search

sim- / tgy

tgy -- Open Source Firmware for ATmega-based Brushless ESCs
http://0x.ca/tgy/
689 stars 386 forks source link

Afro 30A w/ NTM 28-26 1200kv fails under high load #54

Closed erikmansson closed 5 years ago

erikmansson commented 10 years ago

The afros are flashed with the newest firmware and I've tried a lot of different settings. But they all, more or less, fails under high load. Fails as in loosing the sync and just flipping out for a brief moment. Sometimes they brake so hard that they start to reverse. Running 3s LiPos. Only happens with the prop mounted (8x4.5) at about 70% throttle or more.

I've tried changing the timing, turning I2C and UART on and off, changing POWER_RANGE, COMP_PWM...

Are the motors crap? Or am I doing something wrong?

SilviuDadus commented 10 years ago

Hi Erik, odds are the motor is crap. I have the NTM 35-30 1400 kV motor that does the exact same thing without the prop! You can try setting the SLOW_THROTTLE on and see what happens. Also you can try the BLHeli software which has a very different approach and in some cases it worked better than Simon's firmware.

erikmansson commented 10 years ago

Tried flashing BLHeli. Ran great for about half a second, then the famous white smoke appeared. From the motor. I think we can safely say that the NTM motors are pure crap.

SilviuDadus commented 10 years ago

Hi Erik, sorry to hear about the motor. I've tested a NTM motor, a HobbyKing Donkey motor, a E-Max CF2822 and a hard drive motor with the BLHeli software and never had any issues like your white smoke. You got a lemon this time..but from what I hear NTM motor have some construction and QC issues, too bad though, they look really good and don't heat up much...

sim- commented 10 years ago

Hello! There are two other possibilities you may want to consider.

There was what appeared to be a batch of ATmega8s with dodgy comparators. See https://github.com/sim-/tgy/issues/50 ...and, there is an outstanding bug with re-aligning to a running motor (powerskipping at start), the code path for which can be triggered by brief loss of signal, which can be triggered by noise on the PWM input. (Also see above issue.) I need to fix that path, but you could also try twisting your PWM cable to see if the result changes.

There are some issues with those motors for sure (the back-EMF has some third-order harmonics: http://0x.ca/sim/esc/motor_backemf/ntm1200k.png), but the 1200kv is pretty quiet and the one I got is balanced.

erikmansson commented 10 years ago

I gave the rest of the NTM motors to a friend of mine, and they now run okay with the HobbyKing 30A escs. Also simonk, of course. So there's two possibilities as I see it. One being that my Afros are bad, and the other being that he isn't pushing the motors hard enough to make it fail.

I'm pretty sure PWM isn't the problem. I run the same wiring with another more powerful setup, and it runs awesome. Also Afro 30A, but not the same as the ones I had problems with. I actually run those which didn't work in a tricopter now and it works good.

While we're at it, I have major issues with the RC_PULS_REVERSE option. It makes my escs totally unusable - sync loss, startup cogging, no stable rpm. What can be the issue here?

erikmansson commented 10 years ago

Just wanted you to let you know that both the Afro 30A escs and the NTM motors were bad. The ESCs run like crap where other escs with exactly the same fw run fine. And the motors had bad bearings.

I'm running other Afro 30As and T-Motor AT2216 1250kv on 4s without a single problem whatsoever. Just for fun, here's a vid of me flying that config :) http://www.youtube.com/watch?v=dNNav44rOr0

I do, however, still have problems running them with RC_PULS_REVERSE.