FLYING R/C Lunar Module Quadcopter project

[georgegassaway]

Please note: A lot of my posts describing this project/build are very long, and many contain a lot of big images. As such I strongly recommend that anyone who replies, to not quote a complete posting, but to only quote the portion relative to any comments or questions. Thank you.


When I got into model rocketry in 1970, there were some “dream projects” I thought would be neat to do someday. A few of them, I’ve ended up doing.

But one of them, seemed so wildly farfetched……it could only be a “dream” and never a reality.

As the Apollo program was going on with lunar landings, I had an idea for an “ultimate dream model” - A model of the Lunar Module that I could fly by R/C, hover, and land (I wasn’t into R/C at all at the time, that too was a dream for someday, maybe). For 1970 an R/C Lunar Module was an insanely impossible model to even think might be possible to do.

But what was even more farfetched at that time was to be able to have R/C small and light enough to be able to control such a model. And most critically have the ability to keep itself level, using technology that in 1970 would have cost millions and weighed dozens of pounds (for the Gyro-based guidance system itself) instead $30 and weighing well under an ounce. Back when a “computer” would require its own room in a house, not fit in your pants pocket or on your wrist.

Well, decades later, all sorts of massive changes for computers, R/C, micro-controllers, and incredibly tiny and accurate gyro and accelerometer sensors (cel phone and “Wii” technology). Improvements in R/C gear, smaller, way better batteries (LiPo’s). And some incredibly brilliant development of a Rube Goldberg flying model contraption called a Multicopter. Using a gang of electric motors and model plane type props to make a helicopter model fly incredibly well, thanks to the use of the tiny gyro and accelerometer chips on the same board as an Arduino type micro controller, programmed to make it fly.

I got to try one out about 3-4 years ago, but didn’t really get into it until early 2015. I got a $10 nylon frame for a “250” sized racing quadcopter, found/ordered the parts it would need (like Speed Controllers, motors, props, Flight Controller) and built it. It was incredible to fly, a lot of fun.

And then…. I got to thinking about the old ultimate dream model from 1970.

I could make a flying R/C Lunar Module, using Multicopter technology! OK, so the original idea was rocket thrust….. but that’s still not practical for me in any case. While in theory an extremely fine-throttleable hybrid rocket motor for hobbyists could maybe be made, lots of hobby dollars and time, and someone else’s expertise to make that happen someday (and even with gimbaling, it would need some means of controlling the roll axis).

So, I planned to build one. I even went so far as to make up a very crude octagonal box out of poster paper, to represent the shape of a Descent Stage, and mounted that on top of my 250 Quad. Well, it flew, but it flew like it was “drunk”, from the mass and drag on top, and the way that in the top view the octagonal box blocked some of the airflow from the four propellers, the thrust was reduced and it just did not handle well. But I didn’t expect it to fly great, I just wanted to get some idea of what it would be like. I planned to build a Lunar Module Quad a few months later, fall of 2015. But other things came up, and so I pushed it a year to this fall.

Once I did want to “start building”, well, not so fast. First, a lot of things to figure out. What size? And was I going to build that crazy jigsaw assembly of an Ascent Stage piece by piece from custom hand-drawn patterns and hand-cut pieces? No way! This is a flying project for fun, not a detailed scale project. So for the Ascent Stage, I would use patterns for a cardboard cutout model, printed onto thick poster paper, at a scaled-up size.

OK, even a cardboard cutout model of something like this is not something simple to throw together in an hour, but still far easier than drawing up and cutting out pieces from scratch, and painting the various colors. There are several cardboard model options in 1/48 scale. So again, what size? 1/24 would be smaller than I’d like. 1/12 would be really nice, very impressive size. But pretty heavy, requiring some really big powerful motors, more expensive Speed Controllers (ESC’s), more expensive props, and really big expensive batteries.

After running some numbers and some guesstimates on the likely mass for the structure, plus the mass of the motors, ESC’s, battery pack and such, 1/16 scale seemed to be about right for what i wanted to do. I figured it might weigh about 1400 grams of so all up weight. I checked into RC groups to ask some questions, compare what kind of motors, props, ESC’s and battery capacity was used for a model of about the size and weight of mine. Based on that, I then tried the components in the multicopter versions of “Ecalc”, which kind of like the electric powered model aircraft version of using “Rocksim” to determine flight performance:

So what I chose to get for it were:
Motors (out runners) MRM Titan 2212-980KV
ESC’s - MRM Zeus 30 Amp BLheli
Props - 10 x 4.5

For battery power, I was planning on using two 2200 mAh 3S (1.1 volt) Lipos, connected in parallel, to act like a 4400 mAh 3S battery pack (Have a good number of 2200’s to power electric models like Radians). If it needed some more capacity, I could use a 5000 mAh battery but hoped that 4400 mAh would provide sufficient fight time.

For a Flight Controller (F.C.) , a Mini-APM, as I have used before.
https://www.rcgroups.com/forums/showthread.php?t=2182722

That controller runs “Ardupilot” and can be programmed using “APM Mission Planner”
https://www.ardupilot.org/planner/index.html

 

[georgegassaway]

So with the size and major components determined, then to plan out the actual construction. I used some drawings of the LM to determine the size more accurately for drawing up a top view pattern for the Descent Module. Some of the drawings would also be useful later for determining size or patterns for other assemblies.

I had a plan for doing this project in three phases.

Phase 1 was to get the basic model, pretty much the bare Descent stage octagonal structure, flying as a Quadcopter to just see *IF* it worked decently. If it didn’t work for basic concept reasons, such as poor flight characteristics, then it would end there. If it did work but had some damage during early testing…. at least it would not be damaging more difficult parts to be added later, or ruining the appearance.

Phase 2 - Passing basic flight testing, to begin working on permanent additional assemblies such as the Ascent Stage (made of poster paper printed from scaled-up cardboard model patterns), mostly-accurate looking legs, and Base plate (hatch) with descent engine nozzle.

Phase 3 - Make-over to make it look good and realistic, adding gold foil or gold mylar, adding black coverings, markings, SOME details (like ladder), RCS nozzles, and some other extra goodies.

But this is NOT intended to be a really accurate “scale model”. There are lots of very accurate scale models of Lunar Modules sitting on shelves, tabletops, inside glass cases, in museums, and so forth. THIS is for FLYING. So it only has to look like a LM in the air (ignoring the arms, motors, and props), not super-accurate.

My goal is for it to be the BEST multirotor R/C flying Lunar Module in the world. But just by the fact that it exists, it will then become the best……. because from my google searching, it seems nobody has done one!

Although I do have to give credit, a gentleman in Australia, Peter Aylward, built a 1/20 model using two coaxial rotor blades. It was an existing coaxial Copter frame with the fuselage converted to a Lunar Module. He made this a few years ago, apparently by November 2012. He only made two posts on RC Groups about it, no updates on the model.

[video=youtube;R726M2X6J5U]https://www.youtube.com/watch?v=R726M2X6J5U[/video]

 

[georgegassaway]

Back to to the 1/16 Lunar Quadule…..

I made up a crude mockup of the Descent Stage using foam board. That was useful in getting an idea of the actual size at 1/16 scale, and for planning out how to do the build and where/how to mount various components. As much as possible would be mounted inside of the Descent Module, including batteries, ESC’s, wiring, and so forth. To have access to all of that….. it would be necessary to have a large opening in the bottom, to be covered over later by a base plate that would include a simulated Descent engine nozzle. If this was a car……the removable base plate would be like a car hood to access the engine compartment.

The arms for holding the motors/props would be 10mm (about 3/8&#8221 square graphite tubing. That would be sturdy/stiff enough for the job as well as allow running the motor wiring inside of the hole inside.

Here is a photo of the foam board mock-up at right next to a Phantom 1 quadcopter.

Basic structural design of the actual flying Descent Module is all-wood: basswood top bulkhead, basswood lower bulkhead ring, vertical balsa columns joining the two, and 1/16” balsa sides.

Using the drawing pattern, cut out pieces for assembling the base ring.

Used three layers of 1/16” basswood. It was very sturdy.

For the top bulkhead, glued 1/16” basswood sheeting edge-to-edge to make two 10 x 10” sheets. Applied laminating epoxy and laid one cross-grain over the other to create a sort of 2-ply 1/8” sheet. Vacuum bagged it to securely bond them together flat. In the photo below, it is inside of the vac-bag (one gallon ziplock bag), some wrinkles visible because this was shot after the vacuum was removed.

Used the marking guide to cut the 10 x 10 2-ply basswood sheet into the shape for the top bulkhead. Also bent some 1/16” music wire to use for the upper nearly horizontal “V” struts to support the top of the legs. Made up two sets of bent wire for each leg, the two secured together by a wrap of thread soaked with thin CA. In the photo below, this is looking at the underside of the bulkhead and four of the strut sets. An extra set is off to the left, upright, not connected to show them as separate pieces

Those 1/16” music wire struts were anchored securely to the top bulkhead, using glue and with some 1/16” hardwood strips covering over them. 4-40 bolts were used for helping to anchoring them in place. This was done because in a very hard landing or crash, I did not want those to rip out easily as repairing them owed be very difficult. Also seen in this photo was a dry-fit of the 10mm square graphite tubing for the arms, and four bolt locations that are used for anchoring a number of components top to bottom.

In this photo, next to a Phantom 1, a dry-fit view with the upper bulkhead held a few inches above the lower bulkhead frame. Also notable is the “4-cornered star” assembly in the center, which has four blue objects in it. That is the anti-vibration mount for the Flight Controller, the four blue things are rubber pieces that go between the base “X” of the mount to the upper “X” part that the Flight controller attaches to. The anti-vibration is used for the F.C. since multicopters do have a lot of vibration, and vibrations are not good for the accuracy of the F.C.’s sensors

 

[georgegassaway]

More on the Descent stage structure. Eight vertical columns were made out of 1/4” triangular balsa glued to 1/4” square balsa. They were glued to the bulkhead corners, and a jig was used to help to assure they were vertical. Also visible in this view, near the corners, are eight 4-40 blind nuts to be used for attaching the lower near-horizontal struts for the legs.

The upper bulkhead was glued in place on top of the 8 columns. The 10mm square graphite tubing for the arms were spot-glued near the edges, not glued fully across since vertical bolts near the center secured them most of the way.

And here it is after adding the 1/16” balsa sheeting for the sides.

A view inside the lower bulkhead frame. The four bolts near the center were later changed to longer ones. On the inside, those bolts run down from the graphite tubes, to a spacer, to the Power Distribution Board (PDB), to a taller spacer, and finally a wooden plate to support the battery which also is used to secure a couple of velcro straps wrapped around that wooden plate and battery so it won’t fall out.

Wiring harness. At lower right corner, the Deans “T” connector for the Lipo battery. Upper left, a couple of connectors coming from a 5 volt voltage regulator red that shrink over the regulator). At bottom, slightly left, two black ground wires for future use, and two connectors to provide power directly from the battery pack (rated 11.1 volts, usually over 12 volts fully charged). The other wiring in four places are to the 30 Amp ESC’s to power and control the motors. The white wire coming off of each ESC (twisted with a black wire and connector) is the control wire that will be plugged into the appropriate connector from the Flight controller, to control the speed of each motor. The ESC’s outermost red, yellow, and white wires, with “bullet” connectors, are the three wires to run a brushless outrunner motor.

Closeup view of the PDB. It has thick copper traces to carry the current from the battery to distribute it to each wire soldered to it, the traces running according to the + and - signs printed on it.

 

[georgegassaway]

Getting closer….

The Descent Stage with the motors mounted onto it. Also, used foam board to make up an almost profile rectangular dummy of the Ascent Stage. I added this in part to simulate the aerodynamic drag effects in forward flight. Also, for visual orientation of where “front” was (Used a front view of a color/makings drawing, printed hit out and bonded it to the foamboard). No attempt to make it look accurate in 3D since that’s a Phase 2 thing, while Phase 1’s objective was to just get it off the ground for flight testing, regardless of the looks.

Another view, nearly 90 degrees, showing the dummy Ascent Stage. I didn’t even get around to filling in the sides, but that was OK since I do need to see the Flight Control board’s status LED’s to see if its ready and if there is GPS lock. Later I’ll add some LED’s in another location to see that status. So yes, in this view, you can see the Flight Controller mounted in place onto the anti-vibration mount. It did not have the GPS receiver at that time.

Now, not shown is the installation of the wiring inside of the Descent Stage. Basically, ran the 3 wires from the motors inside of the 10m square tubing, exiting thru a hole drilled into the tubes a couple of inches inside. The wiring harness shown previously was secured in place using the four center bolts, with spacers. The wires from the motors were plugged into the ESC’s. Now, for a Quadcopter using an “X” configuration, two motors need to spin clockwise and two need to spin counter-clockwise, at the correct locations. By default they all spun the same way. I had to swap two of the three wires to get a specific motor to spin the opposite way…one of the benefits of using bullet connectors so the wiring can be swapped, and no big deal.

The receiver was hooked up to one of the +5V connectors.

The Mini-APM Flight Controller (F.C.) has a LOT of wires, many of which are not used for this project. Of the R/C signal input wires, six were connected between the 6 channel receiver and the F.C. Had to double-check later to confirm that the correct channel axis from the transmitter (such as throttle and pitch) were being read by the F.C. on the correct axis…… for example throttle stick on the transmitter causing roll inputs to the F.C. would be BAD! Also, the output wires from the F.C. t control the motors, four of those were pugged into the correct ESC control wire (white) to control each specific motor. This also is critical to be done correctly, if two control wires to two ESC’s were swapped the model would go out of control.

Fortunately, the APM Mission Planner software helps tremendously with the set-up. The model type is chosen in this case an “X” Quadcopter. Calibration begins, first with the accelerometers. The model is placed flat upright, a key pressed than click a button, then rotate the model to the left, right, nose down nose up, and upside down, clicking at each step, to calibrate it. This is how the F.C. knows which way is up and can detect tilt errors (along with the use of accelerometers and massively impressive computing to keep track of its orientation once in the air). Compass gets calibrated. And then the R/C channels are calibrated. The transmitter sticks are moved to their full extremes, and also toggle switches being used (I’m using switches on channels 5 and 6). The software takes not of the responses. This is when the correct channel response is confirmed, both for the axis (pitch on transmitter is pitch onscreen) and the direction (pitch-up is confirmed as pitch up and not down).

Flight Modes are also set up, selected by toggle switches. I did not have the GPS module in it at first, so it flew mostly on “Stabilize” mode (#6). With GPS installed, then it could do RTL (Return To Launch, landing where it was when powered up) or other things like “Loiter” (hover over the same spot regardless of wind).

For those who are really interested in the nitty-gritty of the software set-up and settings for the Flight Controller using APM Mission Planner, check out this video:
[video=youtube;30cCs4aHdB0]https://www.youtube.com/watch?v=30cCs4aHdB0[/video]

With the APM F.C. programmed, it was almost ready for flight. But the ESC’s needed to be calibrated. The computer cable was disconnected, and the flight battery plugged in to power everything. The process is documented in this youtube video link:

[video=youtube;gYoknRObfOg]https://www.youtube.com/watch?v=gYoknRObfOg[/video]

It was during the ESC calibration that I really determined which way the motors were turning. Then I swapped 2 of the 3 wires on two ESC’s to make them motors spin the other way as needed.

For power, I decided to use a single 3000 mAh 3S (3 cells in series, 11.1V) battery so I could evaluate the flight durations. Since the model would weigh more after adding various other parts, and might also weigh more if I used a bigger battery later, I also added a second 3000 mAh battery for ballast. This brought the flying weight up to about 1400 grams…. which was my guesstimate for the model. I added a mount to hold the batteries and to also hold two velcro straps to hold the batteries.

And I attached the propellers, 10” diameter by 4.5” pitch, making sure the CW and CCW props were on the right motor locations.

 

[georgegassaway]

So, it was ready to fly…… but had no legs. But it did not need “legs” to fly, just something to help keep it from falling over. So, I drilled some holes into two 3/8” dowels and bolted those in place on the bottom of the Descent Stage.

Took it outside, set it up. Turned on the transmitter. Plugged in the battery in the model. After initialization of the APM F.C., it was ready.

Would it work or not?

Moved the rudder stick full right to arm the motors. After a few seconds the props began to spin slowly, showing it was armed and ready to throttle up.

And then…. THIS!

YES! It took off with plenty of thrust, I had to throttle back. It even climbed a bit at half-throttle. So, it’s NOT an overweight struggling to fly model. The research had paid off. Did a little adjustment to the trims. There was a little bit of wind but not much, good conditions to fly in. It flew nicely. And at one point I realized there was something in the sky, Oh, yeah, THAT!

I got that image from a GoPro I was wearing on my cycling helmet. I wish I could have had someone else use a telephoto camera and I could have arranged the model position and camera aspect to make the moon look larger. Well, i’ll try to get a better moon shot later when the model is completed and looks realistic.

Here’s the first flight video:

[video=youtube;0jPOZ84QZB0]https://www.youtube.com/watch?v=0jPOZ84QZB0[/video]


It flew for over 9 minutes! Well, I made an error with the audio voltage alarm so it did to go off to let me know the voltage had gotten low. But I had realized it had been in the air a long time and I was not letting it get very high. So when it required more throttle to hover, that was a sign the voltage was getting critical so I started to land right about when the voltage got too low to keep it hovering anyway.

So, VERY happy it flew successfully, and pretty well. It was a bit twitchy, I need to do some tweaks to fly smoother. I plugged in the second battery and flew some more. Another good flight, and that is the one where I got the screenshot with the moon.

And so, as of the moment it took off successfully, AFAIK, it became “the best RC Lunar Module multicopter in the world”. But it’ll end up looking better over time ……

 

[georgegassaway]

So, that was Friday, November 4th. That night, I made up some crude temporary legs, so it would look a bit more realistic for flying on Saturday. The structural attachment was much the way I intend for the real legs, of course the crude “V” lower struts will be the complex X and V struts on the final model. The legs are 1/4” wood dowels, intended to break first in a hard landing or crash so as to absorb impact energy and hopefully not cause worse damage to the rest of the leg structure. Also the horizontal lower struts were attached in 2 places by 4-40 Nylon screws, intended to snap the screws rather than damage the lower bulkhead ring. Also seen in the photo below, a yardstick to indicate size, and the props were changed from Gemfan props to APC props which are better quality and perform better (same size, 10 x 4.5)

Also, I added the GPS receiver. It needs to be as far as practical away from EMI sources as well as anything that can can affect the compass sensor, so it was attached to a tower structure above the APM Flight controller. The red/blue parts of the tower came from Nylon spacers that are used a lot for kit and homebuilt multicopters. Also near the bottom you can see the APM F.C. on its vibration mount and some of the wiring running thru holes inside the Decent Stage.

And so, flight #3:

And video:

[video=youtube;6B3C26J1qN4]https://www.youtube.com/watch?v=6B3C26J1qN4[/video]

Actually that was more like battery session #3, there were a coulee of shorter flights. And the voltage alarm was working, but it was set too high. So I kept flying but near the end kept it close and low in case the voltage dropped too low, which it did at the end. So I need to dial in the voltage alarm setting for something I can trust and use properly.

So, some landings were not so great. I was trying too often to land back on the take-off square and sometimes cut the throttle a bit too much when I wanted it to land on it. And I need to get in some more practice…. with another model. Anyway, the photo below shows that one of the 1/4” dowel legs broke, and it was pretty much like I hoped it would break in a hard landing. Also, the 1/32” plywood discs I used for the landing pads (feet), those broke off too easily. On the final model i’ll use some custom cast foot pads, and know now where I should reinforce the joint so that the 1/4” dowel legs will fit inside and not break off too easily. Also, I will make up spare legs with pads, so if a pad breaks off or a leg breaks at a flying session where I want to keep flying, I can just swap out a leg.

Here is video of a short flight which demonstrates a nice soft landing. I had accidentally cracked the joint of a landing pad, so it was loose. Not long after liftoff I could see that pad wobbling. And then it fell off. So I wanted to land soon so I could have the model near to where the pad landed, to make it easier to find.

[video=youtube;li5SY3_i6YA]https://www.youtube.com/watch?v=li5SY3_i6YA[/video]

 

[georgegassaway]

And test flight 4, or again rather the 2nd flight on the 4th battery used in testing (using the 2nd battery on Saturday), after the short hop above.

[video=youtube;ahv_7IUDFyg]https://www.youtube.com/watch?v=ahv_7IUDFyg[/video]

That video ends a bit after 7 minutes, before it landed.

There is a reason why it ended up short. Because I edited it.

Sigh......

 

[georgegassaway]

HOUSTON, WE'VE HAD A PROBLEM. Main Bus B Brain Fart.

Because I did something really stupid. It hit a tree (not a close one) and fell 30 feet to the ground. It can be fixed, but it’s going to take some time. Visually the worst are many of the 1/16” balsa side pieces broken loose. But those are not that hard to fix, and the sides will later be covered over by gold foil and/or gold mylar, or a black covering material to match the appearance of the real LM’s.

The most serious damage is that one of the 10mm graphite arms split into 4 pieces, and is cracked near the edge of the Descent Stage. No way to repair it to be study enough, it has to be replaced. And that’s not going to be easy. But it is possible, and I do have two spares.

A good thing is that as far as the leg structures go, the upper portions using 1/16” music wire are intact. Of the lower horizontal “V” struts, some damage to those, notable for planning how to do the final realistic version of those struts. One of them did have a 4-40 nylon screw shear as hoped. One of the white plastic tubes that the 1/4” dowel legs slide inside of, was bent. But those are easily replaced, not glued in.

It was very good to see that the two heavy batteries did not budge! A concern i had thought of before was if during a crash a battery came loose and cannon-balled itself into the structure. But both stayed together and did not shift. Of course later there will only be one battery, I was using two in part to add ballast to simulate future assemblies not on the model yet (mostly mass of Ascent stage and legs)

 

[georgegassaway]

So, where does this leave things? For the moment….. leaving the model to sit awhile as I catch up on other things, and just do not want to deal with fixing it for a few days. But I will get back to it, and begin repairs.

When flying it Saturday, I got GPS lock, but it did NOT get into any of the GPS modes when I was flying it. I realized later when I had set the mode switch positions….. that I had not pressed the “Save” button to confirm those changes, so it was in “Stabilize” mode as the original default the whole time. If I had it set correctly, then I could have engaged “Loiter” mode so it would hover over the same spot, so I could do some landings more accurately. Also could have used RTL to automatically land close to the take-off square, although the accuracy it not dead-on, typical RTL landings are about 1-2 meters off.

So besides repairs, I have more tweaking to do with model control. Not just modes, but settings such as tuning the PID response. Also I saw this afternoon when I had it hooked up to APM Mission Planner for some screenshots, that for response to the transmitter it was set for maximum, “twitchy”, which is great for aerobatics and crazy Racing Drone flying, but not what I want. So I can dial that back. And do some other adjustments.

Also I need to get my old 250 Quad bak in flying shape. I had removed its APM F.C. and GPS module. But I have spares for those, so I ought to get that flying again so that I can test and confirm the fight adjustments mentioned above. Although being a smaller and lighter model some settings will not respond the same, as PID’s are very dependent on model mass, size, and moment of inertia. Still, it can help out with a lot of that. And also I can just simply practice flying it better, I have not done a lot of Quad flying for over a year.

So, after repairs to the Lunar Quadule, I’ll do some more testing to complete Phase 1. While the 3000 mAh Lipo pack flew it for a good amount of time, I may very well get a couple of 4000 mAh Lipo packs to fly it with. Due to the mass increase, the flight time would not be 33% longer (4000/3000), it would be about 25% longer. Those I may not get any new batteries for some time.

Beyond that, to begin on Phase 2. Now that I know it works, to finalize the leg design, and begin assembly of a master for the lower horizontal strut assembly. That master will be used for creating a 2-piece mold for casting copies of it. The castings would be too fragile by themselves, though. But the master and mold will be designed to allow reinforcement of music wire to be embedded inside, sort of like steel reinforcing rods in concrete. I’ve used that kind of trick before with weak cast parts that needed strength in key areas (Shuttle model ET aft strut assembly, and forward BiPod strut). The struts will not be shaped like the real ones, as again this is fora flying model, not a display model. The drawing below is to show the basic geometry, on the right you can see the merged "X and V" shape of the lower horizontal struts after deployment (no the legs will not be folding/deploying!). While currently on the model I just have crude "V" struts to serve the same purpose structurally.

Scale-up the patterns for a cardboard Lunar Module Ascent Stage, and arrange them to fit onto paper I can print out and take to get printed onto thick poster paper. And build the Ascent Stage. It won’t have details like RCS thrusters until Stage 3. But getting that Ascent Stage built out of poster paper will be a big milestone.

Make up the base assembly, or “hatch”, to cover the big opening in the bottom. I likely will build a balsa pattern for that and vacuum form it. Also may end up vacuum forming a Descent engine to mount into it.

Electronically, add a piezo beeper plus a red and a blue LED in some out-of-the-way location, connected to the analog outputs of the APM F.C. The F.C. board has tiny LED chip lights, red to indicate motor arming status, and blue to indicate GPS lock, which I will not be able to see with the complete Ascent Stage in place. But the F.C. has analog outputs that among other things can be used to light up external LED’s to show that status. Also, a piezo beeper can be added to indicate the arming status of the motors, so I can hear when they are armed, not needing to look for the red LED. Also will add an actual on-off switch, it is very inconvenient to plug and unplug the battery shortly before takeoff and shortly afterwards, especially when that has to be done upside-down to access the bottom opening.

So by the end of Stage 2, it should look the right shape, with all the major parts.

Stage 3, add some (but not LOTS) of details, such as RCS thrusters, ladder, and “porch”. Gold Foil and Gold Mylar all over the place. Some black covering for the areas that were colored black. Add split tubing over the1/16” music wire upper struts, to make them look more accurate, and cover or paint those black. Add “UNITED STATES” markings.

And add at least one cosmetic/visual special feature (there will be no “flight” special feature involving the ignition of anything or the separation of anything. This goes up on rotor power and comes down on rotor power, staying in one piece, unless something goes wrong).

Timeline for this? I do not have a hard deadline for it to be completed, but I expect the next 3 months or so. I will post some further updates on the repairs, other testing, and when Phase 2 really begins.

Again, I suggest trimming any quotes in replies, to quote the relevant parts you are referring to.

(Apollo-9 “Spider”, first manned LM flight, tested in Earth Orbit. Hope nobody had to read this text to realize that was not the moon, or the moon looks "odd" in that pic….but…&#8230.

 

[burkefj]

George I'm not normally a quad fan but that was pretty neat, I think once you get the foil on it will look really great in the air.

Frank

 

[rstaff3]

Very cool idea. Can't wait to see it all finished up.

 

[MaxQ]

Vey cool...very very cool!

 

[grouch]

That's awesome George.

 

[neil_w]

Very cool George. Now all it needs is a bunch of Der Red Max decals and it'll be perfect.

(I promise I will not retread that joke again, at least until the next time).

 

[cerving]

OK, next step is to build a Saturn V, lift it up to a decent altitude, eject it, and have it land at a designated location. Piece of cake.

 

[tab28682]

You really need to plan ahead to do a little unannounced demo at a larger launch.

Warn the event management, get permission, get a GPS fix of a good landing spot a safe distance out from the LCO table.

Then launch from a small distance away, unnoticed, hang out nearby as covertly as possible with your TX and have it autonomously arrive at the launch to the amazement of all present. Take over and land.

Might take some planning for a safe arrival pattern so as to not overfly anyone.

Would be pretty awesome....

 

[georgegassaway]

Wow, that is a pretty neat idea. And indeed if I did it I’d coordinate with the launch manager and RSO.

What I would like to do for flying demos, is to have a means for playing an audio file on my iPod thru some speaker system at the launch. And have a custom edited audio file set up which would mostly be select moments from the landing of Eagle on Apollo 11.

This site is the best multimedia presentation I’ve seen (and heard) documenting the audio (plus scrolling text transcripts), who is speaking, flight data, and film showing the lunar surface outside the window, among other things.

https://www.firstmenonthemoon.com/

With practice, I’d try to land at the moment of the “Contact light” call, and disarm and stop the spinning props when they call engine stop. Run that audio until at least past “The Eagle has landed”.

For the front end of the audio, for takeoff, use some of the audio from the Saturn-V launch, starting up the motors at “ignition”, but not throttle up to take off until “zero”, which the audio would then shortly say “Liftoff”. OK, not quite right for a Lunar Module but effective.

Also perhaps some music, perhaps from “the Right Stuff” and/or Apollo 13. And possibly “Also Sprach Zarathustra”, which most people call “the theme music form the movie 2001 a Space Odyssey”.

I have plans to make a very specially edited video, which would also make use of much of the audio. Also hopefully some nice video scenes beyond just taking off or landing as I’ve done so far. I’d love to use a 2nd Quadcopter with GoPro to get video of the lunar module with both at about the same altitude.

Have some other neat things in mind, for some literally the planets have to align. Well…… moon phases, the of day it is visible, and weather/sky. This one accidentally happened, but the moon is so small and a crescent. It would require a telephoto camera at a distance and the Lunar Quadule at just the right spot to get a good video of the model and moon where the moon would look a lot bigger than this.

If the camera was zoomed in tight on the moon, and I could fly the model in front of the moon, for something sort of like this but the model at about the size of the moon....not the small relative size of Elliot, ET, and bike.

 

[georgegassaway]

So, I got my old 250 Quad back in flying shape and did a bit of flying. It was pretty windy though, and I didn't have the GPS enabled, so I wasn't able to practice the style of flying the Lunar Module.

But I did attempt to get a "selfie" with the moon. Because the sky was clear again today and the moon was out at JUST the right time and place try to do photography/video with the model flying past the moon. I set up a camera on a tripod and zoomed in as close as I could.

It was tricky to try to get into the frame of view though, the wind really didn't allow me to set it up to hover, and trying to get into frame did not work well for trying to get it right into an exact spot to try to hover from far away (the model was at least 200 feet away, maybe 300). I posted a short clip of the video, slowed down by 2X. The audio includes the low voltage alarm beeping in slow motion.

I shot it at close to sunset. I have some lights on that model, allowing flying it at night. Two very bright white LED's on the front, visible when flying towards me, sot of like headlights. And a strip of green LED's on both sides of a "tail boom" sticking out the back, which is a great indicator of which way it is pointing and the primary light source (only time it is not visible is when looking nose-on, when the bright white LED's are visible).

Anyway, once phase 3 is completed, I hope I can get some good video of the Lunar Module flying with the moon in the background, but much more slowly (and without motion blur) than this. But as I said before, the planets literally do need to align to try this, plus Mother Nature cooperating with clear skies.

[video=youtube;CgW-agZgffM]https://www.youtube.com/watch?v=CgW-agZgffM[/video]

Update: MUCH better results Friday.

Had the GPS working well for Loiter mode, and the wind was pretty low. I do need to "tune" this model, its using a new Mini-APM controller that does not have the PID values that the old one had. In any case, also was able to get in some good practice at the kind of landings I want to to with the Lunar Module Quad.

[video=youtube;Svk2hOs-n84]https://www.youtube.com/watch?v=Svk2hOs-n84[/video]

 

[georgegassaway]

Have begun repairs. Trickiest moments were removing the 10mm square graphite tubing which were CA'ed in several spots.

Photo below shows the removed arms, and the damaged Descent stage. I do have the missing 1/16" balsa pieces, leaving those off until the new 10mm arm plus undamaged arm are reinstalled. Will also be doing some reinforcement of the structure between the upper and lower bulkheads.

Expect to have it ready to fly by Monday. Looks like the wind may be low enough to resume test flying it (wind kicking up bad again Saturday and Sunday). The kind of crazy gusty wind I flew the 250 Quad in for the last two days, no way would I fly the Lunar Module in wind like that, too risky (lots of side surface area). The way the moon rises later and later day by day, near sunset, the moon may not be high enough to get a flyby video as in the above post. Of course what I really want for a highly edited special video is to do that when the LM is complete and looking realistic.

I'll take more photos of some of the repairs and post them.

FWIW - here's a photo of my 250 Quad. GPS receiver is on top of the "tower" (Similar thing as the Lunar Module Decent stage has mounted on top, above the Flight Controller, inside of the Ascent Stage). Forward flight is in the orange direction, rear of model green (tailboom in back is for a green LED strip, in daylight it helps show which way it is pointing and at night it is the primary light source, usually). Low Voltage Alarm secured to the tailboom with clear heatshrink. Pieces of foam "pool noodles" cut up to use as landing leg cushions, zip tied in place.

So, that 250 quad is sort of like my way of having a Lunar Landing Research Vehicle (LLRV) to practice with. Except it handles wind way better than the real thing did.

[video=youtube;091ezcY-mkU]https://www.youtube.com/watch?v=091ezcY-mkU[/video]

 

[georgegassaway]

Worked on repairing the model. Was able to pry loose the 10mm square graphite tubing for the arms. Replaced the split arm with a new one, after drilling various holes as the original had. Then slid both arms into place and glued them back.

The Descent Stage needed some more structural improvement. So I cut out some 1/8” balsa “V” braces to glue inside. In the photo below, one is in place, the other three are waiting to be glued in.

The various 1/16” balsa side pieces were later glued back in place.

After being repaired, the anti vibration mount was reinstalled with the Mini-APM Flight controller

Took the opportunity to add a bit more to the wiring harness. Two most significant were an on-off switch (3PDT gang-wired) seen near the lower left with the Deans “T” power plug, and an extension (far right) for connecting the Voltage Monitor to the battery pack’s monitor connector, in a manner that allows the battery monitor to be “off” when the switch is off.

As seen in the middle, above, is the battery mount, plywood with some yellow vinyl tape and Velcro to help keep the battery from sliding. Not shown are the two velcro straps that wrap around that mount and the battery. Underneath the plywood battery mount is a holder for the Power Distribution Board for the Wiring Harness.

Below, close-up of the Flight Controller with case added. Also inside of the case is a piece of “breathable” foam (like the kind used for foam brushes), over the barometric sensor to help prevent any airflow for affecting the sensor. This also shows a couple of the wiring cable sets plugged in and running inside. One of those, the model uses a lot of the wires (RC input uses all but two, the other only uses 6 (Motor output 1-4 and Ground and +5V). A couple of other are plugged in later, one for the GPS module, and a really big one for analog outputs when I'm using for three functions (Arming LED, GPS Lock LED, and status beeper)

 

[georgegassaway]

So, new legs, ready to fly on an afternoon with not much wind.

But then, a big problem. It accidentally took off in Return To Home mode (thought it was in Loiter mode), RTH first climbs to a minimum altitude before beginning descent, so it was climbing quickly, and I did not realize at the time why (I still do not understand why it took off so fast in RTH mode, I did not think that takeoff was possible in RTH). I toggled a switch to go to regular stabilize mode, and was shocked to hear the motors stop! It hit the ground before I could get it to full power. I do not think I have had that happen with anything before, not a fast surprise takeoff in RTH mode. I’m going to run some tests with my 250 Quad to see if I can duplicate that problem...... from a higher altitude. But also I’m going to test the 250 quad to totally replace Stabilize mode with Alt Hold mode, which I think will prevent this from happening in any case (Alt hold never totally shuts off the throttle in the air, at lowest throttle on the transmitter stick it comes down at a pre-set maximum descent rate. Well, I am going to test with the 250 Quad to confirm for sure)

So, after the crash, I had this to deal with.

Other than losing the legs, some other leg related damage (crude lower "V" struts), and some damage to the Descent stage….. it was still flyable. Well, at first one of the motors did not run, but a signal control wire to the ESC had come unplugged in the crash, after plugging it back in it was flyable. Another factor was that the weather for later would be too bad to fly, very windy, for at least the next 10 days. And on Friday I’m having a hernia operation which will put a hold on some things for awhile.

So while the winds were low, I decided to fly the damaged model to learn what else I could and shoot some more video. I had wanted to use “Autotune” to tune the PID settings, but when I activated the switch for Autotune, it did not do it, so I have to check out why it didn’t.

Unfortunately my helmet-mounted GoPro was aimed a bit high, so for some of the landings I didn’t get the landing in view, or not all of it. A fixed camera on a tripod helped with some of that but not as good as the helmet camera could have been. Although I did get one very nice close-up sequence on the fixed camera, after this begins.

One upgrade I added was a white flashing LED on the front of the dummy Ascent Stage. The real LM had a strobe light for use during rendezvous. The white LED is driven by a flashing LED in series, and as such it flashes quicker and is on for longer than a real strobe would be. I hope to add a tiny Arduino at some point, to do 3 special things for the model. And one of those would be to flash the white LED at a slower rate and for a shorter time, to be a bit more like a strobe. Now, in real life, they would not have had the strobe on for the lunar landings, but in some sky lighting conditions that flashing white light can be helpful in seeing the orientation. Also the light helps to show / remind that the model is powered up.

So, here is an edited video from the day’s flying. First scene is the crash, and yes the audio is “mysteriously” missing. Then a nice slow approach and landing, though the helmet camera did not get the actual touchdown. Third clip is a slow descent then holding at about 10-12 feet up. On the Voltage Monitor which has red LED digits, I could see and read the voltages of the three cells of the battery as the data cycled thru, pretty neat (can’t read them in the video). Final scene is from the fixed camera, a good view of control.

[video=youtube;dWxrIkJqrAs]https://www.youtube.com/watch?v=dWxrIkJqrAs[/video]

The slight wobbling, I think that tuning will take care of most of that, once I get the Autotune process working and have a calm enough day to run the autotune sequence. In Autotune (nothing to do with music!), the model hovers and drifts with the wind as the F.C. puts the model into a sequence of fast pitch and yaw inputs, as the accelerometer sensors determine the response of the model and therefore to tune it automatically for the best P.I.D values for that model. Requires several minutes to complete. Can always take over piloting to bring it back but it’s just a bad thing to try on a windy day when the autotune process gets interrupted to fly back, re-engage, drift away quickly again, over and over. And it might have to land due to low voltage before the process is complete. Today was a really good day for trying to get it Autotuned, so a real missed opportunity that it didn’t activate.

FWIW - link to Autotuning in APM
https://ardupilot.org/copter/docs/autotune.html

For the next few days…… I’ll do some testing with the 250 quad, as it can handle wind pretty well and also can take a lot of “hard landing” abuse if something goes wrong. I do not think that I’ll fix the crude legs again, I may go ahead and assemble the master part for the lower “X-V” horizontal braces, and make a 2-piece RTV mold so I can start casting those and do some test flying later using the actual intended parts and see how they work out. BTW - for sure, if the parts were 100% resin, they would snap very easily. But I will be inserting music wire of different diameters, with some bends, and some bent “hooks” on the ends, in various locations, into one of the mold halves before casting, to reinforce everything. They will be a bit more like music wire struts with resin cast around them to give the desired thick round tube appearance. Have done something like this before, though not this complex.

So I’m sort of in a fuzzy area of flight testing for Phase 1 but working up some “Phase 2” things that I realize need to be flight tested to see how well they work out. If the XV lower struts prove to be too fragile for a given landing speed that was not THAT hard, I want to know that long before going much farther.

 

[TangoJuliet]

I dread seeing your "accidents" George, but your methodical approach sets a great example for other scratch builders to follow.

The one thing that I don't like about electric R/C is the frequency with which a "power on" situation becomes a "full power on" situation. I've known many R/C aircraft modelers over the years who've been bitten by props during the battery connection. Be very careful with those four whizzing props. You're lucky it didn't come flying at you.

 

[georgegassaway]

The one thing that I don't like about electric R/C is the frequency with which a "power on" situation becomes a "full power on" situation. I've known many R/C aircraft modelers over the years who've been bitten by props during the battery connection. Be very careful with those four whizzing props. You're lucky it didn't come flying at you.

Hmmm, I've never had an electric plane that could power up on battery connection alone. Although I could see how some poorly designed/cheap ESC might cut corners, or if the model had a direct On/Off engine power set-up (like a servo pressing a switch) instead of real throttle. OK, some really small cheap ones (the kind that use a single 3.7v 70-80 mAh battery) might also do that, but they could not cause significant injury.

The ones have used, throttle must be full down for the ESC to arm after power-up. Now if I bump the throttle stick after it has armed, yes, that's a problem. I've been fortunate I haven't had that happen while my hand or anything else was in the way.

Except for the tiny indoor type multicopters, almost all of the multicopter flight controllers require a special arming sequence. For the Arducopter software, the throttle has to be in low, AND the rudder stick moved full right for 5 seconds before it arms. Then on arming, the props begin to rotate at a very low RPM, too low to cause injury, but definitely proof that it is armed and ready for takeoff. I will review my video from Monday later, and see if there is a good clip showing the arming process. If so, I'll post it. If not, perhaps shoot some new video just to show that, arming only.

 

[georgegassaway]

I made a video of the arming process, and also the wind was not too bad so I flew the LM some more, and did the AutoTune sequence.

I did not document this before, during the recent repairs I temporarily added a LED and Piezo beeper module to indicate status (mounted behind the Ascent Stage, on top of the Descent Stage). These are great things to have on any multicopter that uses a F.C. capable of controlling the LED's and beeper for this purpose. I’ll replace this with much smaller LED’s in some realistic locations (spoilers), and put a separate piezo beeper inside. After startup, when the Flight Computer has completed initialization, the left LED blinks red to indicate the F.C. is operating but that the motors are not armed. When armed, the LED is solid red. The right LED blinks blue to indicate that there is no GPS lock. When it has GPS lock, it is solid blue. The piezo beeper give audio status, beeping during arming, during disarming, and to indicate other things such as loss of GPS lock. Signals to these are sent via analog outputs 4, 5, and 6 from the F.C.

As for arming safety, my transmitter will not even complete boot up if the throttle stick is not full low (many mid-level and above Tx's are that way for "powered" models, but not all Tx's). It will beep and show a message that throttle is high, I have to move throttle low for it to complete its boot-up.

The video shows the arming process, motors spinning, shutdown, and disarming. Also, it has some clips from Tuesday’s flying to set AutoTune. AutoTune does automated inflight pitch and yaw changes, the accelerometers detect the responses and work up the best P.I.D. values for each axis. The P.I.D. values are different for models due to the moment of inertia of the model and how the ESC’s, motors, and chosen props affect the control responses.

So, I flew the model a couple of times, with no issues. Replaced “Stabilize” mode with “Alt Hold” mode as the main mode to fly in.

Key feature of Alt Hold is the way it responds to the throttle stick, because it it not actually controlling “throttle”. It is controlling Altitude, using the barometric/altimeter sensor. Middle stick is hover. There is a bit of a deadband for middle throttle so say 45 to 55 percent stick position can be the same (user programmable). To go up, move the stick up, the more above middle then the faster it climbs, but there is a user programable max climb rate. And to descend, move the stick below middle, the closer to minimum throttle the fast it descends (and the maximum descent rate is also programmable).

How is that different from regular throttle? Aside from hover at 50%, let's say I go to 30% throttle. In Alt Hold mode, it will descend at a fixed descent rate, no faster, no slower, unless I change the throttle stick. In modes that use regular throttle, 30% throttle means about 30-40% less thrust than at hover, so it not only descends, but it descends faster and faster and faster. If I went to 50% throttle after it was falling at say 30 mph, it is going to still fall at 30 mph due to momentum, just not accelerate down to fall faster anymore. Except due to air drag it would slow down some. So the only way to try to make it "hover' soon after descending at 30mph is to throttle up way above 50%, like 70 to even 100%, then as it starts to be close to zero descent rate (nearly hovering), adjust the throttle stick to about 50%, making minor adjustments (fortunately this model does hover at about 50%. Again there are programming options to help tweak that so that 50% throttle stick can produce hover-level thrust, even if it was heavy and needed say 70% throttle to hover).

So, I really should have been flying it in Alt Hold mode to begin with, that would have prevented Monday’s accident. It’s just that I usually fly my 250 Quad in Stabilize mode since that is more maneuverable/faster than Alt Hold mode, and when I do want to fly more precisely I’ll go to Loiter mode, which behaves like Alt Hold mode plus GPS making it hold position and not drift. And I should be flying the LM more gently than a 250 quad type anyway.

Yeah, I know, a lot to understand in the above. Shows that even "Rocket Science" isn't enough with things like this. It's Aerospace science, adjusting program settings (via some excellent software), physics, and more...... including lots of learning how to get the most out of the potential capabilities. Although on the other hand, a lot of the mid to upper RTF commercially made copters have a lot of this stuff in it, to the point that the user (Too many are "users", too few are "Pilots") does not know any of the distinctions and does not have to set anything. Most copters that are mostly intended for shooting video use some version of Alt Hold rather than classic throttle). In any case I hope a few reading this got something out of the explanation about the difference and how Alt Hold does its thing.

So, the video below also shows a flight with arming, and takeoff, then mid-flight with AutoTune beginning to activate, to the end. It drifts downwind so I have to interrupt to fly it back upwind a few times, before it finally completes AutoTune. And, a landing with disarming.

[video=youtube;sWtjbbrsxH8]https://www.youtube.com/watch?v=sWtjbbrsxH8[/video]

NOTE - In the first clip, when the motors/props are armed and turning slowly..... they were turning FASTER than they appear to in the videos. This is due to a stroboscopic effect of RPM vs camera shutter frame rate. But in any case the rpm and motor torque was low enough that if someone put their hand in the way, it would feel more like a light to medium whack with a ruler than cause serious injury. Not a good idea to be hit regardless.

 

[georgegassaway]

Forgot to post this photo. Yes, always take photos of models BEFORE you fly them.

And this fixed camera screenshot of Mondays flying, after the "Hard Landing" that caused some damage but was still flyable enough.

If you wondered how I got video that followed the model while also piloting it, or why the heck am I wearing that bicycle helmet, I have a GoPro mount on it and made good use of it. Link to a bike ride video using that Helmet/GoPro:

https://tinyurl.com/puk3ra6

 

[georgegassaway]

Had been checking around for various kinds of gold foil and such to get to use with the model later.

And then ......

Uh, yeah. Almost like a Faraday cage.

The foil, aluminized mylar, and all other metallic coverings are going to tend to interfere/block the signals to the receiver inside of the Descent Stage. So now I’ll need to plan to have the receiver inside of the Ascent Stage (which will not have much metallic coverings/foil, which can be worked around). The receiver I’m using is a very good one with two antennas, Spectrum 6255, meant for use in planes/gliders with carbon fuselages. Used those a lot the last few years, rock solid.

The one complication about having the receiver inside the Ascent Stage is that it may have an effect on the compass on the GPS module.

But at least I realized this now, before doing any flying with foil on the model, so I can plan around it. And it won’t be getting foil or mylar for some time yet, that will be among the last things done with it.

Another good thing that is mostly unique to multicopters, or at least more advanced controllers like Arducopter, is that if the receiver loses signal, the Flight Controller will keep it flying level. After 5 seconds of no radio response, I have it set to default to doing a RTL/RTH (Return To Launch/Home), fly itself back to where it took off from and land. That’s one reason why the GPS compass needs to be working well.

 

[KenECoyote]

Fantastic build Frank! :clap:

Sorry to read about the issues, but with RC, you have to accept that you will crash a LOT more than rockets. I've been into RC helis since reading about the GMP Cricket in the 70's, but sometimes I'm still crashing them like it was yesterday...:wink:.

Regarding the Lunar Module, I've been wanting to build a rocket version of this for a while now and figured out some possible ways; however that will take a LOT of mindsimming and so it's a waaaay back burner project for me. Good to see someone else is also interested in seeing this fly!

 

[georgegassaway]

Recovering OK from inguinal hernia surgery last Friday. Finally over the hump for the pain, had gone from bee-sting sharpness at times when I moved, to a more dull feeling (no problem sitting or laying down). But definitely better today.

Meantime, did some more work on the model. Started on the lower leg strut assembly. Drew up patterns, used double-sided tape to help hold parts as the liquid cement dries. When done, the whole lower “XV” strut assembly master part will be used for creating a split 2-piece RTV mold, to cast copies. Music wire of different diameters will be bent and cut to shape to use inside, sort of like re-bar in concrete, because the resin itself would be too fragile.

Also ordered an ultrasonic "sonar" module. I'll test it out to see how well it may assist with vertical descent during landings. This one should be effective for about the last 10 feet, or whatever altitude it gets a good sonar response. Above that it'll use the barometer. The real LM had a landing radar to determine altitude above the surface, so the ultrasonic sonar would be doing a similar thing, audio rather than RF pings.

https://ardupilot.org/copter/docs/sonar.html

LATE UPDATE: Completed most of the horizontal strut assembly. Allowing to dry overnight. Will later trim part of it to glue on the mounting lugs that will eventually attach the struts underneath the Descent Stage, using 4-40 nylon screws (mounting hole locations at the center of the yellow squares that are mostly hidden right now). So after adding those mounting lugs to this, and adding some glue fillets to the underside of several joints, let dry another day before using it to make up the first half of the RTV mold. Might have the first 4 sets of "new legs" by Sunday, the cast copies of this assembly, and other legs. Won't have the shallow dished type landing pads for awhile.

The two little wood dowel-like things at the bottom are dowels used as spacers, they are not glued to anything.

 

[georgegassaway]

Wednesday afternoon update: Completed the assembly, added the attachment lugs.

BTW - the diameter of the "V" strut that attaches to short 3/8" tubing that the leg will run thru, the larger diameter near that 3/8" tube is misleading. On the real LM, it is a larger diameter at the other end, near the "X" brace. But I'll simulate that larger diameter later with a soda straw slipped over, saving some weight over casting it much larger there. There's going to be other things/foils covering those parts anyway.

I had bought a 16ounce block of clay at Michaels ($40% off) to use for making the mold. Forgot it in the car, had to warm it up for awhile to make it soft enough to press out wider to use for creating the mold. Carefully pressed the strut assembly into the clay, ideally half-way. Had to press some of the clay "up" cover over parts of the wood lug pieces that had been added, as those are not flat so they stuck up from the clay at a shallow angle. Used the rounded end of a brush handle to create a number of shallow holes into the clay. Those are keying holes that will be used to align both halves of the mold after the 2-piece mold is complete. Have a couple of 1/8" rods sticking upwards, for later use of removable rods to cast the mounting holes into the lugs.

Glued together pieces of basswood to use as a fence to contain the RTV.

In the middle photo, a brush was used to brush the RTV along all of the interfaces between the clay and the strut assembly. Also to fill in the keying holes in the clay, otherwise there might be air voids.

In the right photo, the rest of the RTV has been poured. It'll take at least 24 hours to cure properly, so I might pour the 2nd half Thursday night.