Hi Richard,
By any chance did you accidentally reverse the lipo polarity on the fcb at any point? This sounds like a symptom of a bad pwm input/output from reversed polarity resulting in partial damage to the processor.
Cheers,
Jim
Quadrocopter and Tricopter Info Mega Link Index

Originally Posted by Richard_s

Hey guys. My parts for DIY 939 arrived, and I hacked together bamboo/balsa 11cm M2M quad. 20 ~ 22 grams without battery, about 29 with stock one. I used wltoys 939 motors (two CW and two CCW), stock 939 battery, and stock props. I get some nasty behavior here... with fresh battery, just after lift-off, quad starts to rise pretty fast and no matter what I do with throttle stick, it just goes up to the ceiling. Only zero throttle stick stops this, quad stops motors abruptly, flips nad crashes . I use 9x transmitter, with linear throttle setting. After maybe dozen of such tries, as battery gets lower a little, I can actually fly the thing. However, it shakes - oscillates quite a bit, but flies fully controllable until LVC.

Any ideas? I'll snap some photos later.

Wobble... it is more like oscillation. Unbalanced prop would be noisy and could be felt by running motors while holding quad, but vibrations from motors/props are minor, and props are quiet. I selected best four from two sets. I suspect bamboo bending under load, and fc overcorrecting, or something. But such short arms bending... dunno.

I did not reverse polarity. I only have stock batts and those connectors are impossible to connect reversed, even briefly by accident.

I'll try to dial in some soft throttle curve, any ideas for starters?

bamboo, balsa, ply, paper tubes

I have reversed those stock battery connections before. I did it on purpose because the batteries I was using were reversed. But yeah, you would probably notice since it's not an easy fit.

If you think it's the throttle on the TX then I would start with just simple expo at first before bothering with an actual curve. Maybe 30% expo or something to start with.

Very nice frame Richard,
I'm gonna copy this one soon as I received more motors today from bangood... hope you don' mind
Olivier

Looks GREAT, Richard!

Originally Posted by LittleMo

If you think it's the throttle on the TX then I would start with just simple expo at first before bothering with an actual curve. Maybe 30% expo or something to start with.

Ok, I'll try expo first.

Originally Posted by ztoon13

Very nice frame Richard,
I'm gonna copy this one soon as I received more motors today from bangood... hope you don' mind
Olivier

Sure, go ahead. There is so small contact area between 1.5mm ply and motors/arms, that I used hotglue to hold it together, filling corners between parts. Not best glue to use, but after some crashes it lets go and nothing breaks. I just scrape the glue off, and re-glue parts back together.

Yeah I don't see anything wrong with that frame. I don't think it's vibration. I just checked my own quads and the board is actually loose on the frame and it still works fine. Or maybe it works fine because it's loose.

So maybe a bad motor, bad board, or it's the TX... I guess

Originally Posted by rusty case

Looks GREAT, Richard!

Thanks! Would look better if it actually flew...

Oh, I forgot to mention you might want to pull the antenna away from those motor wires, maybe it's getting interference. Seems unlikely but possible.

Richard,
Regarding your oscillations: Is the fcb directly on the wood, or did you place some isolation material such as foam or servo tape under it? I can't tell how well the rubber band is holding it to the center piece, but it could be allowing excessive vibrations to the fcb. Something to experiment with anyway...

On your throttle...I know this seems obvious, but gotta ask. Is your throttle channel reversed?... It could be a Tx end point issue where there is a dead step at the high and low endpoints that are set.
Cheers,
Jim
Quadrocopter and Tricopter Info Mega Link Index

Originally Posted by jesolins

Richard,r
Regarding your oscillations: Is the fcb directly on the wood, or did you place some isolation material such as foam or servo tape under it? I can't tell how well the rubber band is holding it to the center piece, but it could be allowing excessive vibrations to the fcb. Something to experiment with anyway...

On your throttle...I know this seems obvious, but gotta ask. Is your throttle channel reversed?... It could be a Tx end point issue where there is a dead step at the high and low endpoints that are set.
Cheers,
Jim
Quadrocopter and Tricopter Info Mega Link Index

I tried fc directly on wood, and now it sits on foam tape, can be seen on pic. Made no difference. Rubberband is pretty tight, it holds battery on the belly.

Throttle channel is... pretty normal I guess . I fly airplanes, KK 3copter, MWC 3copter, v911, and no problem. Throttle is linear from min to max. I'll see what will expo do.

when first trying a 9x i got "hit the ceiling syndrome" too. it required some tweaking of rates to get throttle and other sticks under control. the stock wlt radios have dual rate button and the defaults are nothing like a 9x. i went back to the v911/v939 units because 9x was just too big so dont have details on those settings handy.

as far as wobble a couple of my builds did that for no obvious reason. best solution was to make another frame with slightly different dimensions or weight. same motors, board, bat, etc but worked like a charm. no explanation. i suspect their control algorithms are not the same pid stuff used by kk, mwii, etc and one of the drawbacks seems to be instability at certain peculiar arm lengths and weight distributions. even tiny changes made a world of difference. longer arms always help with wobble too. specially with v929 board. my advice is throw together another quickie. maybe lighter and a little bigger dimensions but not putting a lot of work into it just in case. might try different lipo too.

ive now had experience with 5 reverse battery boards both types and have yet to see flaky behavior. either they worked or didnt. ones that didnt got going by fixing blown run. in all cases the tantalum was burned to a crisp and although it seemed to work w/o i replaced with 10uf ceramic just to be safe. i also saw no difference between boards firmly fastened down or suspended loosely by wires. you never know though. theres so much going on in that mcu im amazed to see that almost all my attempts worked.

could still be a bad board though. having two is a big advantage and imo considering $8 price tag no excuse not to.

Great info Dave, thanks!

Expo didn't solve anything, quad responds to throttle maybe after first third of movement, what looks like linear part of expo curve anyway . However I got a little lucky, I could oppose "ceiling shot" ONCE. Man, this thing is touchy... .Some serious throttle curve tuning will probably be the cure.

Other channles seem fine, I can fly it after voltage drops off a bit. I noticed it oscillates aroud one diagonal axis (running thru two arms), so, two motors are "still", and two see-saw about that axis. I'll give this frame another shot, I'll re-measure all arms, and maybe shorten one of wobbly ones, or something . I REALLY want this to work, I was waiting for postman like kids wait for santa

forgot to mention bad motors. swapping those around can be a good diagnostic. even better to have spares there too. good luck.