Servers by jDrones

Tradheli from a beginners POV

arducopter
tradheli

(Jakob Schmidt) #21

So after 2.5 week of almost solid rain, we finally get a sunny weekend…except it’s blowing more than 30km/h and gusting even more.
However, with daylight savings, it’s now light enough after work that I can get some flying done! No wind and while orientation was a little tough with the sun in front, it didn’t matter too much for what I had to do.
My mobius misbehaved again,though.This time a loose micro SD card, but I got my iphone setup for the second half. (Until that ran out of battery).

Much better than last time. I had increased P on both pitch and roll and it was much more responsive. I cranked it up twice, decided that it was getting just a little twitchy (No oscillations, though) and dialed it back down a little.
Loiter also seemed better, although it tended to drift and then catch itself whenever yawing?.

The other issue was also that I needed to add forward + right stick in order to prevent it from drifting when in stabilize. I initially thought that maybe the accelerometer wasn’t calibrated correctly and was trying to find a level surface. to calibrate it on, hence the frequent stop & go on the video. It was only after I finished, that I realized that I did not re-level the swashplate after I removed the trim on my transmitter!!

Other issues/lessons learned:

  • I need to set the collective so that hover in stabilize matches hover in loiter. I almost crashed it into the ground towards the end when switching back to stabilize.
  • Keeping neutral collective when spooling up greatly reduces the chicken dance.
  • Batteries are sagging more than I like under load. I wont buy new ones just yet, but I should probably start planning for it.
  • I do like the simplicity of operating a heli vs a plane: Remove blade holder, unfold blades, mount lipo, press safety switch, go. No fiddling with mounting wings, extension cables, etc etc.

Weather forecast is again looking crappy for the next week or so, but next time I’ll go to a more open area and start experimenting with some auto-missions


(Chris Olson) #22

Just set the IM_STAB_COL params so the hover collective is between IM_STAB_COL_2 and IM_STAB_COL_3 and that will make the switch from Stabilize or Acro to/from the altitude controlled modes far less dramatic.

I never did use a fixed-wing because they tend to get bounced around in the wind a lot, and not usually a suitable landing spot in the places I fly. Although I did experiment with one commercially built fix-wing and was less than impressed with it. For the two years previous I used multi-rotors and was not happy with their performance. I finally bought a piston 700 flybar without even a tail lock gyro on it, went to an old-timer in the RC club who has been flying helicopters for 35 years and asked him if he’d give me flight lessons and teach me how to fly it. I’ve never looked back. After I learned how to fly it, I set it up with a Pixhawk and tried it in the field. It was like, WOW - this thing makes a multi-rotor look like a toy!

The downside to helicopters is that they have a steeper learning curve, requires a pilot with more experience/training, and are more mechanically complex and higher maintenance. Operating cost/acre is probably higher than fixed-wing. But the helicopter can get the job done when a fixed-wing can’t fly due to wind. And the heli is so stable that I just use a fixed camera mount. I tried a gimbal but it didn’t work. The gimbal couldn’t handle the wind loads at 50mph flight speed, nor could it handle the G-forces in the turns. Those confounded rubber ball dealies that isolate the gimbal gyros from vibration gave out within 8 flights and the camera ended up dangling by the safety straps. I decided that was too much machinery to have hung out in the breeze on a helicopter coming out of the last waypoint on the turn right on 50mph, lined up on the next pass :grin:

Your project is looking good!


(Ktrpilot) #23

This looks cool, I’ve just finished setting up a Rex 500 with a 4 blade head on it, a rebuild after an epic crash when a failsafe didn’t quite do what it was supposed to. Stock params and initial testing looks really promising a little soft in the control response ( because of low head speed) but no sign of any wobbles and logs look pretty tight. Running a head speed of about 1650 at the moment but that could probably be bumped up to about 1800. I havnt done more than just hover at this point and I have a feeling that at that speed the tail might struggle to hold, a slightly higher head speed should also increase the response. Chris I like what you have done with your landing gear, I had a set that I made before I crashed lwhere the heli sat on its original gear in a cradle and was held in place with Velcro, it worked well but was pretty heavy… your way looks cleaner and probably reduces the weight quite abit, compared to what I had.


(Chris Olson) #24

On my Trex 700 nitro I stacked plates like Jakob did, and then broke the mount when I lifted a 24lb load on a 15 foot tether to the top of a 190 foot elevator leg with it. So that was the first one I built the heavy duty aluminum landing gear for.

Never weighed that gear, but I don’t think it’s any heavier than stacking carbon fiber plates to use the stock gear. And the aluminum is way stronger. I can attach a four-point lifting tether to the corners of the underneath cradle on the 700 and easily lift 25lbs with it with no flexing of the main frame or gear.


(Jakob Schmidt) #25

Carbon is about 40% lighter and 10x stronger than aluminum. Aluminum is obviously much easier to work with, especially when doing non-flat shapes.
Apart from some initial concepts, I haven’t done any work on mine yet,


(Chris Olson) #26

That common comparison is gram for gram of material, and not real world cross section of parts design. The problem with carbon fiber is that it does not stand up to tensile stress well because it’s too brittle. Aluminum does, and will easily take repeated shock and impact loading, where carbon fiber will just shatter.


(Jakob Schmidt) #27

Which is why you add a layer or 2 of kevlar. A thin layer of rubber/plastic will then be enough to protect contact points from shattering.


(Chris Olson) #28

Cool. My 600 must not have had that kevlar in it. Had a mishap switching flight modes and it totally disintegrated hitting frozen ground hard on the landing gear. Once the frame sides shattered and the main bearing blocks came loose it self-destructed. I got a carbon fiber frame side plate for a pattern and I’m having new .060" 6061T6 frame sides waterjet cut for it. That’s what my 500 has is aluminum side frame plates. I sort of prefer that - if it has a mishap and hits hard, simply take it apart, hammer everything flat and straight, put it back together and go flying again.


(Ktrpilot) #29

After the last crash I had, I seriously considered retiring it and simply upgrading to a larger bird. It is the first of the 500 sized dominator series that came out and parts for it were surprisingly hard to find. I’ve been flying FPV/ setting up for amateur AP with it and I’d always intended that when I finally got a 500 flying properly I’d upscale to something bigger. But I was always interested in seeing how a 4 blade head would work out so went that way instead. I’ve changed a lot of other bits and pieces out as well including changing to a minim OSD and adding full 3axis head tracking to the FPV cam, I’ve gone away from using go-pro to a run- cam for a gimballed camera so I have the ability to start and stop video recording or taking pictures during flight from the Radio, it’s way lighter and in my opinion the difference in picture quality is negligible unless you have a picture from each next to each other. I’m going to upgrade the tarot gimbal I have to a 3 axis one but still deciding which one yet, then I can have it piggy back off the head tracking also. Im at the point where I need to make an extended set of landing gear to accommodate the gimbal and trying to figure out how I’m going to mount the gimbal to It.

Chris I was reading another thread where you were talking about adjusting IM_Stab_ COL to adjust where the hover position is as opposed to adjusting the H COL MID and a few major light bulbs went on with regard to a few issues I was having before the Rebuild so I’m looking forward to having a play with that as well


(Chris Olson) #30

Yeah, I don’t remember where I got that from, but I think H_COL_MID should always be set to true zero collective because the code uses that in some other calculations for proper flight so the flight controller “knows” where real zero collective is. There are instances in wind where ETL will cause the heli to “float” even at below hover collective when it’s landing, and the flight controller may want to use negative collective to make it come down. If it doesn’t know where that is, some other things like your minimum landing collective may not work and could possibly cause a disarm before it gets on the ground because it’s not coming down at the collective pitch the flight controller thinks it should, so it gets “fooled” into thinking it’s landed.

I know some folks adjust that to have some positive collective at H_COL_MID to get the heli to hover at mid-stick, but based on my experience with it, I don’t recommend it.


(Ktrpilot) #31

It made a lot of sense when I read it but it didn’t click until I’d read it, like I said I had a few light bulbs come on. I’d never been able to really set the bird up to hover at mid stick by offsetting H_COL_MID it always increased alt in an autonomous mode at mid stick and subsequently then always rapidly descended upon switching back. In hind sight I also don’t think I took into account properly the fact I had asymmetric Blades. The second thing was, although the desired and actual attitude followed closely with little error as far as trend went, they were always offset so they never over lay one another if that makes sense and I could never figure out why and I have a feeling now that the offset H_COL_MID would have had a lot to do with it. I’m away at work at the moment, for the next two weeks but I can’t wait to get home and keep tweaking and see what I come up with


(Jakob Schmidt) #32

I went through my swash plate setup again and did find that while it would be level at 0, it would not be level at max & min. By adjusting the length of one of the links, I got it level in roll and got it reasonably close in pitch and figured I could tweak the rest from the trim setup in the heli menu.
I got it out again today and while I got the pitch issue mostly sorted, it still wants to pull to the right.
You can see in the log that I get pitch trim sorted, but the roll did not want to play ball and there’s a huge between roll/desired roll.
Not quite sure what to try from here. Mechanically everything appears straight (at least to my inexperienced eyes).
I also played with the hover roll a bit and that had little influence.
Log here, if anyone fancies taking a look

On the more positive side, I finally figured out how tight to tighten the main blades to stop the chicken dance at spool up and are slowly getting more comfortable with the whole thing.
No video and I have now declared my mobius officially dead.


(Chris Olson) #33

Yeah, the blade holder bolt torque is kind of important. On the bigger heli’s it has to be pretty tight because if one blade lags during spool it will tip the heli right over. On the littler ones they can be looser, but if they lag during spool up they should be adjusted the same amount, and fly out and find their centers at the same time so as to not get a lot of vibration.

I have always adjusted the AHRS trim values to get my heli’s to hover pretty much hands-off in Stabilize. I don’t know if that’s the right way to do it, but I’ve never had a helicopter fly bad because of it. And I think my logs show some difference between desired and actual too. But it always follows and I think it’s because my Pixhawk is not mounted perfect in the frame.


(Jakob Schmidt) #34

Aha!
Turns out that for whatever reason, there already was some AHRS trim values that could be the culprit?
Not sure where they’re coming from. I may be able to get out later today, if not I got tomorrow booked off (from wife) for flying!


(Chris Olson) #35

Yeah, and I don’t know how those are calculated. I think it’s done when you do the accel calibration and the software tries to compensate for what it senses the error to be as far as Pixhawk being mounted in the frame. But I’ve always adjusted 'em. Positive values will make it roll right and nose up. Negative the other way. Using those, I’ve always been able to tune it right in perfect. .017 radians is 1 degree of change.


(Jakob Schmidt) #36

I think it’s time for an update!
After some 4 hours of flying time, I now have a very nicely flying heli!. It was a rather steep learning curve, as while I have a fair amount of experience with pixhawks/ardupilot, it’s all been fixed wing and my only rotor experience (apart from previous mentioned MCPX) has been racing quads.
So, not only did I have the mechanical side of a helicopter to deal with, but also a tuning process which was a bit finicky, at least in my untrained hands. It also didn’t help that I simply didn’t know what to look for in terms what was acceptable behaviour.
However, thanks to the extremely helpful people on the tradi-heli group who kept pointing me in the right direction and I eventually ended up with a heli doing this:

I flew a longer mission today and it was rock solid. (But, the balance lead from the lipo was right in front of the camera, doh!)…but a friend captured some of the earlier tests from his Mavic:

Here’s a short list of my main lessons from this:

  • You need to study both the documentation and the forums. I know that there’s some conflicting information, but read and compare to your own hardware.

  • Make sure that your helicopter is in good mechanical order . Get all the links straight and spend as much time as needed getting the swash plate and main blade pitch evened out. Yes, it will fly with an imperfect setup, but if you are starting out like me, it really helps eliminate mechanical issues.

  • Get it tuned for stabilize first before attempting loiter/auto. I know it’s soo tempting, but mine got pretty wonky in loiter a couple of times (before I knew what I was doing).

Again, a shout out to the helpful people on forums.


(Jakob Schmidt) #37

So, the moment I had completed the first auto mission, I got started with the new landing gear.
I wanted to keep the process simple, even if it then meant a less-than-perfect result.
I made a mold out of Dollartree foamboard, applied a thick layer of epoxy mixed with filler, which was then sanded out and then waxed.
The layup was 2 layers of carbon, 1 layer of kevlar, 1 layer of coremat and then 2 layers of carbon. Then covered with packing tape as a poor mans release film as I wasn’t convinced the mold would hold up in vacuum (and also that really did not want to mess around with it

And they came out reasonably nice!

I had mocked up an aluminum version which weighed in at 64 grams each and these came out at about 34 grams each. A vacuumed version would be even lighter.

And hey, I can even fit the canopy now!

While I do have move stuff around a fair bit, this does mean that I can carry loads more battery, a gimbal (or mapping camera) and still have room to spare!


(Jakob Schmidt) #38

I finally got my first custom canopy fitted. Made a few mistakes along the way, that will be fixed in the next version.

I then also made a Pan & Tilt turret for the FPV camera. Using a Runcam Micro Swift, I could keep the diameter to 40mm and keep it fairly compact.

This was then synched up to the Storm32 Gimbal Controller. Now, one of my pet peeves with the gimbal functionality in Arducopter, is that the max speed for using rate mode is quite slow, so instead, I ended up writing a lua script that creates a rate mode on the Taranis, which means I can set the gimbal joystick speed to 0 (direct mode) and also use that to drive the Pan & Tilt.

Now, if only there was a way of switching from RC control to FC targeting on the fly…

Weather has been terrible for the last 6 weeks, but should finally be clear this weekend.


(Jakob Schmidt) #39

So, over the winter I have managed to get a flight here and there and it has just confirmed how good a platform arduheli is.

I have now run a few mapping missions in ideal conditions, but today was blowing a bit (19kmh/gusting 39km/h according to the met office) and yet the Protos 500 flew it without any problems whatsoever at 10m/s. (which is the current practical limit for the camera I got mounted).

I’ve done a fair amount of experimenting with fixed wing mapping and while I got some decent results, I never got anything like that in gusty conditions.
Granted, a fixed wing mapping plane (talon) would roughly fly at around 160watts. My Protos flies at ~330watts, but I got VTOL capability and very high wind resistance. There’s no assembly, no launching/landing drama…and they sound much cooler.

The wind then died down and I decided to do some more speed test on a tight figure eight.


I have the speed mapped on channel 6 and increase it through the mission.
Running a headspeed of 2100rpm and still got it up to about 50km/h. It does struggles through that last turn, but I’m still impressed how it fights it way through it. I then play around in loiter for a bit, but unfortunately my Storm32 has (again) lost it’s settings and is handling pretty rough.

Currently finishing off a 600 sized build and I’m very excited about the future of Arduheli.