I was reading in another topic about setting ATC_RAT_YAW_VFF to 0 and COL_YAW to 0.
This is on my 800 gas UAV heli that carries several different weight payloads. Oh, and its flying again with no problems after replacing the intermittent servo.
What would be some good default settings for the ATC_YAW parameters to start tuning with?
What would the process steps be for tuning YAW?
Thanks
Doug,
I haven’t spent a whole lot of time tweaking my tail tune. I do believe that you need to set neutral position of the tail mechanically for hover that way if you are using I gain, it minimizes the amount of integrator that builds up and should reduce overshoots. @ChrisOlson posted on this topic here with a different method of setting Rate P and D gains that uses a low I gain. I think the big take away is that if you carry too much integrator in flight it will introduce overshoots in the response. Also that too much VFF gain will do the same thing. Hope this is helpful
Regards,
Bill
Here is what we’re running in our survey helicopters. Note no VFF, it would appear I set the I gain to half of the P. No scientific reason for that I-gain setting, that I recall. It must’ve looked like a good idea at the time and it worked.
These helicopters have vertical stabilizer so may not be representative.
When setting the P-gain with zero VFF make sure you have 100% tail pitch movement by testing with your RC radio. P-gain too low and you won’t have full travel.
ATC_RAT_YAW_D 0.005
ATC_RAT_YAW_FILT 25
ATC_RAT_YAW_I 0.11
ATC_RAT_YAW_ILMI 0
ATC_RAT_YAW_IMAX 0.5
ATC_RAT_YAW_P 0.22
ATC_RAT_YAW_VFF 0
ATC_SLEW_YAW 4500
@Shotfire I do not have any of those survey helicopters here, Kayla has all those. But I do have my Raptor 716, I check that and I am actually using the same settings in it. That one does not have vertical stablizer and is smaller 700 class. So maybe those can be used as starting settings.
I think you can consider when tuning that the P is a proportional correction, like a pilot stepping on the pedals to correct a yaw error. If the error is like 10 degrees, it will apply a smaller amount of pedal. If it is 45 degrees, it will apply a bigger amount of pedal. As I mentioned, with no VFF this must be big enough to provide full tail travel.
The D-gain provides more of an instant correction to make it track the desired in your log. For instance, if a wind gust hits our vertical stabilizer and knocks it off track, the D will jump on that and get it back very fast.
The I-gain can cause “yaw hunting” where if you have a video camera on it you’ll see the yaw constantly going side to side in the video. Turn that down until that goes away. I must’ve arrived at 50% of P for that.
With our UAV helicopters I don’t think we want a real fast tail or yaw rate. We are not flying aerobatics and want it smooth for video or camera work, or external sensor work, or even long-line work. That can be limited with the rate setting for degrees/sec of yaw that you desire. And the slew rate also adjusts it. On the rate deg/sec setting I don’t think I’d go much below 180 deg/sec. That is 2 seconds to make a complete turn at full pedal, which I think is pretty reasonable for a 800.
I don’t recommend using the COL_YAW at this time. I’ve never had good luck with that setting. Just let the P and D handle it. Again, we are not doing “pitch pumps” or any of that with UAV and going for more of a scale type handling characteristic. So, typically, collective changes are pretty gentle with UAV.
Glad you found your problem with that bad servo. Those can be a pain. Sometimes you don’t know if it’s the controller itself with a bad pin, or the servo. Swapping them around and flying it one way to find it, but that requires re-setup of the head every time. If the controller or carrier board is bad you have to eliminate bad cable or connections in the pins before you replace an expensive controller. It is just a pain tracking one of those intermittent problems down. And sometimes even more difficult to identify it. We’ve had 'em on both cyclic and tail before and they eventually quit working altogether when it’s a bad servo. A failed cyclic servo is easier to handle than a failed tail servo. I’ve had some pretty hairy landings with a stuck pedal situation with a bad tail servo with the helicopter spinning like a top, then cut the power to arrest it and get it on the ground without wadding it up.