One confession - when I was screwing this together I just grabbed whatever screw I could find so some of them may have been a bit too big. I’ve since acquired the correct screws.
I may try to print these replacements with less infill - just because I’m impatient. I’m also hoping the PETG is a bit stronger than PLA but without the hassles of ABS. This is my first 3d printer (a Sovol SV01) and I’m having a blast. I know a lot of people spend hours calibrating and modifying things but so far everything I’ve printed has been functional and while maybe not ‘perfect’ it’s good enough.
Thanks for pioneering the way :). I know nothing about 3d printing other than what I’ve picked up from @cglusky through my devFrame experiments, ordering stuff online. Just to note all the stuff I’ve ordered online has been ABS and 20% infill, and it’s incredibly strong. I actually find the low infill provides like a ‘crush’ structure that deforms instead of breaks on impact.
I think it has more to do with properties of PLA vs ABS then infill. PLA is just more brittle. But it’s also easier to print. So if you had a bunch of spares you could keep flying. Trouble occurs when failures happen in flight. So there is additional risk. I do not have experience with PETG but have a friend who has been running it on prototype ROVs with success.
I cheat my solid parts with 90% infill. 50% infill on other parts could be lower but would need much more testing. Just easier and safer to error towards more infill.
Looks like your arm tube clamps need to change print orientation. Z axis of print should follow arm tube. Helps reduce cracking when applying clamping force.
If you print the arm tube clamps with right z orientation and keep plenty of motor mount and landing gear spares PLA could work for short term prototypes? Even if that is true I am not sure it is something we can recommend in docs just due to potential failure of motor mount in flight.
I had some hard time assembling arms tube mounts and box ends together, my tapping screws do not tighten well. What do you think about changing the tapping screws to standard M3 screws? It would require some design modifications but allow to assemble and disassemble the frame as many times as you want (could be useful to dismount arms for rover as well).
@cglusky Perfect timing - was just going to re-print. Rotated this in Cura and will see how it goes!
I just used PLA because I was impatient and I had a lot of it available. And I have been banging on this thing pretty good so overall I’d re-use PLA - with the warning that you do frequent inspections pre-flight, post crash.
Trying PETG now and hopefully that will be a nice balance between ABS and PLA from what I’ve read.
We started with M3 and threaded inserts. Decided that being able to source fasteners locally and easily was more important than handling a large number of fastener cycles.
Could you explain in more details what kind of issues you are having with using #6 screws? I am using pointed pan heads with no issues at all. So I am wondering if a true self taping point creates issues. For soft material like wood and plastic pointed is usually preferred. But it’s weird as what we are using also falls under sheet metal fasteners.
I have built my 3,5 Kg hexa also with a 3D printer. I would prefer ABS rather than PLA. I have used 80% infill but with lower part sections. Good part design is necesary to reduce weight and increase rigidity. I have also tried Nylon-CF filament but, in my case and with my printer, I had better result with ABS. I also used rounded arms, but I won’t do it again. With ABS and CF arms, you need a lot of preassure or a tape to hold the arm and to avoid motor tilt. A squared tube will solve the problem. I remember that when I built my hexa (three years ago) CF squared arms were hard to find and even more expensive, but now they are easier to find.
Part of the reason we are restricted to 2kg is use of rather small round arm tubes without physically indexing the motor mounts. It works well enough for this use case. Square tubes would make things easier to build but they are still a bit harder to reliably source vs round. And square CF tubes (wrapped; not pultruded) are more expensive last I checked. I think you would have to pay someone to do a large production run of small square CF tubes to make it work at our price point. We could go with small aluminum square but also hard to find; especially in thin walled.
Just checking in to see how things go? Docs are stalled for a bit longer while I attend a class. Next up will be prepping for first flight and then tuning.
Matt, If you can get me the overall dimensions and dimensions between holes I can at least add a platform mount for your GPS. A cover is a little harder and not really required for simple stuff. Can add cover later if required.
Checking in. Nothing like a 14 week software bootcamp to make you appreciate hardware. Not sure how long it will take me to get them done but devframe docs continue. Up next is initial tune and test. Especially since winter is around corner for this area.
Anyone still working on a build? Project is still alive but slowed by … well … uhm … it is 2020.
Seriously though, it might be 2021 before I tick all the boxes here but it’s still on The List. Meantime, any insights from those still building devFrames is much appreciated.
Very cool! Thanks for sharing. Really like the fact that you are experimenting with the slotted platforms and vertical cross mounts. Also think you are first Cube tester. Been around since the 3DR days and still never played with one. Once I get devFrame to v1 I have parts to go bigger with a Sony A6000 on a front mounted gimbal and hybrid or stretch X. Think that one might get a Cube.
Has anyone done anything with a gimbal? I’m thinking of hanging my FC from the middle using vertical mounts like above and then hanging a gimbal from the center hanging off the CF tubes. I’ve also created some longer legs - was having issue with the short legs in the grass.
That looks fantastic Jim. Originally had adapters to use cheap off the shelf plastic legs from the Flamewheel 450 family. Need to see if those files are still around. The current legs are pretty short for grass.
One of my planned builds is to see about mounting a Canon point and shoot center bottom with GPhoto2 running on a companion, but was looking to use a similar static vibe dampener like you show for your GoPro. Affordable mapping is the use case there so no multi axis gimbal planned.
I have a 3 axis gimbal for a Sony A6000 but it’s slated for a front mounted hybrid X that will be a bit larger than the devFrame. Was actually the initial inspiration for devFrame and been collecting dust for too long.
Not sure when I will get back to devFrame but the plan is ~March some time. Just started a new job that should allow me to get back to build/test/crash around then. I hope.
My cheap gimbal FINALLY got here - was bigger than I thought but I got it mounted. Haven’t flown it yet though
Thinking of swinging the front mount tubes back so they are in “H” mode to make sort of a dead-cat frame - that would give me a bit more room in the front for the gimbal. Anyone run a config like that before with Arducoptor?
@Matt_C You should be able to use a Box End Assembly in the middle of the frame to give you two more arms. So, stretch X with another Box End in middle for two more motors. Breaks up the internal volume a bit but that is likely quickest hex and should still be really strong.
@thecrumb hybrid X should work OK with X as frame type as long as you keep your CG based on motor centers. Have never flown one on ArduCopter though.
I’ve since acquired the correct screws.
