have been attempting a cross country trip to Smith Water Canyon in search of water, but it is just not working out.

As we’ve probed this area there have been a few issues in regards to cell phones that caught my attention. First is that there seems to be better cell phone coverage in the upper regions than the splash maps show, at least according to what we’ve seen in the field. Second, the idea that Bill’s cell phone suddenly goes dead just as he gets to a cell zone strikes me as a remarkably unfortunate coincidence. But maybe. And finally, the measurements by Mike Melson suggest that distances measured by the Verizon system out of the Serin Drive cell tower are fairly accurate path lengths.

So here’s what’s been troubling me. Anyone coming over the top runs into some pretty good areas of cell coverage, we’ll out from 10.6 miles. And some of it is even out further than 11.1 miles, the distance considered within reasonable error. So if Bill did come over higher up, the ping should have been very strong and occurred further out than it did.

Now that changes if the travel route is lower down, and more northerly. In playing with Google Earth, I found a possible path that slowly descended from the upper reaches of the Quail Mountain area into this area without passing directly through the cell coverage zone. It skirted it just to the north, essentially just above Quail Spring. Someone traveling this path would have a continuous downhill route (important if injury is involved) and would pass just below the cell coverage area and thus could generate a single transitory ping at the proper distance.

I usually work out a GPS search route in advance of heading to the field and somewhat slavishly follow it to ensure I get the coverage I wanted. It’s sort of a mechanical, brute force method that essentially ends up doing a line search over time. I wanted to try something different this time.

What I did was ascend the slopes southeasterly of Quail Spring until I intercepted the general path I had found. I then turned to Smith Water with the view of someone who needed to get into that canyon for water and followed the terrain to best do that. I just followed my gut. I knew where the cell coverage zone was above me, so I deliberately made an effort to stay below it. I did note that from the vantage point of my approach that bowl southerly of the mouth of Smith Water could possibly be mistaken for Smith Water itself by someone unfamiliar with the area. That led me to climb the ridge between the bowl and Smith Water proper, which is almost on the 10.6 mile radius. It was along here I scanned the north slopes of Smith Water Canyon up to the height of a reasonable climb up the side.

I eventually reached the saddle between the bowl and Smith Water Canyon, having covered all the obvious possibilities. I turned back easterly and traversed across the face while ascending the south slopes generally going over areas already partly searched. From there I descended back down toward Quail Spring and back out.

Frankly, I’m stumped. Perhaps I’ve succumbed to my own delusion (usually it’s others who are its victims) but I really felt Bill was in this area. I mean, my gut still tells me that but my head says, “Look at all the red lines on the stinkin’ map! We’ve filled it in very well. He’s not there!” (My head can be rather annoying at times). Now I will concede we haven’t covered ALL possible spots in that locale, but I’d guess it has to be close to 85% of it.  So I just don’t know what to make of it. It’s like I’m working an arithmetic problem and keep getting 2 + 2 to equal 5. I know it’s clearly wrong, but just can’t see why. I suspect it will be a while before I head back out there for another search. Obviously I badly require new ideas and need to pout about this a while. I. Don’t. Like. Being. Wrong.

GPS mileage for this trip: 8.5 miles

Cumulative GPS mileage to date: 694.1 miles

GPS tracks for this trip in Google Earth kml format

GPS tracks for this trip in Garmin gdb format

GPS tracks for this trip in gpx format

JT52 tracks

JT52 tracks shown in blue, in context with the others to date. Black is from the original search, red tracks since then and the green dashed line the 10.6 miles radius from the Serin cell tower.

A detailed view of the JT52 tracks in southeast portion of the mouth of Smith Water Canyon.

The general route is shown in red. The green shaded areas are the approximate areas of cell coverage from the Serin cell tower. The black lines show the approximate distances to the cell tower.

Yes, it’s snake season! But not to be concerned, only a nice, pretty Rosy Boa.

Aside from the general geekery of the gadgetry, my primary interests for fooling around with drones/UAVs have to do with mapping and aerial photography. I love aerial pics. This all combines in a especially interesting way for me using drones to image archaeological sites. So I figured it was time to try. As far as archaeological sites go, New Mexico is a target rich environment. The damn things are everywhere, so what better place to try a bit of imaging? So I selected two sites to experiment with on an extended visit Jeri and I made to New Mexico.

The first, unnamed site

The first site is a very unusual and intriguing location I had visited a number of times previously. I’m not going to go into any detail on just what or where it is because of its unusual nature and its sensitivity. I will say it is on BLM land and has been studied very little, if at all. The area of interest is high on a rugged canyon wall (hence the appropriateness of a drone for imaging). I have noted some ceramic sherd scatter as well as lithic debitage at the location. Wood from the site has been radiocarbon dated to the period between 800 and 900 AD. The site sits above 6,000′, which would be an interesting challenge for my composite H-Quad quadcopter, which had only been tested at sea level.

Jeri and I visited the site on April 21 and 22, 2013. Both mornings had unfortunate high winds which greatly hampered flight operations (stinkin’ Spring in New Mexico!). The winds on the morning of the 21st were the better of the two days but I would normally consider both days too windy to fly. They were, however, the only windows of opportunity available to me. Jeri was assigned photographic duties to record any potential aircraft carnage of which I am prone to.

For the April 21st flights I was able to do two video flights and one still image flight. I took video with a Swann FreestyleHD sports camera (a FauxPro) and stills via a Canon A2200 14 megapixel camera running CHDK software. This software forced a shutter speed of 1,000 of a second and took an image every 5 seconds. Both cameras were positioned on a forward facing mount, angled down 15 degrees from horizontal. My area of interest on the steeply sloping terrain and weather conditions were not conducive to using a downward-looking camera mount for mapping purposes.

Because of the intimidating canyon wall confines I was not comfortable initially flying using my video goggles so instead flew line-of-sight on the morning of the 21st. Unfortunately this prevented me from flying as high as hoped since I needed to keep the quadcopter in sight and height estimation was difficult. In hindsight I realized didn’t really get high enough. The second video flight was somewhat higher but still fell short of where I wanted to be. Because of the turbulent conditions I had to keep considerable distance from the canyon walls so detail suffered. That, and the confines of the canyon walls themselves added to the difficulty.

The morning of April 22nd was even worse in terms of wind conditions but I chose to proceed. I’m stupid that way. The first flight was video and after being successful (or lucky) the previous day, I decided to risk flying it using video goggles. This ensured I could fly high enough to see all the way to the top of the canyon walls and gain the higher views the previous morning’s flights had lacked. When I did operate the quadcopter above the canyon rim it caught the full force of the northwesterly wind, headed like a bat outta hell to the southeast and required considerable angling of the copter to hold position against the wind. Thus a lot of the video ended up somewhat tilted as I have yet to add a self-leveling camera gimbal.

The second flight of that morning, the stills, was…entertaining. And not in a good way. Shortly after launch, while the quad was high along the canyon wall, the wind suddenly became so severe I momentarily lost control of the craft. I recall seeing flashes of sky in the goggles and at one point a rather small bush on the slope filled my view. This is not a good thing. As I could clearly hear the quad I ripped off my goggles and decided to bring it back down visually. This would have been a good course of action if I had known where to look, but the quad was hidden against the earth-tone background of the canyon wall. I could hear it, but couldn’t see it. This was especially not good.

At this point I’m figuring it’s not a matter of if I’m going to crash, it’s just when and the rest of the morning was going to be spent in a recovery effort. So as a last resort, gambling its orientation was somewhat level, I punched the throttle hard sending it screaming straight up from wherever the hell it was. After a few very long seconds it popped into view above the canyon rim and I had visual again. A few more sweaty minutes and I had it safely on the ground. Thus ended those flights.

The bright side to this episode (aside from the usual cheating of aircraft death) was that I got some acceptable video and some decent stills on the area I was interested in. Certainly good enough for a very first try. I also found my sea level quad flew just fine at this relatively high altitude.

Laying out the goodies and getting set up for launch. Note the FauxPro cam on the front of the quadcopter as this was a video flight.

The quadcopter is up pretty high in this pic and starting to have issues with the wind. I’m flying it line-of-sight. When I switched to FPV goggles I would eventually get it up a lot higher.

Jeri got a pretty good telephoto shot of the quad examining the slopes.

In this image I’ve pulled off the goggles and am attempting to wrestle it to the ground in the wind, visually. I won….eventually.

Another “wrestle it to the ground” picture. With winds, takeoffs are easy. It’s landing that can ruin your day.

Pejunkwa ruins

The second, much less exciting, archaeological imaging test was at the Pejunkwa ruins in the Santa Fe National Forest, not too far from Jemez Springs the morning of April 30, 2013. I had previously visited the site and reported on it here. Since it was so easy to get to, and somewhat flat, I thought it would make a good test. Also it was at 7,500′ which could be even more of a challenge for the quad.

The winds that day were also uncooperative, but not crazy like at the first site. It was probably a smooth, constant 15 mph out of the northwest. I wanted to try doing some mapping style imaging so I put on the downward looking mount for the still camera. This puts the camera beneath the frame of the quad, with the extended lens just inches from the ground. From previous, bitter experience, I found that if I lifted off the quad or landed it on dirt, the rotors blew dirt into the extended camera lens and jammed the mechanism. . To avoid this I had taken to laying out a white plastic tarp for takeoffs and landings in dusty areas.

My initial plan was to fly sort of a grid pattern above the pines and try and cover as much of the open rubble piles as possible. I used my FPV goggles for this flight since it provides a heads up data display, overlaid on the image from the craft, and I could maintain the constant altitude I needed to get a good panoramic shot.

After liftoff I could clearly feel the wind’s effect on the craft. I ended up holding it at 300′ above the ground, much higher than I had originally planned (I often make this stuff up as I go along).

The wind was also causing problems with the quad’s orientation. I needed it to fly as level as possible to ensure the camera was pointed straight down. However if I wanted to fly into the wind I had to angle the craft quite a bit and the camera was no longer looking downward. After a couple of minutes of that I found the best approach was to fly fast and hard up into the wind, then let the wind drift it back downwind holding a fairly level orientation. This didn’t give me all the images I needed, but I got some that were useful.

Surprisingly, the quad operated pretty well at that altitude, although I noticed I had to use substantially more throttle. Unfortunately at the end of the first flight I landed stupid, nicked a prop and had failed to bring any spares along with me. So I only had the single flight at Pejunkwa.

Looking at the images was something of a disappointment. Lots of off-angle shots and the software I was using for image stitching, Microsoft Image Composite Editor (pretty decent and free) was just doing a crap job with my crap images.

Then I found AutoPano Giga from Kolor. Wow…It is the best image stitcher I have ever seen. It was able to take my junk pictures and turn them into a mediocre, but acceptable, composite image of the ruins area. It can take a series of images, at different rotations/orientations and combine them into a seamless mosaic…all friggin’ automatically. It’s magic I tell ya!

Here’s what I was finally able to produce from 31 individual images of Pejunkwa using Autopano Giga. I’ve shrunk it down to around 1.6 megs in size from the original 37 meg jpg. Some areas are darker as there were clouds moving across the sky and some images were taken while in the shade. Yeah, I know, it’s still pretty crap, but not as crap as it should be considering the individual source images. I prefer to think of it as a highly polished turd of a panorama.

An aerial of the Pejunkwa site, with north being roughly up. The “launch tarp” is the white square at the upper right of the clearing. The clearing is actually the crumbled roomblocks.

Pejunkwa Ruins (Adapted from “Overview and Synthesis of the Archeology of the Jemez Province”, New Mexico, by M. Elliot, 1986

Lessons learned

• Imaging flights need to be in fair wind conditions.
• Image intervals of 5 seconds are too far apart (I have since tweaked the setup to take images at a much better 3 second interval).
• A camera gimbal would be a very good thing to have.
• Always have spare props along.

I don’t know why they say there is a learning curve to all this. Curves are smooth. This is proving to be the learning jaggy-pointy-peaks of death.

It had been a while since Jeri and I had been on a School for Advanced Research field trip and we were looking forward to this one and reconnecting with old SAR friends. Since we were coming from a different direction this time, we skipped leaving out of Santa Fe and met them along the way in our own vehicle.

We hooked up with them on the morning of April 19 at the Southern Ute Cultural Center and Museum in Ignacio, CO, where we were given a tour of the facility. The structure’s architecture was about as good as I’ve ever seen, clearly done by someone very creative. The exhibits were also excellent with a high level of information transfer. I have a strong loathing of most newer national park visitor centers as I think they dumb down the experience in the interests of giving short attention span people what they want: Pretty pictures and video screens. The Ute facility was done the way visitor centers should be, a stunning structure with real information inside.

Leaving that facility, and after a lunch stop at Hesperus, we stopped at the newly constructed Mesa Verde National Park visitor center for a behind the scenes tour. While it was operational for the public, it hadn’t formally opened and some landscaping was still in the works. Having just come from the Ute facility, I couldn’t help but contrast the two structures.

The new Mesa Verde visitor center was certainly nice, but it looked like it had been designed by a committee of committees. There was a bureaucratic blandness to it all. Inside, it seemed most of the spanking new public space was devoted to either the gift shop or the area to queue up to buy Mesa Verde tour tickets (Truly an “American” national park…oh wait…it didn’t have a snack bar). There were precious few of the artifacts recovered from Mesa Verde on display. Well…..There was a window in a wall that allowed the public to see some of the real collections area…where everything was carefully stored away in boxes and out of view. Not very satisfying.

It was, to me, pretty underwhelming. I had only been to Mesa Verde once before. Many years ago Jeri and I went early in the day and took a tour, chafing under the corralling of the park service and dodging the hordes of mouth breathing tourists. Then later that same day we went to Chaco Canyon for the first time and after that, well…. we never had any interest in returning to Mesa Verde. Been to Chaco more times than I could possibly count though.

After overnighting at the Blue Lake Ranch (VERY nice place, BTW!) the group headed out to meet our guides who would take us into the Ute Mountain Tribal Park. It could best be described as “Mesa Verde Adjacent”, as it wraps somewhat around Mesa Verde. However it’s owned and run by the Utes, and their sensibilities are just a wee bit different from those of the park service.

First, you can’t even go in without a guide, but it’s not an expensive thing to arrange. There are no nice concrete trails or steps (sorry ADA people) and it’s all dirt roads. The trails are often narrow and perhaps even dangerous in the minds of some. And the ladders, made from pine trees, used to gain access to some of the cliff dwellings will absolutely freak some people out. It is a very rustic experience. Finally, there is refreshing lack of what the national parks folks like to call “interpretation”. You know, those kiosks and signs telling you what you’re supposed to think about what you’re seeing. At the Ute Mountain Tribal Park, what it, is. Think for yourselves. But if you have questions, ask the guides and you’ll get an interesting story. It may even be right, but it will be good.

But it is all so totally worth it! It’s like what Mesa Verde must have been 60 years ago. As long as you don’t do anything stupid, the guides permit you to walk through the ruins to your heart’s content. There are also numerous pottery sherds and other artifacts that visitors are encouraged to pick up and examine. Try that in a national park and you’ll see the rangers reaching for handcuffs. Oh, and our group were the only ones at these particular ruins that day and we never saw another person. In summary, it was sort of a Bizarro World of a national park where you are welcomed to get down and dirty with the ruins. It was friggin’ fantastic and we’ll definitely be going back.

Looking across Lion’s Canyon to the Eagle’s Nest ruin. Note the 40′ ladder to the right and below of the ruin. We’ll see that again later, MUCH closer.

Rick, our Ute Mountain Park guide explaining the Lion House ruin on the far side of the canyon from us. We were later to walk through it. These structures we were visiting were originally built in the 1100s and completely abandoned by 1250 AD.

A sense of scale. Some of our SAR members walking above a very precariously located set of ruins.

That’s the Tree House ruin behind the trees to the left. We are starting down the first set of ladders to get to it.

A closer look at part of the Tree House ruin.

Now at the Tree House ruin.

The inside of a kiva at the Tree House Ruin. It was in excellent condition with the walls still plastered.

The guides encouraged us to pick up and examine whatever were wanted as long as we put it back where it was and didn’t do anything stupid. This is a collection of mostly pottery sherds and some lithic debitage. Toto, I don’t think we’re in a national park any more….

A really nice granary door showing the sealing.

One of several T doorways among the ruins. These seem to have some sort of connection with the Chacoan people, where these doors seem to have originated. No one quite knows why they are shaped this way.

Here’s a shot looking back from the Tree House ruin at the series of ladders we used to descend to the trail that took us to the various ruin sites.

Jeri doing touristy stuff at the Lion House ruin.

At Lion House a collection of manos and metates…..which we again were encouraged to handle. Too cool!

On the trail between the ruins. Why is Jeri smiling? Because she’s not at Mesa Verde!

Looking down into another kiva showing part of its roof. Note the rather unusual deflector construction on the left side of the kiva. Typically a single stone slab was used block the ventilation from the fire pit. This one is fancy!

Maize cobs at the bottom of a kiva. These are perhaps 1,000 years old.

Ummm, remember that 40′ ladder in the first pic? Well this is it. Rather intimidating and the only way to get to the Eagle Nest ruin.

Now at the top of the ladder, the route into the Eagle’s Nest ruin requires you to duck walk under the overhang to the right. NOT for anyone uncomfortable with heights!

Another granary at Eagle’s Nest. And someone’s ass.

A kiva at Eagle’s Nest with some of the wood roof support still in place. Note the deflector on the floor to the left is the usual type of deflector construction.

This was yet another of the many School for Advanced Research field trips Jeri and I have taken.

Although not terribly “archaeological” the first day was spent at the semiannual Hubble Trading Post Indian art action. There were a lot of exceptional deals on Navajo rugs and assorted pottery. As a result we spent way too much money. There’s some nice stuff to be had there.

After blowing our discretionary cash at the auction, Jeri and I headed back to Canyon de Chelly to do the White House ruins hike. Except for the trail to these ruins, you can’t go into the canyon without a guide. The trail is about 3 miles round trip and a 600′ descent then climb back out on the return. I can’t imagine it would be a fun thing in Summer when hot and full of touristas, but it was great for us. The ruins weren’t any great whoop, but it was a good excuse to be hiking in the canyon. It’s a real interesting trail with two tunnels on the route.

Canyon de Chelly is an odd duck in terms of its status. It’s a national monument, run by the park service. They are responsible for preserving its antiquities. But the canyon is part of the Navajo reservation and the land is owned by the Navajos. A certain number of Navajos live in and around the canyon, driving in and out. There are Navajo roadside vendors selling assorted native jewelry and trinkets, as well as inside the canyon to tourists. Compared to other national parks/monuments, there’s a lot of litter and general crap around. A very weird vibe to the place. It struck me as sort of a “Mad Max” version of a national park.

The SAR group dispersed into a number of vehicles with guides and we headed up the lesser visited Canyon del Muerto. We traveled fully to the end of the dirt road, at Mummy Cave ruin. Along with our group was the park archaeologist, Keith Lyons who provided commentary. In all instances the ruins were fenced off and we weren’t able to get close. This was sort of unusual for a SAR trip, as they usually have really special access to sites. However on the way back out we stopped at a series of petroglyphs known as the Standing Cow petroglyphs. Turns out that particular parcel of land was owned by one of our Navajo guides, Sylvia Watchman. She invited us through the fence and up to the family’s ancestral hogan. Sherds and lithics were everywhere, and we spent some time just staring at the ground looking at all the “stuff”.

Following a couple of the vehicles getting stuck in the stream bed (lots of water crossings), we made it back out after a very loooong day.

The view from the trail down into the White House ruins.

The White House ruins at the canyon bottom. These are the only ruins the public can visit without a guide. Not too shabby and a nice 600′ drop/gain to get there.

More ruins, the Lodge ruin, I think. Note the neat swastika petroglyph above the ruins.

The park archaeologist, Keith Lyons, points out features of the Mummy Cave ruins to our SAR group.

Jeri pondering the warm welcome of the park service. Don’t even think of doing anything interesting.

A hogan at the Standing Cow petroglyphs. This property belonged to one of our guides, who was raised there. She invited on to the land for a close look.

More at Standing Cow and some of the many petroglyphs.

So what’s a guy who’s overly fond of southwest archaeology doing in Illinois, just east of Saint Louis,  looking at mounds of dirt? Good question.

Turns out these mounds of dirt are the remains of what was probably the greatest city in America north of Mexico. Yes, bigger and badder than Chaco Canyon. Until recently I had never heard of Cahokia, but the more I found out the more I wanted to see it. Curiously, its rise and fall seemed to be timed similarly to the Chaco/Mesa Verde areas, although there’s no evidence of their interaction.

Unfortunately Saint Louis is freeway close to nowhere I usually go. But then due to a variety of circumstance, I ended up with a Southwest Airlines ticket to anywhere I wanted, so off to Cahokia I bounded.

Cahokia was substantially occupied from around 700 to 1400 AD, by which time it was completely abandoned. It reached its 50 year peak, beginning in 1050 AD, with a population of perhaps 20,000 people over 6 square miles. It was the Chaco of the Midwest….and then some.

An excellent book on Cahokia is “Cahokia, Ancient America’s Great City on the Mississippi” by Timothy Pauketat. Pauketat suggests that there is some evidence that growth and expansion at Cahokia went off the charts starting around 1050 AD around the same time a massive supernova was observed in the sky, the same one recorded at Chaco Canyon in the Penasco Blaco pictograph. Pauketat also goes into detail about the substantial indications of human sacrifice at Cahokia. Apparently it’s not just an Aztec thing.

In his book, “A History of the Ancient Southwest“, archaeologist Steve Lekson muses about a possible New Mexico/Cahokian connection. Many are familiar with the Spaniard Coronado and his expedition north from Mexico in 1540 seeking Cibola, the alleged seven cities of gold. After passing through many New Mexican pueblos and finding, well only dirt and stone, Coronado eventually met an Indian guide near Pecos who claimed he could show Coronado the way to Cibola. The guide lead Coronado to the northeast, well into Kansas, where Coronado finally lost patience with the guide (or was just sick of Kansas), killed him and headed back for Mexico.

Now conventional historical wisdom regarding the affair says the pueblos were just fed up with Coronado and had the guide lead him on a wild goose chase to anywhere but here. But that seems like a pretty risky thing to do when dealing with a prickly Spaniard with pointy steel weapons. Lekson says some historians believe the Indian guide was actually a member of the Wichita tribe. If that’s the case, it may well be that the guide could have been acting upon stories told to him by his grandfather, of a rich city to the east. So perhaps Coronado was in fact being led to the remnants of Cahokia. Saint Louis = Seven Cities of Gold? Don’t know, but it’s a good story anyway.

Now admittedly I had low expectations for the place. It apparently wasn’t worthy of national park status and was only a lowly state historic site. And it was, after all, Illinois. Having originated in Illinois myself I understand full well what the state’s capable of. I had seen so many impressive southwest ruins that this place couldn’t possibly be anything too interesting, despite what I read. Turns out I was very wrong.

My first inkling this was going to be something great was driving onto the massive grounds towards the visitor center. The place was immense with mounds everywhere. Monk’s Mound (named after some monks who lived near it many years ago) is over a hundred feet high and covers 14 acres! You could probably take all the great houses combined in Chaco Canyon and drop this sucker completely over them (A Cahokian house dropped on the wicked witch of the Southwest??). And it was only one of, well….craploads of mounds.

And the visitor center was a stunner. Maybe the best I’ve ever seen. Packed with real artifacts from Cahokia and copious information. An amazingly cool exhibit was a reconstructed village in the middle of the visitor center. A small cluster of thatched huts and that sort of thing. But the trick was they surrounded it with semi-reflective glass walls. This gave an infinity effect of the village extending off into the distance. Combined with the audio sounds of animals and children playing, it was extremely effective at creating a sense of place. I usually find such attempts rather sad and cheesy. This was neither.

Leaving the visitor center I started exploring the grounds. Now I will freely concede all there really is to see are dirt mounds of various shapes and sizes. However things become very impressive when one has knowledge of what it took to construct these suckers and an understanding of the community surrounding each of them. Oh, and most of them probably still had considerable human remains and other artifacts still buried within them. As much as I’m impressed by Chaco Canyon and its outliers, in some ways Cahokia impressed me more.

The only way to really get your mind around the size of it all is to climb Monk’s Mound. Figure ten stories worth of stairs. But once at the top you get a sense of just how big Cahokia really was. Except that it was even bigger. Originally it stretched all the way across the Mississippi River to Saint Louis, where in the 1800s a mound the size of Monks Mound was scraped away in the name of progress. And you can see Saint Louis, way off in the distance. Damn…..This was quite a place!

So if you’re anywhere near Saint Louis, or even if you’re not, Cahokia Mounds State Historic Site is more than worth a visit.

The rather amazing Cahokia Mounds State Historic Site visitor center. This thing is huge.

Wow, a visitor center that has real stuff on display! And what cool stuff it is!

Yeah, them’s Clovis points! About as old as it gets in the US, around 13,500 years ago.

More very cool displays.

Some of the Cahokian trade goods.

A few of the hundreds of ax heads blanks (also known as “Celts”) found near Cahokia. They would be worked into sharper edges at a later date.

Chunkey stones! Chunkey was a field game played by Cahokians that involved throwing wood spears at rolling stones the size of hockey pucks. And much betting. It may have been a contact sport, a la hockey. Crazy Cahokians….

A collection of flintknapping tools and partially completed stone tools.

Cahokian pottery. Not as flashy as some southwest pottery, but functional. And some pieces were pretty damn big.

More Celts, some already sharpened into working ax heads.

A view of the largest Cahokian mound, Monk’s Mound, from the visitor center. It doesn’t look too big because you can’t see how far away it really is.

A reconstructed portion of the 2 1/2 mile pole wall that once surrounded the main area of Cahokia. They must have wanted security from someone as it was rebuilt 4 times and incorporated “ramparts” pushing out every 85′ or so to allow bowsmen to fire on attackers. This is seriously defensive stuff.

A couple of the twin mounds south of the main Monk’s Mound.

On approach to the main mound, Monk’s Mound. The people on the upper stairs give a sense of scale. The cars are traveling on a four lane highway that cuts past the base of the mound, a very weird sight.

Looking back at the visitor center from the top of Monk’s Mound. The visitor center is 0.4 miles away. One of the Twin Mounds is visible to the right, due south. All that green open space is part of the site.

Off is the distance to the west is Saint Louis, 7 miles away.

Looking to the east of the main mound are a couple more sections of the reconstructed defensive wall.

Woodhenge, a Cahokian solar calendar. This thing is over 400′ in diameter with 48 posts. It’s lined up with the main mound in the far background.

I can’t really take much credit for this idea. I first saw it as a posting by BloomingtonFPV on the FPVLab.com forum on January 6, 2013, so he deserves the bulk of the credit for this. He did a LOT of work on it. Here’s a link to that thread as there’s a lot of good, detailed info there beyond what I’m presenting here and you should read it also .

Ummmm, the problem was I couldn’t get the damn thing to work. Delving deeper into the code posted it looked like pieces were missing, especially in the Arduino code. It looks like something maybe got hosed during its posting. So I jumped into the code and put in what was needed to get it to work and did a little streamlining. I also added annotations so that others could better see what it was doing. I know I had a hell of a time figuring it out myself so you’re benefiting from my failings. Furthermore, while it was originally written for a Canon SX230 HS camera (a GPS enabled camera) I’ve tested it on not only on that camera but also a Canon A2200, a smaller, lighter and cheaper non-GPS camera. And it works. It will probably work on a lot of Canon cameras.

Why you want this

If you are using a Canon camera for aerial photos, a model supported by the CHDK software, this device allows you to take off with your camera lens safely retracted. Then when you are where you want to be to take images, throw a switch on your RC transmitter and the camera comes alive, extends the lens and starts taking pictures at an interval you’ve specified. When you’re ready to land, you throw that same switch on your transmitter and the lens retracts and the camera shuts down. This is ideal to protect a belly mounted camera in a fixed wing aircraft. It also is useful in the case of a multirotor setup where the camera is facing straight down beneath the aircraft and susceptible to blown dirt and grit from the rotor blast.

The complete Arduino system, minus the camera ’cause you know what that looks like. You may only need the FTDI board once for the initial programming.

How it works

Most RC transmitters send out Pulse Wave Modulation (PWM) signals (If yours isn’t PWM, you can stop reading now). By assigning a three position switch (or an analog knob) to a spare channel on your transmitter you can send PWM signals that range from low to medium to high.

An Arduino microprocessor sits between the RC receiver in the aircraft and the Canon camera. The Arduino is programmed with the “USBController.ino” to monitor the PWM signal coming out of the aircraft’s RC receiver. If the PWM level is less than 1250, the Arduino generates 15 millisecond pulses on the USB input to the camera. If the PWM level coming from the receiver increases to between 1350 and 1650, the Arduino starts generating 45 millisecond pulses. Finally, if the PWM rises to 1750 or above, the Arduino generates 90 millisecond USB pulses.

Here’s the cool part. While the Arduino is doing its thing the Canon camera is operating the CHDK software and running the script “PictUAV2.lua”. It watches the signal coming in on the USB port. When it sees only 15 millisecond pulses it just sits there in standby, sucking its thumb. Pulses greater than 30 milliseconds but less than 90 cause it to extend its lens and begin taking pictures at rapid intervals. And when the pulse lengths increase to 90 milliseconds or greater, the camera shuts itself down. Sweet!

What you need to make this happen:

• Arduino Pro Mini 328 – 5V/16MHz (Available from Sparkfun.com for $9.95 ….that’s way cheap!). You want the 5 volt version, not the 3.3 volts as the power is coming from the RC receiver and is 5 volts. This is a tiny friggin’ unit!
• Mini USB cable to chop up
• Canon Camera running CHDK software
• “USBController.ino” to program the Arduino
• “PictUAV2.lua” script to run the camera under CHDK
• A programmer for the Arduino if you don’t already have one. The FTDI Basic Breakout – 5V from Sparkfun for $14.95  will do it and allow for programming of other Pro Minis.
• Maybe some header pins like these  to make connecting to the Arduino easier. You certainly need some to connect the FTDI board.

Putting it together

Here’s the super simple wiring diagram:

Pin guide for Arduino Pro Mini 328 – 5V/16MHz

Solder in four header pins in the Arduino’s following holes:, 10, 12, VCC and GND. You’ll also need to solder in 6 header pins for the FTDI programmer along the bottom of the board as shown in the image above. Ummm…You do know how to solder, don’t you?

Power for the Arduino and the PWM signal are obtained from the aircraft’s RC receiver. Using a typical servo cable connect the +5 volts (middle wire), signal (usually white or yellow)  and ground (usually black or brown) to the Arduino header pins at the locations shown.

The USB connector you need is the very common Mini Type A. Chances are you have some of these in a box somewhere ready to chop up. You’ll only use two connections to the Canon camera, contact 1 (ground) and contact 5 (+5 volts).

There are two ways to figure out which wires go to which USB contacts. The first is to do a continuity check between the contacts until you find the correct wire. Hold the mini USB plug so that you are looking into its opening and the five copper contacts are at the bottom. The “top” of the pin is narrower than the bottom. Looking at the contacts from this perspective, contact 5 (+5 volts) is the one at the far right and contact 1 (ground) is the pin at the far left.

The other way to get the correct wires is to just look at the color of the wires in the USB cable bundle. The wire to pin 5 is SUPPOSED to be red and the wire to pin 1 is SUPPOSED to be black. Those are the standards. If the USB cable you’re slicing up is extraordinarily cheap and made by slave labor in a remote country….Then who knows. But my colors were correct.

So contact 5 on the USB plug (probably the red wire) goes to pin 12 of the Arduino and contact 1 of the USB (probably the black wire) goes to the GRD pin of the Arduino.

Wow, you’re all done with the hardware!

Software

I have links to the two software items below. I put each on a separate web page so just copy and paste them at your end with whatever text editor you like.

To program the Arduino you’ll need the free software from Arduino  and the FTDI Basic Breakout board. The reason you need the FTDI board is that this version of Arduino is designed to be tiny so they left out the programming interface. You could use a larger Arduino if you like and not have to use an FDTI interface. The Arduino website has all the programming info you’ll need. In summary you’ll install the Arduino software on your PC, plug in the Arduino to be programmed via a USB cable and the FTDI programmer, load the USBController.ino code into the Arduino software and upload it to the Arduino.

The Canon software is simply written into a text file named PictUAV2.lua and put into the “SCRIPTS” subdirectory of the SD card in your Canon, running the CHDK software.

I will admit that I’m far from adept at using the CHDK software and can’t offer you much in the way of help there. There’s a bit of a learning curve. But it’s real powerful stuff. And if I can learn enough to get it working, so can you. The CHDK software is here , be sure and get the version for your specific camera AND firmware. If you’re not familiar with CHDK, you’ll be doing some reading….

• USBController.ino for the Arduino
• PictUAV2.lua for the Canon camera

Setting up the Canon camera

Once you have your Arduino programmed and the PictUAV2.lua placed in the SCRIPTS directory of the camera’s SD card (AND the write protect on the card enabled….CHDK is funny that way!) you’ll need to change a few settings in the camera’s CHDK setup. These worked for me for both the SX230 HS and A2200 but it’s possible other models could be just a little different.

With the camera off, press the “Playback” button. You should see a CHDK splash screen appear but the lens will not extend. Press the “Func Set” button, and you’ll get a full screen CHDK menu. Select “Load script from file” and then select PictUAV2.lua. Then, back at the menu, set “Autostart” to “On”. Next, select “Remote Parameters”. You want to set “Enable remote” to “On” (This activates the USB port to act as a remote). Now select “Back” to go up a level to the first menu. The last thing you need to do is make sure the “Save params” option is checked. This ensures the changes you just made will be retained the next time the camera powers up.

Having done the camera setup at the bottom of the menu, below the line labeled PictUAV2, are the parameters you can set for the actual program’s operation. Most important is “Secs/frame” which sets the interval at which you want images taken. Note that some camera models have a minimum interval at which they can run. For some reason my SX230 can run at a shorter interval (2 seconds) than my A2200 (3 seconds). Probably due to the SX230 being a better camera.

To be honest I haven’t really played much with the other options but it appears you can set the length of time you want the intervalometer to run, whether or not to focus for every shot (for aerials a single focus will save time) and if you want to turn the display off (which saves battery life). More choices than you know what to do with.

Operating the system

The RC receiver should be on and providing power to the Arduino. I suppose it’s wise to press the reset on the Arduino to make sure it’s booted up properly, but mine has always been fine with just the power up. You mileage may vary.

Activate the Canon by pressing the “Playback” button. The choice of the button can vary for different Canon models under CHDK, so check the CHDK Wiki for whatever model you’re using.

If everything is working as it should (Hey, it happens occasionally) after you see the CHDK splash screen there will be a box in the lower left showing “Autostart”, “PictUAV2 Started”, “PictUAV2 loop” and a = 0. The lens should still be retracted. If the USB cable is already connected you’ll see the a value to be dancing around between 0 and 1. This is it reading the 15 millisecond standby pulses. If you’re seeing this, all may well be golden.

Now throw your selected switch on your RC transmitter to Record. The lens should now extend. You’ll see a notation pop up on the camera’s display that recording has started and the a value will increase. And, if all is well, the camera should be taking pictures at the interval you’ve selected.

Now, to shut the beast off, move your transmitter’s switch to Shutdown. You should see the a value increase to 90, the lens retract and everything go off. Did it do it? If so, buy a lottery ticket. You’re on a roll. Note that once it’s off you can’t turn it back on until it’s on the ground and you manually press the camera’s “playback” button.

Oh, one last thing….Since the script is now set to autostart, suppose you need to break it to do other things with the camera. If you turn it on now it just starts taking pictures like a maniac. If so just press the shutter. You’ll see “Interrupted” appear on the display screen and the incessant picture taking will cease. Or it should. If not smack it against a wall.

Anyway, that’s it. It’s a nifty way to loft a camera and keep it reasonably protected. I mean other than it being flown around hundreds of feet over the ground.

Copy everything below the line and using a text editor paste it into a file titled, “USBController.ino”

Back to the Remotely actuated intervalometer for aerial mapping page

———————————————————————————————–

//  Canon Arduino Controller

//  Arduino pin 10 is input from RC receiver
//  Arduino pin 12 is output to camera USB
//  Arduino pin XX is +5 volts from RC receiver
//  Arduino pin XX is ground
//  Set up initial variables:
int led = 13;
int PWM_In = 10;
int USB_Power_Out = 12;
unsigned long duration;
int ActiveState;
int STANDBY = 0;
int RECORDING = 1;
int SHUTDOWN = 2;

//  The following sets the PWM levels from the RC receiver:
//  Low PWM=standby, medium PWM=record, high PWM=shutdown
int StartState = 1250;
int RecordStateStart = 1350;
int RecordStateEnd = 1650;
int ShutdownStart = 1750;
// End set up initial variables

// Setup routine runs once when you press Arduino reset button:
void setup() 
    {                
    // Initializes pin 12 and pin 13 (LED) as digital outputs
    pinMode(led, OUTPUT);     
    pinMode(USB_Power_Out, OUTPUT); 
    //  Initializes pin 10 as digital input  
    pinMode(PWM_In, INPUT);
    Serial.begin(9600);
    digitalWrite(USB_Power_Out, LOW);
    }

//  This loop routine runs until shutdown:
void loop() {
   duration = pulseIn(PWM_In, HIGH);

//  Check whether input signal is for standby, record or shutdown:   
   if (duration <= StartState)
     {ActiveState = STANDBY;}  
   else if (duration >= RecordStateStart && duration <=  RecordStateEnd)
     {ActiveState = RECORDING;}
   else if (duration >= ShutdownStart)
   {ActiveState = SHUTDOWN;}

//  If Standby mode, output 15 msec pulses:   
     if (ActiveState == STANDBY)
        {
         digitalWrite(USB_Power_Out, HIGH);
         digitalWrite(led, HIGH);
         delay(15);
         digitalWrite(USB_Power_Out, LOW);
         digitalWrite(led, LOW);
        }

// If Record mode, output 45 msec pulses:   
     if (ActiveState == RECORDING)
       {
        digitalWrite(USB_Power_Out, HIGH);
        digitalWrite(led, HIGH);
        delay(45);
        digitalWrite(USB_Power_Out, LOW);
        digitalWrite(led, LOW);
       }

//  If Shutdown mode, output 90 msec pulses:
      if (ActiveState == SHUTDOWN)
       {
        digitalWrite(USB_Power_Out, HIGH);
        digitalWrite(led, HIGH);
        delay(90);
        digitalWrite(USB_Power_Out, LOW);
        digitalWrite(led, LOW);
       }

  delay(200); 
} 

Copy everything below the line and using a text editor paste it into a file titled, “PictUAV2.lua”

Or….I seem to be having difficulties posting this block of code while still maintaining the proper indentation and line spacing which makes it sort of crap to read. And it’s pissing me off! So if you want to download the file with the proper indentations, click here and do a “File save as”. Sorry about that…I’m just dense.

Back to the Remotely actuated intervalometer for aerial mapping page

____________________________________________________________________________

–[[
@title PictUAV2
@param s Secs/frame
@default s 2
@param h Sequence hours
@default h 1
@param m Sequence minutes
@default m 5
@param e Endless? 0=No 1=Yes
@default e 1
@param f Focus: 0=Every 1=Start
@default f 1
@param d Display off frame 0=never
@default d 0
--]]
print( “PictUAV2 Started ” )
– convert parameters into readable variable names
secs_frame, hours, minutes, endless, focus_at_start, display_off_frame = s, h, m, (e > 0), (f > 0), d

props = require “propcase”

– derive actual running parameters from the more human-friendly input
– parameters
function calculate_parameters (seconds_per_frame, hours, minutes, start_ticks)
local ticks_per_frame = 1000 * secs_frame — ticks per frame
local total_frames = (hours * 3600 + minutes * 60) / secs_frame — total frames
local end_ticks = start_ticks + total_frames * ticks_per_frame — ticks at end of sequence
return ticks_per_frame, total_frames, end_ticks
end

function print_status (frame, total_frames, ticks_per_frame, end_ticks, endless)
local free = get_jpg_count()
if endless then
local h, m, s = ticks_to_hms(frame * ticks_per_frame)
print(“#” .. frame .. “, ” .. h .. “h ” .. m .. “m ” .. s .. “s”)
else
local h, m, s = ticks_to_hms(end_ticks – get_tick_count())
print(frame .. “/” .. total_frames .. “, ” .. h .. “h” .. m .. “m” .. s .. “s/” .. free .. ” left”)
end
end

function ticks_to_hms (ticks)
local secs = (ticks + 500) / 1000 — round to nearest seconds
local s = secs % 60
secs = secs / 60
local m = secs % 60
local h = secs / 60
return h, m, s
end

– sleep, but using wait_click(); return true if a key was pressed, else false
function next_frame_sleep (frame, start_ticks, ticks_per_frame)
– this calculates the number of ticks between now and the time of
– the next frame
local sleep_time = (start_ticks + frame * ticks_per_frame) – get_tick_count()
if sleep_time < 1 then
sleep_time = 1
end
wait_click(sleep_time)
return not is_key(“no_key”)
end

– delay for the appropriate amount of time, but respond to
– the display key (allows turning off display to save power)
– return true if we should exit, else false
function frame_delay (frame, start_ticks, ticks_per_frame)
– this returns true while a key has been pressed, and false if
– none
while next_frame_sleep (frame, start_ticks, ticks_per_frame) do
– honour the display button
if is_key(“print”) then
click(“print”)
end
– if set key is pressed, indicate that we should stop
if is_key(“set”) then
return true
end
end
return false
end

– click “print” to turn on/off display
function seek_display_mode()
click “print”
end

– switch to autofocus mode, pre-focus, then go to manual focus mode
function pre_focus()
local focused = false
local try = 1
while not focused and try <= 5 do
print(“Pre-focus attempt ” .. try)
press(“shoot_half”)
sleep(2000)
if get_prop(18) > 0 then
focused = true
set_aflock(1)
end
release(“shoot_half”)
sleep(500)
try = try + 1
end
return focused
end
– This is the program core, a loop:
Started = 0
print( “PictUAV2 loop ” )
a = -1;
repeat
aold = a
a = get_usb_power()
if (a ~= aold) then
print(“a = ” .. a)
end

– If pulses are 90 msec or greater, shutdown directly and end.
if (a>8) then
print( “shutting down ” )
shut_down()
sleep(1500)
end

– If pulses are greater than 30 msec start recording
if (a>3) or (Started == 1) then
if Started == 0 then
print( “recording Started ” )

sleep(1500)
Started = 1
if focus_at_start then
if not pre_focus() then
print “Unable to reach pre-focus”
end
end

start_ticks = get_tick_count()

ticks_per_frame, total_frames, end_ticks = calculate_parameters(secs_frame, hours, minutes, start_ticks)

frame = 1

print “Press SET to exit”
end
print_status(frame, total_frames, ticks_per_frame, end_ticks, endless)
shoot()

if frame_delay(frame, start_ticks, ticks_per_frame) then
print “User quit”
break
end
frame = frame + 1

end
until ( false )
print( “PictUAV2 ended ” )

I had been doing more studying on using drones for mapping/photography purposes and I was eager to try it again after my less than satisfactory attempt at the Pejunkwa ruins. I had managed to decrease the interval rate on my Canon A2200 from 5 seconds per image to 3 seconds giving me more images per flight. BTW, one of the reasons I really like this camera (other than it’s cheap and light at 133 grams!) is that it doesn’t have image stabilization. This might seem counterintuitive, as image stabilization usually works rather well….if you’re hand holding the camera. But in many cases the vibrations present on a multirotor can totally swamp a camera’s stabilization abilities and result in blurred images. If the camera’s stabilization is turned off, or it doesn’t have it, shooting above 1,000th of a second usually results in a good image. But I digress….

One of the places I have passed many times while searching for Bill Ewasko is (are?) Samuelson’s Rocks in Joshua Tree National Park. It’s something of a landmark about two miles in from the paved road consisting of a low, rocky hill, maybe 50′ high running about 550′ north-south and about 330′ east-west. It’s primarily known for the eight inscriptions carved into the rocks around the site by a (slightly crazy) guy named Samuelson in the late 1920s, through the early 1930s. Anyway, this struck me as a good target to see what sort of aerials I could create and anything more.

On a previous trip, looking southeasterly towards Samuelson’s Rocks from the adjacent hills. They are the rockpile in the middle of the image just right of the yucca plant. 

This is a close up of Samuelson’s Rocks looking toward the southeast from some nearby hills.

So (very) early one morning found me lugging my quad the two miles in to Samuelson’s. I brought batteries for two flights in case something didn’t go quite right on one of them. From earlier experiments I determined a 250′ altitude should give me good resolution so I launched and put the flight controller on altitude hold when it reached that altitude. My flight controller, an APM 2.5, uses a very sensitive barometer to sense height and typically holds it to within a meter. There was a very light breeze but learning from past bitter experience, I choose a day when conditions were generally calm. I was able to make the two 8 minute flights without incident, obtaining 153 images on the first flight and 152 images on the second flight. I was there maybe 30 minutes max. Left nothing but footprints, took nothing but pictures.

Back home I went through the images and tossed those which were blurred or off-angle (not looking straight down). I then fed those remaining 124 images into Autopano Giga and let it do its thing. This excellent software doesn’t care how the images are rotated, it manages to find matches, slices and dices them in magical ways I don’t understand and spits out a huge panorama. What’s “huge” you ask? Well, it ended up as a 600 meg jpg file. In my world, that’s huge. Looked pretty damn good, too.

Here’s a version I cropped and resampled down to 2.9 megs. It lost some resolution in the further compression but you get the idea. North is at the top and there’s the remains of an old corral plainly visible on the left.

This is a cropped composite of 124 individual images. North is at the top.

So just how good is the resolution in the original? Well, since you asked (you didn’t but I’m telling you anyway), here’s an example. There is a rusty old bedframe up in Samuelson’s Rocks, presumably Mr. Samuelson’s. It tends to get moved around by hikers but it’s always up there somewhere (There’s a picture of it at the bottom of this webpage). If you’ll look at the partial image below I circled the location of the bedframe the day I took the images. If you go back to the big image and look in that same spot you’ll see the bedframe. But I also went through the original full-resolution single images and found one that happened to be centered on the bedframe, which I’ve cropped out and also added below. The maximum width of the bedframe frame is only about 1 1/2 inches. Yet this dimension is plainly visibly. That’s pretty neat!

The location of the mentioned bedframe is in the center of the red circle. 

This is a crop of one of the original 14 megapixel images showing the bedframe at full resolution. The members are about 1 1/2 inches wide.

Had I stopped there I would have been very satisfied. I am well on my way to developing the ability to create high resolution aerial photos of whatever I like. But then in my readings I came across something called “Structure From Motion” (SFM). This is a technique, only a few years old, which can create a 3D model of an object from a series of individual photographs taken by a camera with known settings. In essence, it uses serious computer horsepower to look at craploads of images, finds common markers in the images, and looks at how these markers change with respect to each other as the camera viewpoint changes and creates a 3D model from this. It turns a camera into a poor man’s laser scanner. I don’t pretend to understand how it works, but I suspect it was reverse engineered from alien Roswell technology.

A few things make this a fantastically attractive proposition. The first is there’s free open source software to do this, VisualSFM. The second is that it doesn’t require high precision (expensive) photogrammetric cameras. Cheap point and shoots like my Canon A2200 will work quite fine. And finally, while it was originally conceived to make 3D representations of objects, it will also work well when used on terrain features. This.Could.Be.Interesting!

So I selected the 124 Samuelson images and dumped them into VisualSFM without really knowing what I was doing (Read the manual??! Why??). About 30 minutes of computer chomping later, out pops a pretty cool 3D surface. Now just doing a screen grab of the surface really wouldn’t give a good representation of its scrumptious 3Dness, so below is a short video clip of the Samuelson’s Rocks model being rotated around. I think the holes in the surface are due to areas of shadow. I deliberately took early morning shots when the Sun was low to enhance features and bring out the terrain in the images. Turns out this is not what you want when doing a SFM model. High Noon with full lighting is better. Go figure.

So considering this was done with a cheapo Canon with no ground control whatsoever, I think it’s a pretty impressive result. If I add the known scaling of a few points in the model I can then create a scaled 3D model and wrap the aerial photo over the surface. But I’m still figuring that out.

Seeing as how I have been proving to be very adept at causing damage to my assorted dronies, I’ve been a bit reluctant to let them loose by themselves. The APM flight controllers have the ability to fly a course completely on their own(even land and take off). Setting up the course is as easy as simply clicking the points on a Google Map you want it to fly to, tell it what altitude at each point (and a few other things if you like) and then load it into the APM flight controller. To initiate the program, just throw a switch on your RC transmitter and off it goes. It might even come back, if you’re not one of the unlucky ones. So I figured it was time to take the leap with my quadcopter. Besides, if it augured in I could plausibly argue it wasn’t my fault as I wasn’t flying it.

Whenever I’m testing something sketchy I have a favorite spot on the outskirts of Moreno Valley. There’s nothing but miles of open fields and not a thing to run into but the ground or an occasional meth lab. This works well for a person of my abilities.

I loaded in a flight plan about a mile in total length with 10 waypoints the quad needed to hit. It was roughly rectangular with a lot of right angle turns. This allowed me to fly a decent number of waypoints but kept the quad close to me in case it did something wonky and I needed to take over control. Not that that ever happens. I instructed it to fly to the first three waypoints at 90′ then descend to 70′ for the rest of the waypoints, then return to its launch point and hover at 15′. I brought along three batteries so I’d have three flights.

I lifted off the first flight manually and held the quad at around 15′ getting it stable, then threw the switch. Off it headed to the south as it was supposed to, climbing nicely. I had set up my ground station to watch the quad’s video feed with the on-screen data display as it did its thing. Normally I use video goggles but I wanted to keep a physical eye on the sucker….just in case.

As it headed for the first waypoint I saw it climb past 90′ and kept right on going up. That’s not what I wanted to be seeing. I had my finger on the switch ready to take over control from this demented mechanism. But then it hit the waypoint, although at a too high 125′, and made a sharp right turn, just as it should. But it was still friggin’ climbing. I decided to let it play out and see what it was up in its beady silicon brain to as long as it followed the waypoints. The flight controller has a very sensitive barometer in it and normally holds an altitude within just a few feet, so this was very strange.

It hit waypoint 2 and made the required right turn, but it was now at 211′ and still going up. Then halfway to the third waypoint it finally ceased its climb and the altitude stabilized at 293′. WTF?

So as I’m looking back and forth between the data display screen and the quad in the distance, the mathematical part of my brain starts whispering to me, “Hey moron, do you think it’s a coincidence that it’s flying just over three times higher than you programmed it to??”. Wait….three times….Oh shit….It wasn’t 90 feet, it was 90 meters! And that, boys and girls, is how you crash a two hundred million dollar space probe into Mars. In my defense, the software which is used to set the waypoint altitudes does not specify whether it’s feet or meters. It’s a position I’ll cling to. So it was doing exactly what it was supposed to, fly at 90 meters. It didn’t do it prior to the first waypoint since there’s a maximum climb rate it will use when flying autonomously and couldn’t make it to its designated height before that waypoint. But it sorely tried.

During my metric enlightenment the quad hit waypoint 3, turned right and started for number 4. As it did so, it began descending. At this point I’m now ahead of it and am guessing it’s going to stabilize around 220′, which would be approximately, ahem, 70 meters. That it did, flattening out at 228′, and continued to hold this height rock-solid through the remaining waypoints.

On its last leg it returned to a spot immediately above its launch and started descending from 228′. These vertical descents are surprisingly rough since the quad is descending straight through its wake. If you’re flying manually it’s better to have a bit of forward motion. I manually did a very rapid descent once from high up with no forward speed and the quad suddenly rotated 90 degrees on its side without warning. I was able to punch the throttle and recover but it was sphincter-tightening.

The quad finished its descent and went into an automatic loitering mode at 15′. This is a mode where it just hovers and holds position. It is capable of autolanding but I just don’t trust the little bugger that far yet. I took over control at that point and ham-handedly brought it down to the ground. The next two flights were virtually identical. I find it especially annoying that it flew itself far better than I ever could.

When I got home I had a chance to go over all the considerable recorded data in detail. Comparing the actual flight paths to the planned path showed something strange going on. All the legs between waypoints had weird curves in them. All consistent but all curved. Turns out this appears likely to be a simple compass calibration error, with the compass being off about 25 degrees. Here’s what it looked like:

Actual quadcopter flight paths (blue, green, yellow) as compared to planned path (red). The flight route was generally clockwise along the waypoints numbered in white. North is to the top in this image.

The flight controller has two ways to figure direction, a built in magnetic compass and also with its GPS. But using the GPS for direction only works if it’s in motion. When the flight controller isn’t in motion, like hovering at a waypoint, it uses the magnetic compass. So when the quad started for the next waypoint it headed off around 25 degrees to the right of where is should have gone. Then, as it started to generate a direction from the GPS readings it realized the error and started correcting back to the right point. So the compass and GPS were fighting it out. Now that I know this it’s pretty simple to calibrate the compass. Actually I thought I already had.

So this all was a great success, my metric dissonance notwithstanding. The reason this will be so useful is that the software also has the ability to do automatic flight path creation. You just click on an area you’re interested in on the Google satellite photo and make a polygon around it. Then you tell it how high you want to fly (in stinkin’ meters!) and what sort of percentage overlap you want. Then click “Generate” and it automatically generates a flight path that will cover everything you’ve selected (and do so way better that I ever could). Way too cool.

Here’s a video from one of the flights taken from the HD camera on the front of the quad. It would have looked much better if it were in fact at 90′ and 70′ up:

There you have it, cheating death one flight at a time…..

Date: 6-24-2013

Participants: Tom Mahood

General Search Area: This was a hike on the third anniversary of Bill’s hike with no specific search areas intended. Primarily reflection and reconnaissance, following Bill’s presumed route to Quail Mountain.

I really wasn’t enthusiastic about doing this hike.

I had done the hike in to the base of Quail Mountain on the first anniversary of Bill’s hike, presumably along the route he took. It was hot, probably about the same temperatures he experienced. And I don’t like heat if I can at all avoid it. My ancestry is that of cloudy, cold places and I like it that way. But still, it was a good time to reflect and consider new ideas.

Last year, on the second anniversary, I was out of the state so no hike was possible then. Which brings us to this year. And I really didn’t want to do it.

Why? Well, there’s that heat thing. Now given a good idea of a place to search I’m OK with a summer hike, if that’s what it takes. But there’s that “good idea” part. I feel it’s been a while since I had any ideas I thought anything of since we crashed and burned so badly with my southeast mouth of Smith Water Canyon idea. And this sort of hike would just serve to remind me that I’ve still failed with each step. It gets old after a while.

So I decided to make this sort of a tourist trip, heading in to Juniper Flat, along Bill’s presumed route, and maybe have a look at where a bandanna was found during the initial search. I had the coordinates and knew it was at the base of Quail Mountain but I had never been there.

The bandanna find was interesting since Bill, who was balding, often hiked with them. I was told he sometimes would drop a bandanna during a hike to mark an important junction for the return. However other than a brief mention of the bandanna find in the original search records, there was no further info on it. Nothing about color, condition, nada. It could have been Bill’s or it might not. But the spot was worth a look.

My initial wacky idea was to start off on the hike around a similar time to what Bill might have, after 10:30 AM. But then I realized I’d run into all sorts of traffic getting out there and I hate traffic worse than heat. So I pivoted to the other side of the nasty traffic time and ended up at the Juniper Flats trailhead at 8 AM.

Getting out of the car was…well, a surprise. It was windy and almost felt like I needed a jacket. The forecast was windy and in the upper 80s but the morning was a little crisp. This is late June? I decided it would have to warm up and started down the closed dirt road that heads to Juniper Flat. I just zoned out and 1:35 later I found myself in the middle of Juniper Flat. Well THAT was certainly quick!

I shot a panorama and took a critical look at the terrain. Looking towards Quail Mountain there’s an obvious route right up its face. This is a sucker route. I know from experience it’s not what it appears. There’s a big canyon between the ridgeline and the final slope of Quail rendering that direct route not possible. Or at least really stupid. The smart route is to continue easterly across Juniper Flat and climb the ridge on its east edge. This ridge does a nice gradual ascent to Quail Ridge, just easterly of Quail Mountain. But if you wanted a no-nonsense route up Quail, consistent with what guidebooks describe as the “direct route”, and weren’t really familiar with the area, well you’d take the sucker route. Since this was pointed in the direction my GPS was giving me for the bandanna coordinates, off towards Quail I headed.

Right away I noticed there was a very pronounced use trail headed in my direction so I hopped on it and followed it. I figured it would soon veer off to the right and take a more intelligent approach to Quail. But no, it was headed for the sucker route.

Looking toward the area the bandanna was found from Juniper Flat. The route up to the main ridge is along the ridge spine above and to the left of the bandanna location. You can’t go directly from the bandanna location up and right to Quail Mountain because there’s a large canyon in between that’s not visible.

As I approached the toe of the ridge spine I realized Patrick McCurdy and I had been just a few hundred feet west of my current path during JT16. I further realized that the bandanna location wasn’t on the flat part of Quail Flat, but was actually up the ridge spine just in front of me. A short climb later and I was at the bandanna site.

If anything its location makes it even more ambiguous. It appears to have been found smack on one of the well used routes up Quail, so anyone could have dropped it. Further, it’s not a spot someone would mark as a decision point on the return. So if it was Bill’s, it was dropped accidently. So the bandanna remains well in limbo as a clue. Maybe. Maybe not.

So as I’m looking around the bandanna spot I’m starting to realize a few things. First is that this really didn’t take me much time. Probably some of that was the cooler than usual temperatures but the rest of the way up to Quail didn’t seem that far (I know, another optical illusion). I was previously a bit skeptical that Bill would have proceeded up Quail what with the heat and the later start time. But if he was a “heat person” he probably would have been feeling about what I was feeling then. Also in previous trips up the Quail ridge I had been doing searching of other areas along the way so I was pretty beat when I arrived there. I think it colored my judgment in this regard. This little tourist trip was making me rethink the viability of ascending Quail in June. It wasn’t nearly as bad as I had thought.

I also had the sense that if Bill did head up to Quail Mountain, it was highly likely I was on the same route. It’s what felt right based upon the terrain. In fact, once you start up this route there’s only one way to go. And with the presence of fresh footprints and an occasional rock cairn, it’s clear this was a well used route. If I was planning a trip to the top of Quail Mountain for myself, I’d have been over on that ridgeline much further east. But if I wasn’t real familiar with the terrain, I’d be right where I was.

Checking on my Verizon cell phone, I had a decent signal on the ridge spine, probably coming up out of the Coachella Valley. And there continued a good signal all the way up. So if Bill did take this route he clearly had his cell phone off. This is consistent with the lack of any reported pings after he completed his morning calls while he was enroute to Joshua Tree.

Looking out onto Juniper Flat from the bandanna location in the foreground. That’s the Coachella Valley off in the distance and Verizon cell coverage is starting from here on up.

So up I went, taking the only reasonable route, along the ridge spine. Shortly after passing the bandanna site the ridge spine gets high enough to see that big canyon I mentioned between where I was and Quail Mountain. I can imagine many a climber going, “Oh crap!” at this point. It becomes clear the route actually isn’t straight to Quail Mountain but rather curves away from Quail Mountain, heading westerly as the spine rises to Quail Ridge. It eventually gets up there, but about a half mile westerly of Quail Mountain.

I finally reached the local high point of Quail Ridge and checked my time. It was only 2:45 since I had left the trailhead and I was looking right at Quail Mountain only 0.4 miles to the east, a trivial walk. And on top of that I had stopped several times to look around, think and take pictures. I was beginning to think going up Quail in summer wasn’t as farfetched as I had originally thought.

A panorama on the Quail Mountain ridge with landmarks annotated.

I decided I had already been to Quail Mountain itself enough so I didn’t bother heading over there. Instead I found a comfy rock, sat down and thought about things while taking in the vista. I couldn’t help noticing that to the northwest, seemingly very close, was what I knew to be the southerly rim of Smith Water Canyon. I looked over at Quail Mountain, then looked out to the Smith Water rim and it really didn’t look that far away. I later measured it on the map at 1.5 miles, still not far. Looking back down at Juniper Flats, from where I had just come up, reinforced the feeling that a cross country hike to Smith Water Canyon wouldn’t be too bad. And this is where I sort of scared myself. I started actually thinking about doing it.

Looking toward Smith Water Canyon from Quail Mountain ridge. The ridge in the distance is the southerly rim of Smith Water Canyon. The camera is slightly zoomed to approximate what the view is to a typical human eyeball. This is pretty much what it looks like in real life.

Sure it would be a long hike, verging on epic. Head to Smith Water, drop into the canyon, head westerly in the canyon until meeting up with the California Riding and Hike Trail which runs back to the Juniper Flats trailhead. What I was seeing in front of me didn’t look too bad. The first part, descending off the ridge I was on was wonderfully open and gentle. It was calling my name!

Fortunately I got my binoculars out and scanned the cross country route in the distance. Immediately I saw all the boulder fields I had forgotten about and was reminded of how bad that area really is. But it wasn’t discernible to my naked eye. Also I had started the hike with 5 liters of water and I probably had used half. I figured I’d have enough water to make it to the bottom of Smith Water but just enough. And the idea of partaking of scummy water wasn’t too appealing.

A zoomed panorama looking toward Smith Water from the Quail Mountain ridge. The level of rockiness approaching the crest becomes apparent in this zoom. It’s not visible when viewed normally from this distance.

Taking a mental step back from the edge I started thinking about why this at first seemed a not unreasonable idea to me in late June. I mean, it sounds pretty nuts, doesn’t it? Well, I think mainly it was the wind.

Quail Mountain and its ridge are the tallest things for many miles around. If I was ever up there and it wasn’t windy I can’t remember it. Some places suck, Quail Mountain blows. And this wind, combined with the low humidity, makes it seem comfortable and much cooler than it actually is. However once descending out of that nice wind, possibly heading northwest, you find yourself in a furnace of still, hot air. And oh, by the way, you’ve now entered a huge cell phone shadow area. No coverage until you get near the rim of Smith Water Canyon, which happens to be around 11 miles from the Serin Drive cell tower. An unlikely coincidence in my opinion.

So as I was sitting there, considering my own foolish thoughts, I started thinking about how this could have gone with Bill.

First, I think he either ascended Quail via the route I just completed or he went off someplace so strange no one has yet an inkling of where it is. So let’s assume he went up Quail. If he was a strong hiker, used to heat, he could have been where I was in even 2 1/2 hours. But almost certainly he would have burned through most of the three bottles of water he is believed to have been carrying.

We know Bill knew about Smith Water Canyon, that it had year round water and it connected with the California Riding and Hike trail at its westerly end. Even if he scaled it off on a map, he would have found he was only 2 miles from the water. Looking out to the northwest, he could see most of the route before him and it looked better than what he had just come up. So why not try it? No need to make any phone calls as he wasn’t in any trouble. Just taking the long, scenic route back (’cause Smith Water is cool!) and picking up some water along the water. Sounds like an adventure! So off he heads to the northwest.

Now I’m told Bill was an old school map and compass kind of guy. So he would have either taken a bearing or sighted on a prominent landmark. Maybe a landmark such as the conical peak at the easterly portion of Smith Water Canyon. In any case I think he would have tried to hold to a straight line, something made difficult by the terrain.

Viewing the terrain from the saddle on the ridge this is the general route I’d expect someone not fully aware of the terrain to choose. Black lines are original search tracks, reds are tracks since then and the cyan radius line in the upper left is the 11.1 mile radius from the Serin Drive cell tower. The red shaded area corresponds roughly to the zoomed panorama looking toward Smith Water.

At some point my hunch is he experienced some injury, perhaps as minor as a twisted ankle. Very easy to do in that terrain, especially if you’re trying to hold a bearing rather than follow the contours of the land. And by then the still dry air would have caused him to finish his water supply. Finally, no cell service is to be had. A serious survival situation has now been created.

So what are the options? Well, he certainly could have returned the way he had come. But if whatever befell him occurred when he was well towards Smith Water the prospect of going back up and over the Quail Mountain ridge and making it all the way back to the car would have been VERY daunting. And no water the entire route.

If he turned right and started down the drainages to the Samuelson’s Rocks area and to the even further Park Road, that was many miles. And again, no water. Water equals survival so that’s the priority. Add to that it was completely unknown territory.

So the only option that’s left is to press on to Smith Water Canyon and water. Sure it’s a risk, but it is, as they say, your only chance. This resonates with me because frankly, it’s what I would have done. I’m not saying it’s the best thing to do, but I would have taken that gamble (although I might have eventually tried getting to the highest spot I could find to use my cell phone). The nightmarish aspect of this option is that the closer you get, the worse the terrain gets. But you don’t find this out until it’s too late to do anything else.

Clearly, if this is the scenario, Bill didn’t make it. But it would explain the strange cell ping on Sunday morning. The distance of 10.6 miles puts it either on the edge of Smith Water Canyon or just in it.

I and others have spent considerable effort exploring the southern slopes of Smith Water Canyon based upon this premise, and so far, nothing. I’ve found that extremely discouraging as the obvious locations have been cleared. But perhaps locations we felt were unlikely when viewed from a perspective of approaching from the north, west or east turned out not to be so if approached from the south?

So what did I take away from this trip? I think Bill leaving the Quail Mountain ridge, heading northwesterly towards Smith Water Canyon and not making it is still the most viable scenario. Granted that brilliant idea is hindered a bit by the fact we haven’t found him there. But perhaps he’s in an area we have yet to search. And that amount of area is getting smaller and smaller. Certainly he’s someplace that’s not visited, otherwise he would have been found by now by a cross country hiker.

My sense (my fear?) is that when Bill is eventually found it’s all going to seem so clear in hindsight. And I’m going to whack myself on the head and ask, “Duh, why didn’t I think of that??!”

After my ridgeline meditations I dropped back to Juniper Flat via a drainage that had never been searched and blasted out of there. I was back at my vehicle by 1:30 PM for a total trip time of 5:30, which included a lot of sitting around. So Bill could have conceivably left the trailhead after 10:30 AM and still been out of the park by 5, his original plan.

I look forward to Fall…..

GPS mileage for this trip: 12.8 miles

Cumulative GPS mileage to date: 706.9 miles (breaking 700 is a depressing milestone!)

GPS tracks for this trip in Google Earth kml format

GPS tracks for this trip in Garmin gdb format

GPS tracks for this trip in gpx format

The overall route of JT53

A closeup view of the area covered by JT53 in blue. Original search tracks in black, tracks since then in red.

This is still a work very much in progress. Been crashing it far more than I should, so I don’t know what’s up with that.

The Bixler 1.1. This has ailerons but no flaps.

The small black box at the left is the FPV camera (I know, it’s tiny, but size doesn’t matter). The center stuff is the Hobbyking KK2.0 board running the OpenAero firmware. Above it and to the right is the GPS for the onscreen display. The jury is still out on this setup. An astute observer might notice the amount of dings on the nose….This is NOT a good sign.

Another view of the front of the Bixler. Yeah, that nose looks pretty beat up.

The mid parts of the aircraft. To the left, under the motor is the ESC. The thing on the boom is an ImmersionRC 600 mW 5.8 GHz video transmitter, one of my favorites. The antenna is beneath the boom.

And now for something completely different.

In the early 1970s I was obsessed with finding an old mine in the San Gabriel Mountains of California called the Stanley-Miller Mine. This was probably my first big quest and perhaps what started my ongoing affliction of hunting for weird shit

I had studied historic USGS maps of the area which showed several old trails going to the Stanley-Miller. One of these trails traversed the easterly slopes of the East Fork of the San Gabriel River Canyon, and well…It needed checking out. Along the way it passed by a mine called the Allison Mine. Didn’t know much about it as this was the days of libraries, not the Internet.

Being young, stupid and apparently insensitive to trail miles or altitude gain this seemed like a potentially fun outing. I talked a friend of mine, Phil Budig into going along as he was equally stupid.

So sometime in the winter of 1973 Phil and I hiked up out of Heaton Flat to Heaton Saddle, which looks down into Coldwater Canyon. From there we spotted an old trail which essentially stayed level and went westerly then easterly around a ridge spine into the canyon the Allison Mine was in. The trail was pretty overgrown but mostly decent. Eventually we ended up at the Allison Mine.

OMG was there stuff here. Not just stuff, but STUFF! There was a covered workshop with no walls, which still had tools hanging on the supporting posts. Behind it was a cabin with all sorts of interesting things in it. The cabinets hanging inside on the walls of this cabin were covered with graffiti from the 1940s and 50s. There were phone numbers written with four digits. I kick myself today for not getting pictures of this. It appeared most of the operations were from the 1930s.

Nearby was a second cabin with more artifacts. Adjacent to it was a large one cylinder gas engine, connected to a generator by a belt that had rotted away. In fact, on the property there were four to six of these one cylinder motors of various sizes. There were ore car tracks going into the mine, a trestle and ore dump. Below it in the canyon was a ball mill for grinding the gold ore.

We quickly lost interest in looking for the trail that continued on to the Stanley-Miller and spent the rest of the afternoon exploring the Allison. It was really something.

What was very unusual for this trip was that I had brought along a Super 8 movie camera. It was a really shitty camera which I had purchased for reasons I have long forgot. I also forgot about the film until only a few years ago when I found it in a box. Not quite sure what was on it, I sent it off to be converted to a DVD and was amazed to see the Allison Mine of so many years ago. Of course a shitty camera equals a shitty film, but at least it shows what the Allison was like. Why is this important? Ahem…

Probably around 1976, about three years after my first Allison Mine visit with Phil, I made a solo trip back to the location. This was my second visit to the site. I forget exactly why, as I had already found the Stanley-Miller (Ironically I found it on the day I ran into a ranger that morning who said he doubted it existed any more, but that’s another story and so worth all the hunting!). I went the same route in and eventually saw the ball mill below the trail. Continuing on a little further I reached the stream that I knew ran next to the cabins, but there’s nothing there, just a clearing. Hmmm…Perhaps I’m mis-remembering and the cabins were a little up the stream. So up I go. Still nothing. At this point I’m really confused and a little disorientated. It all SEEMS like it’s where the cabins were, but clearly they’re not so I must be mistaken. But some of the mining stuff was still there and I knew that it had been next to the cabins……Then I realized….The cabins were gone.

I don’t mean gone like in people taking souvenirs over time. They were gone and their locations rehabilitated to blend into the existing vegetation. All that fantastic stuff…vanished. Tools, motors, pots pans, foundations….gone.

Now the public generally sucks and there are always some assholes who will do a number on interesting things in the middle of nowhere. That’s one of the reason I get cagy on this website when I write about unusual places. But to make a site completely vanish and look like it was never there? Well that takes an outfit with substantial resources, an outfit like, oh I don’t know,…..the US Forest Service. At the time I still had the naive notion the Forest Service were the good guys in the funny hats who protected our public lands from the assholes. As I was standing there where the Allison cabins used to be my world view took a major shift and I realized there were more assholes out there than I had even considered.

Now I have no idea why the Forest Service would do such a thing, but since then I have heard of other cabins they’ve removed. Usually the concern was over squatters or something like that. It was deep within the Angelus National Forest. In those pre-Internet days it was hard to get answers and it usually involved a lot of time with telephone books. I had more interesting things to do. But I never forgot it.

So below I’ve linked to the YouTube video of the Super 8 movie I took during the hike. Phil and I were young, stupid and generally idiots so much of the film is cringeworthy. But it shows at least in a grainy way what the Allison Mine once was. These days the Allison Mine is much better known and people visit the site via a couple routes (Google it). And none of this folks know what they are missing. But if they stumble across this video, maybe they will. And should any of them later come across a fantastic hidden cabin somewhere in the woods, maybe, just maybe, a remembrance of things long gone will help them to STFU.

But probably not…..

In the world of Physics, Dr. John Archibald Wheeler was a very big deal.  But if you weren’t in the world of Physics, you never heard of him. Heard of Black Holes? He’s the guy who coined the name (as well as doing the very early theorizing on them). He also built and ran a nuclear reactor at Hanford in the 1940s which produced some of the nuclear material used in the bombs dropped on Japan. He was one of the co-authors of the classic book on gravity (Gravitation by Misner, Thorne and Wheeler). His students included Richard Feynman and Kip Thorne.

I don’t know how or why exactly it came about, but in August of 2000 Wheeler, age 89, was going to pay a visit to LIGO Hanford. He was also going to visit the B reactor at Hanford, the reactor he built during World War II. There were lectures in Richland, book signings and a dinner in his honor. He was going to be brought to LIGO by Kip Thorne and his wife Carolee.  It was going to quite a show.

At the time I had one of the very new digital cameras, the sort with removable flash memory. I think it was for this reason, and no other, I was assigned as unofficial LIGO Wheeler photographer and general stalker. This pleased me greatly as I had read many of Wheeler’s papers, some of which were applicable to Jim Woodward’s work (Yay, Absorber Theory!).

Wheeler visited for two days, August 15 and 16 of 2000, if I recall correctly. The first day would be at LIGO, a dinner that evening, then the following day a tour of Hanford’s B reactor. Through it all I dutifully tagged along taking pics. I uploaded the pictures on to LIGO’s servers and that was it. I thought I had kept copies for myself, but looking around in later years I couldn’t find any. That was certainly stupid of me.

Then in August of 2013, while looking for something in my garage, I came upon an old CD-R labeled “Garage computer backup – 2005″. Hey, don’t give me crap….Doesn’t everyone have a garage computer? Anyway, I had no idea what was on it, but it was 8 years old and I was curious as to what I thoguth was worth saving eight years earlier. Putting in my PC, what should I find but a bunch of pics I took of Wheeler’s LIGO visit so many years ago. I quickly put them on my hard drive and made sure they were backed up. Unfortunately I didn’t find any pictures of the day of the B reactor visit, just LIGO.

So in the interest of obscure history, he are some pics of Dr. John Archibald Wheeler’s August 2000 visit to LIGO Hanford and a few Wheeler quotes.

LIGO Hanford Director Fred Raab showing Wheeler and Kip Thorne’s wife Carolee the interior of the LVEA building. That’s the evacuated beam tube behind them. Laser safety goggles for everyone here.

Doug Cook (I think) showing Wheeler on of the LIGO optics in the bake room.

Time is defined so that motion looks simple.  John Archibald Wheeler

Wheeler being shown one of the Reference Cavities used for laser testing.

Spacetime tells matter how to move, matter tells spacetime how to curve.  John Archibald Wheeler

Wheeler being shown around the LIGO Hanford grounds by Mark “Ski” Lubinski. That’s the protective concrete cover for the evacuated beam tube. That cover runs off in the distance for 2.4 miles.

In any field find the strangest thing and then explore it.  John Archibald Wheeler

Ski and Wheeler at one of the beam tube cover’s access doors.

There is no law except the law that there is no law.  John Archibald Wheeler

Dinner for Wheeler with Fred Raab making a presentation. Kip Thorne is to the right of Wheeler

Another view of the crowd at Wheeler’s dinner. I don’t know who they all are, but they are very, very smart.

Wheeler being recognized

If you haven’t found something strange during the day, it hasn’t been much of a day. John Archibald Wheeler

The poster at Wheeler’s dinner

Black Holes have no hair.  John Archibald Wheeler

Yet another table of physics nerds/geeks/smart people at Wheeler’s dinner. I don’t recall who these people are!

The “A” table where I somehow ended up. On the right that’s Kip Thorne, then Wheeler, then Carolee (Kip’s wife) and Rai Weiss (brilliant scientist and great guy). On the left is Barry Barrish, director of the entire LIGO project, Dave Ottaway (ner’do’well Australian), my seat and Bill Butler (equally ner’do’well, not Australian but looking quite glassy-eyed)

Dr. John Archibald Wheeler, August 15, 2000

No phenomenon is a physical phenomenon until it is an observed phenomenon.   John Archibald Wheeler

Kip Thorne and John Wheeler

Some people think Wheeler’s gotten crazy in his later years, but he’s always been crazy. Richard P. Feynman

Hey, you looking’ at ME??

I had learned a lot designing and building my Composite H-Quad, but it ultimately left me dissatisfied for several reasons (beyond my usual perfectionist streak). My chief gripe was that it didn’t transport nearly as well as I had envisioned. Sure, it would disassemble just fine into three major pieces. But you see….I sorta forgot about the props when I was doing my design. Unless I went to the trouble of removing the props from the motors (a pain in the ass for several reasons) they extended the length of each arm assembly by ten friggin’ inches!

And the props were, well, bendy. On my first effort to field test the quad I packed it in 4 miles in a large daypack. When I pulled it all out for assembly, I found the top and bottom of the pack had bent the damn props that were sticking out. It flew, but it let me know it wasn’t happy about it with copious amounts of Jello on the video I recorded. Interesting lesson learned.

The other growing problem with the quad was that I had beat the crap out of it while learning (somewhat) to fly. It took some amazing high speed crashes into the ground and survived. Mostly. There were growing cracks in the balsa portions of the fuselage and my repair efforts were marginal. There was only so much that could be done and it looked like it was time for the patient to get its affairs in order.

But the kiss of death was the new Arducopter software, version 3.0, that had come out for my APM flight controller. It greatly increased the reliance on the internal sensors, especially the magnetometer (the compass) and the accelerometers. If these items were addressed, the thing could fly amazingly and hold a solid position on its own in high wind. But if the vibration was too high for the accelerometers or magnetic interference was generated by the wiring, the craft could fly away in some random direction or toilet bowl into the ground. The latter is actually a recognized multicopter flying term. You don’t want it to happen. It’s expensive.

In testing, I found my battery wiring was generating a magnetic field equal to over 70% of the Earth’s magnetic field….but in the wrong direction. Toilet bowl time. The only way to rectify that would be to rearrange the power wiring. But in an amazing flash of foresight when I had assembled my quad, I installed all its power wiring into the fuselage then epoxied the damn thing shut. So it was not fixable. Time for a rebuild. And what I find really appealing about this “hobby” (or whatever you want to call this money pit) is that at this point in its evolution if you want something even slightly unconventional you have to build it yourself.

I had seen a few folding quad designs where the arms folded in against the sides of the body and the props sort of “nested” there. I thought this concept provided a high measure of protection and would avoid “kinky prop syndrome”. I was also looking for considerably longer flight times than the 9 to 10 minutes I had been getting, so that meant slower turning motors and larger props. And all the wiring had to be kept well away from the flight controller, and oh while you’re at it, minimize vibrations to the flight controller too. How hard could it be?

I liked the general result I had previously achieved using a composite of G-10 fiberglass sheeting and wood, and this time I chose to use Basswood. Much stronger than Balsa but still very light. So I decided to use similar construction techniques. The arms were mainly composed of 10 mm square Basswood to take compressive loads and skinned with a thin G-10 fiberglass layer on the tops and bottoms to take the tension. This engineering stuff is good for something after all.

This shows the arm construction sequence. Top are the individual Basswood pieces. Center are the G-10 skins. Bottom is an assembled arm.

Setting the arm length was challenging and required a lot of CAD time. To nest, the props needed to be shorter than the fuselage but they also needed to fold outward to the right, balanced locations. Throw into this mix the need to pick the right motors matched to the right props to lift an aircraft weight I hadn’t fully worked out. I think I had one more variable than equations. Doing a lot of analysis, I was eventually able to freeze a design that looked appealing.

A size comparison during the Folder’s early construction between it and my previous Composite H-Quad.

The body had a similar design to the arms, with Basswood compression spacers and G-10 top and bottom plates. The wiring was bundled tightly together to minimize the magnetic field generated but also designed to be removable. See, I can learn.

The folding mechanism merely pivots the arms on aluminum 4-40 screws, with adjacent screws that tighten to lock the arm in place. All the nuts are Nylock types to prevent loosening from vibration.

A side view of the Folder in flight. The “golf balls” are actually foam Nerf balls which act as temporary energy absorbing landing gear during testing. They are easily removable and aren’t necessary if one flies reasonably. I must try that some time.

Another view of the Folder in flight. The small forward FPV camera is visible below and to the left of the HD recording camera. The 5.8 GHz circularly polarized antenna is the black mushroom at the rear.

The Folder in its folded position, ready for transport.

The underside of the Folder in the folded position. The downward looking FPV camera is visible in the upper lightening hole in the white area.

To get the power wiring away from the magnetic compass I mounted the flight controller high on a box containing the battery. This box superstructure also contained other electronics like the GPS, RC receiver and onscreen display board. The box was constructed of 2.5 mm plywood (to help kill vibrations) and connected to the fuselage via four Sorbothane vibration isolation mounts. The large mass of the box, combined with the Sorbothane isolation worked fantastically to reduce vibrations to the flight controller (which was further isolated on its own special elastomer gel mounts). The compass interference was reduced from my previous 70%+ to about 6%. That’s outstanding.

Camera placement was an issue I went back and forth on. I wanted to be able to carry an HD video camera or one of two high-pixel still cameras either pointing forward or downward for mapping. And I needed room for a decent FPV camera. Given the narrow fuselage and the odds that I’ll occasionally land hard, nothing of any importance (like expensive video cams!!) can be mounted underneath.

So of course I ended up with two FPV cameras. A very small one, facing forward tucked under the main camera mount as the primary flight camera. The second one looks straight down through the fuselage. They feed through a video switcher which I can control from my RC transmitter. This allows me to look either straight ahead for normal flight, or straight down to verify position for mapping or other important imaging. That downward looking camera is also relocatable to the rear of the fuselage should I have need for a “back up camera”. And I can think of some situations that could prove handy.

Folding Composite H-Quad parts layout

So at this point I’ve spent well over a month wringing it out through a series of flight testing and tuning. I’m really pleased. It performs well and flies long. I’ve had a few semi-crashes in its very initial flights and it survived them surprisingly well. The only negative I have to say about it is it’s somewhat feral, with tendencies to attack if one isn’t paying attention.

When drones attack: A large, powerful quad + a moment’s inattention during the tuning process = seven sutures. I’m a bit concerned now that it’s tasted human blood.

Final specs:

• All Up Weight (3.3 Amp hour battery and HD camera): 1,823 grams (4.02 pounds)
• Motor to motor dimension: 736 mm (29″)
• Fuselage length: 470 mm (18.5″)
• Current at hover: 12.5 Amps
• Measured hover time at full weight, to 3.6 volts per cell: 14+ minutes

So it’s ready for the field…..I just need to wrap up the other two drone projects in the pipeline.

Since it’s been summer and Joshua Tree has been either too hot or just plain closed due to the government shut down, it’s given me impetus to do something I’ve been wanting to do for a while but have been too lazy. That is, work up a better picture of exactly what areas have been cleared by all the searching to date in the area.

What I had in mind was that instead of just using lines of a fixed pixel width to show the search tracks in Google Earth, perhaps there was a way to show a scaled width. That is, a person hiking through an area is able to observe some distance to the right and left thus clearing a defined swath of area. I figured this should be a simple thing to do with a GIS program….if I had a GIS program and knew how to do it. Instead I had Google Earth and kml files.

After a lot of hunting I discovered the following command:

<gx:physicalWidth>XX</gx:physicalWidth>

If this is inserted into a kml linestring  file at the right location, XX can be specified as the line width in meters. Neat!

I decided to pull all the tracks between Quail Mountain and Smith Water Canyon, as well as those tracks northerly of Smith Water. The original search tracks from June/July of 2010 were left as black and those search tracks since then left as red. All the tracks were edited to have a line width of 50 meters (164 feet). My assumption was that clues could be reasonably seen within 82 feet on each side of a searcher. This was probably overstating it a bit when looking upslope but very much understating the observable distances when looking down slopes, but on the whole a reasonable estimate. Your mileage may vary.

After bringing them into Google Earth they were set to be mostly transparent. By doing so when search tracks overlapped, the color, either red or black, got much darker. Thus the darkness of color in a specific area is a direct reflection as to how well it has been covered.

Here are screen grabs of what the final product was, using both a satellite image and a topo map as a base map. Following those, there is a closeup view of the southeast slopes of the mouth of Smith Water showing the coverage.

Looking down on the extent of Smith Water Canyon showing search tracks and areas clear to date (October, 2013). This is using a satellite image as the base map.

Same coverage as the previous satellite view but using a topo map as a base for clarity.

Looking southeasterly across Smith Water canyon at its southeast slopes. Note the extensive coverage to date of the steep slopes. The lower parts of the bowl, to the left in the image, are fairly open and have been scanned from above.

I found it all very enlightening. Areas near the southeast mouth which I still harbored some interest in look like they’ve been covered quite well. Lots of dark red. There are a few holes but nothing really too intriguing. The significant holes that do remain are in extraordinarily nasty areas on the south slopes in central Smith Water. We’ve made a few passes through there, but always going downhill….very carefully. What to do about those will require some thought.

To really get a sense for the coverage you have to download the kml file, load it into Google Earth and twirl it around yourself. So here it is:

KML file for 50 meter search coverage

As an alternative, I did a fly-through in Google Earth and saved it to YouTube. It’s not that great but if you don’t want to mess with Google Earth yourself it’s a good way to get a feel for the search coverage as well as the terrain (If you have 12 minutes of your life you never want back). Of course I do need to mention I’ve learned the hard way that Google Earth makes the ground surface appear much smoother than it actually is. At the end of the fly-through I added a flight from the Quail Mountain ridge taking the most direct route toward Smith Water. This was the idea I was struck with during my JT53 hike to Quail Ridge. That could be a fun trip…..

Date: 10/22/2013

Participants: Tom Mahood

General search area: The southeast slopes of the mouth of Smith Water Canyon.

Rationale for searching this area:

This was a combination shakedown hike (first of the season) and sort of a mop up of unsearched areas in the SE Smith Water Canyon area. Having just completed the exercise of laying out 50 meter coverage swaths along previous search tracks it was obvious there were a few holes left in the coverage of this area. In essence, filling these in would provide (theoretically) fairly complete ground coverage of the SE Smith Water area.

Impressions of area and findings:

I quickly discovered why the spots I had targeted were holes in previous search coverage. They were mostly pretty ugly pieces of terrain. Still, in a burst of OCD, I dutifully hit them all.

Coverage level:

I’d estimate my coverage to be 80% within 75 feet of my travel path at a minimum, and in some areas a much wider swath. I think my 50 meter swath reasoning holds for this search.

Comments:

Over Summer I deliberately avoided thinking of this problem to give myself a break and approach things fresh. Between that break, the coverage analysis I just completed, and this mop up I’ve come to a few conclusions.

• The idea of Bill heading to Smith Water Canyon to reload on water, and him viewing it as a non-emergency situation, still seems to best fit what we know at this point.
• Bill must be within an 11.1 mile radius of the Serin Drive cell tower, and probably closer to the 10.6 mile radius. Given how the Verizon cell system figures distance from the tower, any place further out is probably a waste of time.
• Smith Water Canyon puts a hard northerly limit on Bill’s travels. It seems very unlikely he would/could have passed beyond Smith Water. Thus he’s to be found somewhere between the 11.1 mile radius and Smith Water Canyon.
• After fairly extensive searching, I’m now of the opinion Bill isn’t going to be found in the SE Smith Water area. He’s somewhere else.

In hindsight, there was a flaw in my reasoning that made SE Smith Water look so attractive. I made the assumption that Bill experienced some injury before reaching the crest of Smith Water, when he was north of Quail Ridge in the canyons with no cell coverage. And to get into Smith Water there’s only two ways to do so without passing though an area of excellent cell reception at the southerly rim of the canyon.

One is to pass through an obvious saddle to the west of the rim high point (imaged below from Quail Ridge on my third anniversary hike….It’s just to the left of the center flat ridgeline) then down a narrow canyon to the bottom of Smith Water. The lower parts of that canyon had been covered well by the Sierra Madre team during the initial search and the upper parts have been covered on two other searches since. So Bill’s not in that canyon.

A zoomed panorama looking toward Smith Water from the Quail Mountain ridge. The level of rockiness approaching the crest becomes apparent in this zoom. It’s not visible when viewed normally from this distance.

The only other way to get to the 10.6 mile radius without pinging the tower would be to travel to the northeast, staying somewhat lower and below the cell coverage zone, and finally arriving at the SE slopes of the mouth of Smith Water. I explained that reasoning here. It made a lot of sense until searching showed Bill wasn’t there. Again, another case of a beautiful theory spoiled by ugly fact.

Thinking things over on the hike out I realized I may have underestimated Bill’s abilities. I just assumed he got injured well before he got to Smith Water, thus explaining the considerable span of time before the cell ping. An injury would have slowed him substantially and could explain the odd timing.

But suppose Bill was doing fine as he reached the southerly rim of Smith Water. At that point he’d be able to see the bottom and “only” had to descend 1,200′ to get to the water. Even though he was in the middle of great cell reception there was no need to turn on the phone or call for help. So down he starts.

Now having been in this area a few times I will swear an obscenity-laced oath that the central Smith Water Canyon area is some of the nastiest and most difficult terrain I’ve encountered in this effort. Yes, worse than the SE Smith Water area. I say this as someone who was well equipped, hydrated, rested and with hiking poles. The chances of someone tired, dehydrated and without poles successfully making that descent seem slim. The chance of becoming seriously injured or incapacitated are large.

Have a close look at the image below. I think Bill’s somewhere in it. It’s part of a larger image I took on our JT44 trip, which was a reconnaissance of Smith Water from the northerly slopes of the canyon. It’s pretty much taken looking due South. The black lines are the tracks from the initial search, and the red lines tracks since then, updated to the most recent. The orange line is the 11.1 mile radius from the Serin tower. There’s a bit of blue line to the lower right, which is the 10.6 mile radius. But the 10.6 mile radius is essentially the canyon bottom.

Looking southerly toward the southerly slopes of the central Smith Water Canyon area showing approximate cumulative search tracks to date.

There are few things that aren’t immediately obvious in the image. The first is that the area above the orange line is a zone of excellent cell coverage. Descending below the orange line cuts it off. Also that peak in the image is the back side of the high point visible in the shot I took from Quail Ridge. Not having a GPS and being a map and compass kind of guy, Bill would be inclined to sight on a distant landmark for navigation purposes. And that peak would certainly serve that purpose to get to Smith Water.

Note that right smack in the middle of the image is the reported “odor location”. Further note that above it and to its left, just below the orange line, is some appallingly nasty terrain (I speak from personal experience as the two red lines to the left of that area were mine and the one to the right is the ever-gullible Paul Caraher). And as a final observation, those nice, vertical rock faces generally point at the Serin Drive cell tower. Can you say, “reflector”?

So from a possibilities viewpoint, I like this bowl area a lot. From every other sensible viewpoint, I loathe it. It’s a scary place to be climbing around.

The reported odor location deserves a bit of discussion. It was initially reported by Team 2 on 6/29 (Tuesday). It was then followed up by the Sierra Madre team on 7/3. No source for an odor was ever found. But something I always found a bit odd was that there weren’t any GPS tracks that went exactly to the reported coordinates (which I’ve triple checked). The image below shows the three search tracks of Team 2 in black and the reported odor location coordinates. None of the team’s tracks actually hit the location. (Note: The black line that’s shown passing through the odor location’s coordinates was from Team 1 on 6/30, after the location had been reported.)

Reported odor location coordinates as compared to the actual GPS tracks (in black) of the search team that reported the odor.

Looking through the search records shows that Team 2 had three members. There were also three GPS tracks turned in for the team. So it’s not like there was someone along who didn’t have a GPS and noted an odor. I’m thinking the odor was actually noted where the one black track jogs off to the SE from the main canyon, stops then returns. This spot is in the canyon bottom easterly of where I’ve been showing the odor location. The red line that passes through where the black line turns back was me on the Smith Water Death March (JT27) descent. I was probably so happy to be back on the flats I wasn’t looking around very carefully. But higher up that drainage remains a good possibility.

Access to this area is difficult at best. To get to the spot in Smith Water just beneath the odor location from the usual Samuelson’s Rocks trailhead is about 4 1/2 miles with a total elevation differential of 750′. To the same spot from the Covington trailhead is 2.3 miles and 600′ of elevation differential. Both routes require traversing meat-eating Acacias groves in the bottom of Smith Water Canyon. And then the climbing begins. This isn’t an area you want to be in unless you really know what you are doing and know how to manage risk.

GPS mileage for this trip: 10.0 miles

Cumulative GPS mileage to date: 716.9 miles

GPS tracks for this trip in Google Earth kml format
(Note: I’m providing the kml file with the new 50 meter swath display that I’ve now adopted as my standard practice)

GPS tracks for this trip in Garmin gdb format

GPS tracks for this trip in gpx format

Cumulative search tracks as of JT54. JT54 tracks are in blue, initial search tracks are in black and other tracks since then in red. The light blue line is the 10.6 mile radius from the Serin Drive cell tower and the orange line is the 11.1 mile radius.

A perspective view, looking southerly, at the cumulative search coverage in the SE Smith Water area. Blue track is JT54. Note that the vertical slopes below the “bottle” location have been extensively covered.

(Note: Written by Adam Marsland)

Date: 10/24/2013

Participants:  Adam Marsland, Anny Celsi

General search area: Slopes and highlands south and east of Quail Spring.

Rationale for searching this area:

JT51 and JT52, conducted by myself and Tom M., respectively, both arrived at the same probable ping site on a slope in the southeast Smith Water Canyon area, just north of a low ridge behind which there is an extensive dead zone leading south to Quail Mountain, thus providing a possible route for Bill to have traversed and then to ping the tower shortly after daybreak Sunday.

Tom has since moved on to a new theory of the source of the ping, but I felt that this spot still fit the known facts the best, and in blundering my way out of the area on JT51, following what seemed at first like the path of least resistance, I wound up in a less searched and increasingly rugged area south of Quail Spring. With night falling on the previous trip, I was not able to search, and in fact in the more rugged areas was so focused on my own survival that I probably would not have noticed Bill unless I tripped over him. Thinking Bill might have selected a similar route for similar reasons, I wanted a chance to get back to that area and make a thorough search of the slopes and highlands immediately south of Quail Spring in daylight.

Impressions of Area And Findings:

There are a lot of rocks.

It was interesting to observe the lightly-searched canyon area roughly midway between Quail Mountain and Smith Water Canyon. The slopes into the canyons near Quail Wash are bad, bad, bad. Very easy to imagine someone coming to grief there…and the slopes get worse and worse the father you go down. Past a certain point, going up is so steep, you are committed whether you like it or not.

Coverage Level:

Excellent. I noted that on previous trips I might have been too focused on scouting the area ahead to fully take in my surroundings at times, so I kept my eyes open and searched any outcroppings or likely hiding places thoroughly. I also had a companion who was as much as a 200 yards away at times, so there was quite a bit more area covered than the GPS tracks would indicate.

Comments:

My fourth time to Joshua Tree to look for Bill, my first time in six months, and also the first time that Tom Mahood’s understandable exasperation, both at the bleak surroundings and at the complete lack of any trace of Bill, started to rub off on me. It wasn’t Tom’s fault; three months on a world tour had left me paunchy and less fit than my last trip out, and I was also psychologically this time more determined to be thorough and, if possible, find something that would shed light on the mystery. And we also, through careful but still faulty orienteering, wound up covering about twice as much ground, much of it outside the planned search area, as I had planned to.

I had a companion for this trip, Anny Celsi, both an old friend and a free-lance stringer for NPR, who thought the search and the crowdsourcing angle might make for a good story. I was glad for the company and for the extra pairs of eyes, though the recording gear made for bulkier packing all ’round.

We got started at 10ish, parking on Park Road in a spot that (I thought) would be easy to find on return, which was important because the cheap GPS program on my iPhone was refusing to work in airplane mode and without a 4G signal to kickstart it. This meant no tracks on the way in, high battery consumption, and also no way to mark the location of the car for the way out. I was able to kickstart the thing once we got to Quail Wash, so at the very least I was able to get usable GPS tracks for the main part of the search, but it did not bode well.

Since I wanted to search southerly of Quail Spring, we made a diagonal swath across the alluvial fan, hitting the old road from Park Blvd. to Quail near Samuelson’s Rock. At that point, a well-loaded, middle aged male hiker passed about 30 yards away, seemingly headed for Smith Water. I had already noted more cars than usual parked on the main road. I indicated to Anny that we should not initiate contact with him, but I did wonder if he might perchance be a fellow Ewasko searcher. The gentleman looked a little like Bill himself. As unnerving as being out here alone can sometimes be, seeing a solitary hiker pass quietly nearby was a little weird too.

After arriving at Quail Wash and getting adjusted, we set out on the search. The original plan was for me to retrace my JT51 route backwards but it took us so long to get underway that instead I decided to just head straight up the steep slopes before us, with Anny heading up the same direction a good distance off. Unfortunately, as much as I had tried to carefully consult with the maps at my disposal, and in spite it being my third trip to the area, I had confused the two ridgelines and we were actually on the next range east. Thus the first and most strenuous part of the hike was fairly useless from a Bill-searching standpoint. We did eventually get more towards the desired area, however it wound up being a less thorough, more sprawling and southerly search than I had intended.

We found lots and lots of bones, most easily identifiable as Big Horn Sheep bones. One set drew my attention as they were larger and actually contained a joint section, but even on the remote chance they were human (which I doubted, and Tom subsequently confirmed), it would be a smaller and more recently deceased person than Bill. At one point while doggedly searching the ridge I heard a telltale rattle and looked up to see an unhappy rattlesnake about three feet away, coiled up and asking me in snake language what the hell I thought I was doing. After a few seconds of stupidly staring at each other he slunk into a nearby bush, and not for the first time I blessed the good situational awareness, combined with inherent laziness, of your average rattlesnake.

Probable Big Horn Sheep bones. For size comparison, boot is size 13.

After that incident Anny and I regrouped in the high, relatively level area just above the long north-south canyon that runs about halfway between Quail Mountain and Smith Water Canyon. This was an interesting area to search. It was further south than I had yet been, a significant (though traversable) barrier to progress if this was the route Bill had taken. For Bill to have come this far south after pinging the cell phone tower was unlikely — it was, after all, almost halfway back to Quail, and this was a distance it had already taken him more than a day to cover (assuming he was injured en route). However, on looking at the map, it does appear that from the canyon area north, we did follow what would have been the easiest topographical route to and from the suspected ping location, so in retrospect, it was still worth a look.

My goal from that point was to head to the theoretical ping location on the slopes of Smith Water, but about halfway there I was perplexed to run across a canyon wash I did not remember from a previous trip. We elected to get in the wash and start down, which proved to be very perilous going. It’s when things begin to go awry with the planned hike that I think one might blunder across Bill by sheer stupidity. Certainly, it was very easy to imagine Bill getting in this same wash, and never getting out of it. We were more fortunate, eventually escaping the rocky, obstacle-laden wash and making our way down the treacherous, steep slopes of the canyon running south from the Quail Springs area that I had gone down (in the dark) on my last trip. As a possible area for Bill to have tried and failed to get down from the mountains, I did want to get another look at this area in the light, so I was glad about that.

I wasn’t glad about much else. The descent was just as bad in the daylight and I had another hiker (and a dying iPhone) to be concerned with. Finally, after much careful negotiation, we got into the canyon area and resumed the same route over the low lying ridge east of Quail Mountain and thence north to Park Blvd. that I had taken on JT 51. I headed to where I remembered I parked my car, and found myself to be nearly a mile off, which had me cursing in the dark, since this has happened every single time I have parked on Park Boulevard (yes, Tom, I will bring a real GPS next time).

My burning desire for a margarita to cool off with was further thwarted as I started up the car and headed up the now-darkened highway to pick up Anny; I could tell by how the car was handling and sounding that I had a flat tire. Sure enough, an inspection revealed this to be the case. We were able to drive on the rim into cell range, and thence to the West Entrance. What transpired then has given me a new theory of the Bill Ewasko case. I now think Bill got out safely, but only was able to call AAA. Bill gave AAA his precise location, which AAA then blithely decided was the address of the Visitor’s Center in 29 Palms. When Bill called back, AAA then chastised him for giving them a bogus address.

OK, maybe that’s not what happened to Bill. But it happened to somebody I know. At least I got to Pappy & Harriet’s…eventually.

This was not a fun day. It was hotter than I expected, and I felt more conscientious than usual. I wasn’t just adding fun lines to Tom’s grid. I really felt that, with all the work that Tom has done narrowing the options, an answer would be near, possibly forthcoming on this very trip. And yet I became increasingly aware that there are just so many places a person could conceivably wind up, even if you have a sense of where they were at one point.

I do have a point of disagreement with Tom about coverage and clearance. I think it’s possible that Bill could be in the area I was looking in, the heavily-searched highlands above the SE mouth of Smith Water Canyon, or even under a bush in the heavily-traveled Quail Wash. All it takes is one nearby feature to be sufficiently distracting, or the surroundings to be arranged in a certain way as to conceal from the searcher. What is eerie to me, though, is the lack of any trace of the man, with the possible exception of a bandana. There’s not a lot of trace of human passage up there. It seems like a water bottle, a note, some scrap, would escape whatever debris field Bill lies in, or would have fallen out of his pocket en route, and would at least give us a new clue to chew on. But there’s just…nothing. Which, perhaps, speaks well of Tom’s new theory that Bill is stuck in a particularly nasty little spot near the Smith Water “Odor location”, since it would be harder for clues to escape from such a dungeon.

On leaving the area, I was so tired, footsore and hot — not to mention sobered by the vast amount of areas that were very unlikely but still conceivably possible for Bill to be — that I contemplated this being my last trip out here. After all, as Tom has pointed out, this is not the nicest area of Joshua Tree to visit. It’s grim and desolate country, not to mention dangerous in the steep areas. But that was before I realized I’d only partially searched the area I had intended to — and that that search didn’t rule anything in or out. In fact, it was completely plausible that Bill could have pinged the phone in the spot I picked out in JT 51, noted Joshua Tree (and possibly Park Road) in the distance, headed south along the easiest terrain, and then expired choosing one of the various bad options to descend from that location. My theory was still totally viable — I hadn’t really proved or disproved anything. I just searched some areas that hadn’t gotten looked at that much.

Arrrgh.

P.S. The America’s Best Value in Yucca Valley was a very pleasant surprise. Too bad we only got to stay there for seven hours or so.

GPS mileage for this trip: 4.81 miles (actual miles covered: 11)

Cumulative GPS mileage to date: 721.7 miles

GPS tracks for this trip in Google Earth kml format

GPS tracks for this trip in Garmin gdb format

GPS tracks for this trip in gpx format 

All search tracks up to JT55 (shown in blue, with 50 meter viewscape). Orange line is 11.1 mile Serin cell tower radius. Black lines are GPS tracks from original search and red lines search tracks since.

All tracks in SE Smith Water Canyon area showing 50 meter viewscapes. JT55 shown as blue line.