Servers by jDrones

Collective Pitch Quad support


(zhangpeng) #41

That is sufficient to set the corresponding positive and negative pitch, can be achieved by the traditional helicopter backflush self-hovering?

Did you say that a single helicopter is not a traditional helicopter? But a uniaxial or coaxial helicopter


(tridge) #42

it seems to work fine for throw mode, which produces a very high rotation rate.
I’d except 3D heli to work fine.

I mean all helicopters, and by “single” I meant one main rotor, ie. a traditional heli


(Laszlo Kiss) #43

Hello tridge,

Can you help me, how this system work? Does it change the RPM of the motor? Or it does everything with the blade adjustment? If you want to increase the height, the motor speeds up, or it increases the blade angel as well? Or the blade angle is used just for the stabilization, and drive control?

I am working on a tilt rotor with 4 props, driven by petrol engine, and all blades are adjustable. I am using the ArduPlane image, and I planed to connect the blade control servos to the ESC output. Can it work? I dont know, where to connect the engine throttle. What is the different in this new solution?

Laszlo


(Chris Olson) #44

Collective pitch quad-heli’s use only the collective pitch for everything. The motor runs at constant speed. It uses differential collective pitch for stabilization, and to go up or down. And it also uses differential collective pitch for yaw control.

In theory, motor throttling could be used to go up or down. But I would consider that bad design, as when prop speed decreases if the motor is throttled down, stability and responsiveness of the collective pitch would be compromised. I believe this was a common flaw in the first types of these models to appear on the market, using a throttle/pitch curve in the FBL units that flew in them, and varying motor speed with the collective lever. Hopefully, ArduPIlot has addressed this issue and uses the RSC system from Trad Heli with a governor to prevent control, vibration and stability issues caused by throttling motors on a collective pitch control system.


(tridge) #45

Chris is completely correct about the control method (of course!)

That is all available with HeliQuad, as it is a heli frame, so all the RSC options are available. So far I’ve just used the simple set-point option though, and that has worked well. I don’t have RPM feedback, so I’d need to add that in order to do a proper governor. It may even be that the ESC in this frame has a built-in governor? I don’t have any information about it.


(Chris Olson) #46

I don’t think the ESC’s in these hobby-grade frames have a governor. But, I suspect people will be building their own and probably using higher-end ESC’s with a built-in governor. And then H_RSC_MODE=2 will work perfectly to maintain headspeed. That will be one of the keys to reducing vibration, as collective pitch can get quite rough at high AoA blade pitch angles if the headspeed drops. That was a common problem with bricking FBL units back in the day with the Stringray 500 model.

I see no reason the heliquad can’t be made to be reasonably vibration-free using a little higher quality components from regular helicopters and a decent RSC system to run the heads at a speed where their resonance does not couple with frame resonance.


(tridge) #47

I’ve solved the vibration now. I used some 25g lead weights from an online auto store (the types of small weights used for balancing car tyres). They come with stick on tape. I put 4 of them on the bottom of the pixhawk and otherwise didn’t change the mounting at all. In todays flight the average vibration dropped from around 60 m/s/s to 7 m/s/s, which is pretty good for such an easy change!
It flew a lot better with that weight added!
The only issue I had today was that the maximum roll/pitch in ALT_HOLD mode was quite limited. It turns out it is the alt_hold angle limiter trying to leave headroom for collective. It seems there is something wrong in the HeliQuad motors backend that it causing it to limit the angle too early. I’ll look into it.
Cheers, Tidge


Slow Heli Reaction in ALT HOLD mode
(Perth UAV - Sam) #48

So you have solved the IMU vibration by adding some mass damping on the PXH, however is the rest of the airframe still very vibration prone or relatively smooth? Are there any noticeable RPMs that cause drivetrain/frame vibration?

I’m wondering about where the RPM sweet spot is for drivetrain, motor and prop losses are on a collective quad whilst avoiding vibration harmonics etc.

On a non collective quad my experience has been that running smaller higher kV motors with fairly small low pitch props can produce good hover lift numbers in the 10g/W range up until preceding blade wake turbulence kicks in. This is completely opposite to what would be required when using the collective quad with a tilt mechanism, where the blade pitch would need to be steeper to match forward cruise speed for best efficiency.

With little knowledge of the heli setup, Is it currently possible to setup two different throttle/RPM control methods, one for hover and one for forward cruise, for a tilt-collective rotor quadplane? Also, would it suffice to provide RPM feedback, or would motor load also help modulate throttle/RPM?

I think the only way collective pitch can improve hover endurance and range is if it can be optimised for each hover and forward cruise. Theoretically, if one measured motor RPM and load/current draw, the throttle and blade pitch could be managed in such a way that the minimum current draw is achieved for each flight phase, whilst maintaining enough control response for flight control.


(Chris Olson) #49

There is three modes available in heli for RSC.
Mode 1 is a throttle/pitch curve set up in the transmitter and either slaved to the collective lever or made adjustable by a switch/knob configuration.

Mode 2 is governor mode with a setpoint that the software sends to the engine’s governor (by far the most common in heli’s). It has adjustable ramp and runup times, primarily used to prevent torque from spinning the helicopter around in a circle on the ground during runup before yaw authority becomes sufficient to counter main rotor torque.

Mode 3 is a software-controlled V throttle curve which varies throttle similar to a curve slaved to collective with Mode 1, but is done in software with a three-point curve corresponding to max negative, mid-pitch (zero collective), and max positive collective.

In theory, Mode 1 can be used with a two-position switch to set two different power values or curves, either using a governor, or not.

With any of them in heli, care must taken to set up the collective failsafe properly for loss of RC signal. Or it will shut down your engine in flight with a radio failsafe and you’ll have to autorotate it. An in-flight restart is usually impossible due to ramp time. I don’t think these quadheli’s can be autorotated like a regular helicopter, so that’s probably not an issue for them. If they shut down in flight, it’s an automatic crash.


(Laszlo Kiss) #50

Hi Chris,

Thanks for the reply.
I am trying the ArduPlane image with Quadplane support, but I have trouble with it, so I think I will install this ArduCopter solution for the quadcopter function test.
Is there any description, how to activate this new function?

Laszlo


(Chris Olson) #51

@tridge would be the best person to ask that. I have not set one up. But I think it is only available at this point in 3.6-dev and you have to to define the frame type. But I don’t know the details of that.


(Perth UAV - Sam) #52

Thanks Chris for the detailed response regarding the various governor modes.

Am I right in assuming that there are no self learning efficiency profiles incorporated into any of these? For example one for best forward cruise and one for best hover energy consumption?

It would be interesting to try and see if auto rotating the collective quad is at all possible. If it is, I’d assume they would all have to remain connected to eachother through the drivetrain to maintain control authority on descent? I wonder what the drivetrain losses are like on that thing, and how yaw control can be maintained.


(Chris Olson) #53

No. There is no such thing

With regular helicopters there is an over-running clutch in the transmission that allows the main rotor to “freewheel” during descent to maintain headspeed and main rotor inertia. The main rotor drives the tail rotor thru the autorotation transmission gears. And proper autorotation depends on the use of cyclic pitch to drive the main rotor - it is not driven with collective. The collective pitch is mostly at feather although some pilots will use up to -3 degrees if the helicopter is close to what is known among helicopter pilots as the “dead man’s curve”. All helicopters have a best autorotation speed, which is normally 1.3x the published best descend speed in the flight manual. During final stage of landing cyclic is used to flare the helicopter like a fixed-wing, with the addition of collective using stored energy in the main rotor to slow the descent. A skilled pilot can set one down soft as it does with the rotor powered.

Being the quadheli lacks cyclic pitch I have doubts it can be autorotated effectively. I do not believe the quadheli to be capable of it because it is lacking one of the key ingredients that makes helicopters efficient in forward flight, and highly maneuverable - cyclic pitch. The final flare using cyclic to bring the helicopter into a nose up attitude builds HUGE headspeed - usually far beyond what is used in normal flight. If you look at the phase angle of the cyclic blade pitch it is easy to see what makes it work in the flare.

Otherwise, precision autorotation contests are very popular among RC helicopter pilots. Here is an example of Mitch Marozas doing an inverted piro in autorotation, snapping the helicopter upright and landing it during the autorotation limbo contest at IRCHA.

Practicing autorotations is one of my favorite things to do when I get bored. I’ll sometimes spend an entire afternoon seeing how far I can fly the heli in autorotation, see how many different maneuvers I can make, practice autos on FPV (LOADS of fun), and see if I can land it precision within a 2 meter diameter circle by turning into the wind and use ETL to give me more lift and maneuverability. It is a little known fact that helicopters fly more like an airplane - even the control inputs in dynamic flight, with the exception of managing collective, are pretty much identical to fixed-wing.

With a larger size quadheli equipped with a proper over-running clutch it may be possible to do a “dead man’s curve” auto which results in a controlled crash. I wouldn’t know without trying it, and I don’t know anybody who has.


(Laszlo Kiss) #54

Hello Tridge,

Does this function support Plus quad layout?
Can you explain, how to install this solution?

Thanks
Laszlo


(tridge) #55

not yet, but we can add it if you need it
Right now it only supports X and H layouts


(Laszlo Kiss) #56

It would be really nice.
Is it so complicate to replace the yaw control, with the tilt rotor - vector control yaw method (Q_TILT_TYPE: 2)?

Thanks
Laszlo


(Chris Olson) #57

@tridge with the vibration issue on your test QuadHeli, and having to hang all the weights on the Pixhawk, I can suggest something that has worked well for me on regular helicopters. I use a 4.5mm solid aluminum plate mounted to the frame on the corners (with regular double-sided tape) of the plate. And mount the Pixhawk to the plate on the corners with 3M tissue tape.

The aluminum plate dampens vibration before it gets to the Pixhawk’s case. When I designed my mount for my helicopters I initially mounted the Pixhawk to the plate with two round 3M tissue tapes, about 1.5cm in dia. I still had a little higher vibes than I wanted so on a whim I tried mounting the pixhawk on the corners instead, using four 3M tissue tapes ~1 cm square. I now get vibes around 3-7 average.

The final thing I did was put a covering over the Pixhawk. Helicopters are quite noisy because of geartrain and engine. The noise affects the IMU’s. So one day I spotted this thing my wife has, to put in the windshield of her car when it’s parked in the sun. It disappeared and she didn’t know where it went because it ended up on my Pixhawks. It has a shiny heat reflective suface on it with a thin foam insulation bonded to it, and some sort of waterproof covering on the insulation that prevents it from getting wet or falling apart. I stuck it to the Pixhawk with some double-sided tape and it reduced the average vibes to 3-5. I think just from noise damping.

This is what it looks like on my 626 helicopter:

New photo by Chris & Kristin Olson

So anyway, these are things you can try for a better mount. How well the mount absorbs vibration to prevent it from getting to the Pixhawk’s case depends on the frequency too. By careful engineering and tuning of the mount I’ve been able to get the EKF quite happy on all my heli’s without using rubber ball mounts that let the Pixhawk move around too much and cause issues of their own.


(tridge) #58

yes, I’ve used that before. Adding the lead weights was a quick retrofit, and I was pleased at how well it works.


(Steve Borenstein) #59

@tridge and others,

Woudl anyone be able to help guide me on initial parameters for a Heli-Quad setup for a LARGE (55 lb) H-Frame? I can’t currently get into a lot of details, but each rotor is about 1.2m diameter using trad heli heads, and it’s gas powered. We are conducting intitial startup tests this week and hopefully start autopilot testing next week. This quad is not 3D, so minimum pitch is zero. I am guessing, but haven’t tried yet, that I’ll want to set the trim value to 1100, rather 1500.

Thanks!


(Kyle) #60

@tridge,

I’ve got an old stingray (H-frame), looking to get a pixhawk on it. It’s been awhile since I’ve tinkered with Ardupilot stuff.

What I’ve tried is to use mission planner to upload v 3.5.5 heli to the pixhawk, now I’m unsure how to load the params, is there a way I can upload using the params file? I tried in qgroundcontrol, to load params from file the file you posted from github, but I don’t think it worked, because after I searched for some param names, and they were unchanged