Also, while in position hold, the heli wanders around a little bit. What parameters can I change to make this a little tighter?
Is this the correct graph to analyze?
Hi Chris,
Did you use the wiki tuning method? This seems pretty low frequency (~2hz). I need to see your current param file to see what you’ve got for current settings. From what I’ve seen too much P gain results in a little higher frequencies. However this is a 1200 size heli?
Here is my parameter file. Stingray 2-9-2019.txt (23.9 KB)
It’s not far off from recommended 3.6 settings.
I think I really only changed PSC_ACCZ_P from 0.28 to 0.20 because I saw it oscillating in altitude hold.
The heli is not oscillating, it’s just wandering within + - ~2 meters in the x,y, and z. Could be because of GPS signal jumping around, but I feel like it can hold a little tighter than what it’s doing now.
I have read the wiki, but it does not describe how to tune position hold mode. Unless I can’t find a particular page?
Depending on where you are in the US right now, it can be a lot worse than that. WAAS is available only intermittently across the entire US and the affected area advertised by USAF Space Command (who runs the GPS satellites from Colorado) is a lot bigger than what they said.
After reading another thread about how I should have my Rate vs Desired Rates match closely, I took a closer look at some of my logs and noticed while in a hover, the magnitudes of the actual rates are huge compared to desired. However, when I give it sharp inputs, it does seem to match. Is this correct? Here are some snapshots:
Looks very nice. You’re bound to get some GPS drift with all those trees around. I don’t know if I’d trust the GPS position enough to hold the old Taranis out like that 
Yes, look at the rates when you make full stick inputs. The amplitude of the peaks should about match. It doesn’t really tell you much when you’re not moving the sticks and not calling for big rate. Yours looks really good to me and if the heli responds the way you want I wouldn’t turn it up any more. These things actually fly fine on just VFF - that’s all that’s used on flybars. The PID loop basically takes the place of the flybar and damps the control inputs to make it smooth and graceful, and provide correction to outside disturbance like wind.
I truly appreciate all the help Chris. And you too Bill!
The trees are a little further away than they appear in the video haha. They also serve as something to hide behind in case the heli decided to go hay-wire lol! Weather is suppose to be pretty lousy for the next week or so but next flights will be in a bigger field 
1 Like
Some forward flight from today: The carb sounds like it could use a tune, but from the video it seems to be holding a pretty solid headspeed (no governor yet either!)
And a loiter video:
Yesterday for the first time, I tried RTL a couple times during my flight and noticed some poor performance with yaw during it. When the controller commands yaw to turn toward the home waypoint, it seems to overshoot and oscillate a little bit back and forth as it travels to the home point. But when it yaws again to point in the direction it originally takes off from, the yaw is solid. Same thing happened twice. Any advice for tuning this out would greatly be appreciated. Here are some pictures from my log:
I’ve noticed that before too, but I never use RTL as a normal thing. You’ll notice that it’s not your tuning, but the autopilot actually does that for some reason. The desired and actual match.
I have no explanation for it other than maybe it wags its tail like your dog does when you call him home 
haha good to know. Yea I don’t really plan on doing RTL regularly anyways, but I was just curious to try it.
I am a 99% user of RTL with 3 aircraft. None of them have tail waging.
Conclusion: They have to come home but don,t like it. Ha, ha.
1 Like
I have no clue why it does that. The couple of times I’ve used RTL just to test it mine have done it too. Maybe it’s only gassers that does it - it needs to shake its tail like a fish to get going?
I typically can’t use RTL because when it comes to a stop over the home point it does a hit the brick wall stop thing. Can get by with it with smaller helicopters. But big ones react quite severe to sudden pitch changes like that. And Copter is not exceedingly smooth with a lot of the autonomous stuff for big heli’s.
And Copter is not exceedingly smooth with a lot of the autonomous stuff for big heli’s.
That’s why Bill and you developed the:
L1 navigation for tradheli
I guess.
Well, that is true. We’re trying to improve on the situation for heli. Copter Nav works, but it’s not ideal. Helicopters were kind of forced to adapt to multirotor methods which leaves some things for heli’s being somewhat brutal and inefficient. The L1 Navigation controller is a nice step forward in providing more ideal navigation flight controls for UAV helicopters, and better fits their flight dynamics.
Things like flaring to come to a stop are different between multicopter and helicopter due to the fact that the only basic difference between a helicopter and a fixed-wing in flight is that on the fixed wing the wing doesn’t move - on a helicopter it is rotary. It is impossible for multicopters to have the same flight dynamics as the helicopter because their little tiny rotors don’t have cyclic pitch, they don’t create the lift because they don’t have the blade tip speed of a much larger single rotor (lift is proportional to the square of the speed of the airfoil), so the multicopter does not respond the same to attitude changes. Even very simple things like the ACCEL_Z_P, multirotor requires much higher gain. While using that same gain on a helicopter, with blade tips traveling at 300+ mph and collective pitch varying the lift of the airfoil, will make a helicopter jump and down like a rabbit that has gone nuts. The helicopter has a vastly more powerful and efficient wing for its relative size. And the helicopter is an inherently stable machine - we fly full-size ones every day without any sort of electronic gyro stabilization. It is impossible to fly a multicopter without electronic stabilization.
These differences make helicopter actually easier to fly if you start with a fresh sheet of paper. And understand that in hover, out of translational lift, they are similar to a multicopter. In flight, once it enters translational lift it transitions to being more like a fixed-wing. So things like the L1 Nav are ideally suited to the helicopter. The challenge was getting it to transition back and forth from the hover/slow speed profile to the true flight profile where it’s dynamics as a flying machine take over. So L1 is infinitely smoother than anything we’ve had before for heli’s.
3 Likes
@ChrisOlson,
I have an opportunity to use my heli in a real-world job and would like to pick your brain about sling loading your helis.
The payload I’ll need to carry is around 4-5 pounds. Is there a rule of thumb on how long my line needs to be underneath my heli? Also, I plan to do this flight semi-autonomously (meaning, either in position hold, or loiter). I haven’t had a chance to do auto missions with this yet and don’t want to mess with that right now.
My flight test cards will be dropping the payload from various heights and various speeds. To me, it sounds like the easiest mode would be to do it in loiter mode (with my loiter speed tuned to my desired forward speed - that way I can have my cyclic stick pegged forward during the dropping pass and know that I’m at my target speed and altitude condition.
The particular sling load I plan to carry will be a dropping mechanism at the end of the line, with the drop-able payload firmly attached to it. The dropping mechanism will be no more than a pound. So heaviest I predict is 5 pounds fully loaded, then 1 pound after release.
Things that make me worry is what might happen after the payload is dropped…
My heli is currently tuned well for this aircraft weight (without payload), so I’m hoping it won’t go crazy. If it does, it seems like my best plan of action would be to immediately switch to stabilize mode. I may be pretty high and far from myself to do acro mode.
What are your thoughts?
My heli is quite large spinning 1100mm blades so it seems like 5lbs is just chump change for it.
I do plan to fly an 8lb dumbbell this weekend to practice for this. Is it safe to assume that as long as I’m keeping some forward speed during my flights and passes that the load won’t swing worse and worse?
Thanks for any and all help!
@Chris_Khosravi I know that Chris Olson has busy with a new business venture and I’m not sure whether he will have time to comment. I don’t have experience with what you are asking but I have read some other posts regarding tuning for various payloads for multicopters. Read through this thread https://discuss.ardupilot.org/t/two-crashes-with-an-overpowered-drone/43109/3 . Leonard is the Copter flight controls designer and provides some insight on tuning with payload in the aircraft and not slung below the aircraft. Not sure it helps but thought I would post.
Thank you Bill,
I have read that thread before. Both Leonard and Chris recommend tuning for lightest configuration, then adding payload (which is what I’m doing).
I’ve just never tried slung load flights so I’m trying to imagine all the things that could go wrong 
-Chris Khosravi