Servers by jDrones

New ArduPilot Acro VBar Mode For Helicopters

(zhangpeng) #21

I was very excited to hear the news and see the progress of the work. I am expecting Ardupilot to have such a flight mode for a long time. I will take the time to participate in the flight test in the near future.

Thanks again to the efforts of developers!

(zhangpeng) #22

The cold winter finally passed. Today, I finally have time to test the Vbar Acro mode of Olson’s 3.6.8 firmware. I upgraded directly from the 3.6.7 firmware. I didn’t change other parameters. The direct test flight has not been flying for 6 months. The ALZRC 420 has no aileron helicopter and CUAVv5 flight control. It switches to Acro mode. There is no longer the problem that the previous tilting disk slowly changes. The aircraft is directly unlocked and take off. The aircraft is stable and stable. It really has the FBL system. It’s great to control the feel! Two places that feel weak: First, the feel of pitching and rolling is soft. Second, Yaw’s handling feel is not as clean as the FBL system. Is it a setting problem or is it a current code problem? I am very grateful to Oslon for his great dedication, and finally let Pix’s manual flight mode no longer be so lame!

(Chris Olson) #23

@zhangsir I must say I enjoyed your description! :grinning:

And I also must say this was a community effort project between myself, @bnsgeyer and @Ferrosan to make acro more suitable for helicopter and not so “lame”. :grinning: The more direct stick-to-swash “feel” without the robotic stop-turn-go stuff makes it a pure pleasure to fly with a natural feel that a helicopter should have.

To get snappier handling in the pitch and roll and adjust yaw, adjust the following to your preference

And if you run up against the acceleration or rate limits adjust these:

To adjust the rate of decay (changes the “feel” of the virtual flybar) adjust these.

This new mode also makes it possible to fly the helicopter with the EKF blown. Tested here, and it works fine. It is only loosely connected to the attitude controller’s attitude solution. Since it constantly decays to actual attitude, it doesn’t really matter what it “thinks” actual attitude is. So it provides flybar redundancy for a FBL helicopter and when I blew the EKF, other than the ground station complaining about it, I could not tell it in the way the helicopter flew.

Other features of it - slope procedure is a piece of cake. You can hold constant cyclic pressure into the hillside just like a real one, and it won’t build up the rate command and tip your helicopter over. No more goofy nudge and center the stick and see what happens. The pilot actually has control of the machine the way Igor Sikorsky intended for helicopters to work.

In the end, getting it to work was a relatively simple thing. But it’s still a masterpiece and it took a lot of flight testing before I was happy with it. It has since been PR’d to master, but not merged (yet).

(zhangpeng) #24

@bnsgeyer@Ferrosan,Sorry, I didn’t read the previous post carefully. It is the result of the joint efforts of several community members. I am sorry again, and thank you for your hard work, thank you!

Thanks to @Olson for the detailed parameter setting instructions, I will immediately look for time to track the actual flight effects brought by these parameter changes, and report it to several hard-working code contributors in a timely manner. Thank you for sharing with everyone in this community. The traditional helicopter firmware application technology, really very honored.

cheer !!!

(zhangpeng) #25

@ChrisOlson,@bnsgeyer,@Ferrosan, The weather was very good this morning. I got up early. According to Chris’s parameter suggestion, I adjusted some of the Acro parameter settings of CUAVv5 and made a simple test flight. I want to say that it is really great, and the feel has been greatly improved. The video below is a clip of a simple test flight that allows you to visually see the maneuver of the operation. At present, the AC RO_BAL_ROLL is set to 1, the action is a little weak when the level is restored, and may need to be improved, as Chris suggested, at a value of about 2.

The Loiter mode was also tested briefly. In the case that the satellite conditions are not very good (the RTK mode is not set, only the GPS is received separately), it is also very good. The fixed height mode is changed from the previous GPS positioning to the barometer reference. High performance is very good.

The only regret at the moment is Yaw. Even if I change the parameters of Yaw in Yacro, the effect of the lock tail is far less than that of the FBL aileron gyroscope on 3D helicopter flight. I look forward to several great developers in the community who can find better ways to enhance Yaw’s handling.

(zhangpeng) #26

It is found that the Virtual_flybar function is not included in the P6484 firmware. There is no PUXHACK NANO, which is a Cuav, which is PUXHVID, NATOVI, I can not know the test 3.6.8, there is way to let PX4V4 use virtual_bar barro?

(StefT) #27

Hi Chris, this is great news and I’m very exited to impliment it. However, I followed the link, which took me to 3.6.8. After upgrading my Cuav v3, I can’t seem to find acro_m as described in your video. Am I missing something? Do I need to force acro_trainer to a different value? Thank you.

(Bill Geyer) #28

@StefT there are no special selections to get feature. put the aircraft in acro mode. You do need to disable trainer mode. then you’ll set ACRO_BAL_PITCH and ACRO_BAL_ROLL to set the feel of the virtual flybar. setting these to zero will make if fly like the stock acro flew like before this change.

(Chris Olson) #29

We changed it in the final version of the code to be default for heli, and it makes no changes to the function of acro for mulicopters. So with heli, it’s automatic.

I see Bill just responded with how it was finally implemented, and the applicable settings.

This seems to be a pretty popular feature, as I think heli pilots are probably the largest userbase of the copter acro.

(StefT) #30

Thank you Bill and Chris. I changed the pitch and roll parameters but did not deactivate trainer mode. That is why I did not see any change in flight.

Unfortunately I have not been flying heli the past 4 months due to other obligations and only start picking up where I left beginning of January. I had a close call this morning with my 800 gasser due to the ghost effect at takeoff (fortunately recovered in time.) as I was testing AUW. I normally feel more comfortable taking of and land in acro mode. If I understand correctly the vbar option will greatly reduce the risk at takeoffs.

Thank you for the amazing effort and hard work you guys put into the development!

(Chris Olson) #31

Yes, it does. If you set the BAL settings to zero it will be like it was before. If you set them to 3 it will be almost direct stick to swash. Most people are going to like it somewhere in between, probably around 2. With a setting of 2 it provides a slow enough decay so you get nice holding of attitude when you move the cyclic, but fast enough decay to give it the more stick-to-swash “feel” of a full-size helicopter.

Set at 2 you will have no problems with buildup of rate command on the ground. You can tip the heli around and back and forth with it running at full power on the ground and it will follow exactly what you command it to do with the cyclic control. It is such a huge improvement over the pure rotate and hold rate command that it’s like flying a whole different machine and makes a FBL helicopter feel and act just like a flybar.

(Chris Olson) #32

@zhangsir Build is now available for CUAV V5Nano. This is done with automated build script, the V5Nano is now added to the script.

(zhangpeng) #33

Understand, thank you Chris, I am about to receive CUAVv5 nano, ready to start testing!

(StefT) #34

@ChrisOlson,@bnsgeyer,@Ferrosan. I have been testing the new vbar mode for a few days and want to say that I am extremely happy with the results. The handling is smooth and provide a similar feel to an “off the shelf” flybarless unit for sport and scale flying. My takeoffs are now with a lot more confidence as well as there are no sudden surprises. I am sure the previous stabilize mode was OK but taking off with a high disk load from an irregular spot was a bit nerve wrecking for me.