Servers by jDrones

EKF blowout cause?


(Chris Olson) #20

That is an excellent setup to adjust the CG. Really cool looking canopy too :grinning: I LIKE that. The stock canopy on most of these heli’s doesn’t allow much room under it for stuff.

I think you’ll notice that when you yaw the heli in Loiter it won’t tend to move around in a circle like somebody is swinging it by the tail with the GPS mounted in closer to the mainshaft. At least not as much as they tend to do when it’s mounted way out on the tail boom.

You don’t get any twitching on your rudder servo with the MavLink radio mounted there? I’ve gotten servos to twitch before just running the MavLink wiring alongside servo wires when the servo’s don’t have any ferrites on the servo cables.


(Jakob Schmidt) #21

Yeah, the telemetry was just supposed to be temporary, while I got things set up, but much to my surprise it hasn’t caused any issues there.


(Jakob Schmidt) #22

Hrmph. The new VTX477s appear to be a lot less efficient that the Edge blades. Vibrations are up too.

Since the Rails don’t come in 470mm, what would be your second choice?


(Chris Olson) #23

Do you have enough tail clearance to swing Rail 516’s at a little lower headspeed? They’re only 1.5" longer than what you have and they are designed for 550’s. If you look thru the internet forums and such you’ll be hard pressed to find anybody that ever flew Rails that had anything bad to say about 'em.

I got three sets of 'em here now on different heli’s and have gotten the same results on all three. Just got this one flying yesterday with 626’s on it. I’m making a different mount for the Pixhawk and still have to put a decent landing gear on it. This is a 550 with the E5S stretch kit and it could swing 696’s if I wanted to, but I’d have to put the E7 main gear in it to fly 696’s.

New photo by Chris & Kristin Olson

(Jakob Schmidt) #24

No, 480 is the max, I"m afraid. I did find a stock of Edge blades in Germany and shipping is reasonable, so I may order from there.

As this is my ‘test frame’, I’m also keen to try out a 3-bladed rotor and have my eyes on this:
http://www.cnchelicopter.com/flybarless-trex-500-etc-500-size-helicopter-cnc-head-3-blades-system/

It’ll have to wait a bit, though. I’ve decided to do a complete re-tune of the Protos (With the old Edge blades). It’s annoying as it was flying really well before the GPS started freaking out, but on the other hand, it’ll be good to do it again.


(Chris Olson) #25

I don’t like to bash any companies but I have had reports of poor quality from the scale builders on that one. There is also an Align head available for thee blade rotor for both the 700 and 500 that the scale guys have spoken highly of. I don’t know if Align makes it anymore but you might be able to find one on eBay or something. Even though US dealers don’t carry it I think it’s still orderable from Align’s Taiwan website. The Align one is more expensive but it uses the same blade holders, thrust bearings, etc as two-blade, just with different feathering shafts and center section.

The scale guys get pretty excited (not in a good way) if they got a head that has slop in it and the helicopter doesn’t handle right.


(Chris Olson) #26

Jakob, here is the one that is highly praised by the scale guys but it is for a 550/600, not a 500, so it has a 10mm mainshaft. And it is in stock in the US Align store.

http://www.aligntrexstore.com/550E-Three-Blade-Rotor-Head-H55H004XX_p_4475.html


(David Ardis) #27

Does this also apply to the Align Blade on a 450 and could this also be part of the reason for the rubber bandy tail :grin:


(Chris Olson) #28

It could be part of it. On the flybar Trex’s, higher headspeed usually fixes that “rubber bandy” tail effect. Take a look in your logs and see what the elevator servo is doing when it pitch bobs. Graph the actual pitch attitude vs RC3 OUT. If the graph show the elevator servo is over-reacting and actually causing it due the inherent instability of the head, speeding up the head to make the head more inherently stable will magically “fix” it.

If the elevator servo is doing nothing, then you got a flybar setup problem, head out of balance, blades not tracking, etc. Setting up a flybar head still remains more of an art than science. Tracking a three, four or five blade FBL head is even more of an art. I’ve spent a whole day before helping one of the scale guys track a four blade head on a Bell scale model until it flew with absolute precision. We used a high frame rate video camera with chrome tape on the blades so we could see which one was causing the issue :grinning:


(Chris Olson) #29

Jakob, if try the method I outlined for tuning, please report back on that. That is the method Leonard recommended for tuning. It has worked on every FBL heli I’ve tried it on. No drama like the wiki method because you always have control of the helicopter. And cutting the D and P gains in half from the oscillation point has yielded a superbly stable heli in cruise flight where porpoising due to over-tuned P gain can quickly lead to loss of control.

Leonard actually recommended setting I gain equal to P gain. But that was not enough I gain in my experience for heli’s for high speed flight. So I set it equal the FF instead, which then gave excellent results.

I have not heard any reports from anybody else that has tried that tuning method, and not enough heli test pilots on the project. If it works for you like it did for me, I think somebody should consider updating the wiki on tuning of trad heli’s.

I even tried it on a APM with 3.2.1 with my 500 that I set up for pylon racing for the Memorial Day Weekend races at the RC club, and it “just worked”. With a lighter 2200 6S battery in it I later flew this heli at 70+ mph running wide open at 2,900 rpm practicing for the races. Banked over on her rotor tips in the turns at full collective pitch set to 15 degrees it didn’t do anything unexpected using that tuning method.


(OAPpilot) #30

Hi Chris
Sorry I have not been back, as the ESC packed in, then had problems with the swash set uo again.
Well I got it sorted out and did your setup this afternoon. It was a bit windy but I thorght I would give a go. All seemed good.When the wind drops I will do a log and send it.
One thing I noticed, When I pressed the button on the Pixhawk and the servo’s became active, The Swash was going the wrong way. Like left was going right, and forward was ging back. When I armed the Heli they corrected themselves.
I’ve attached the Param’s file for you to look at.

2017-08-23 16-24-03.log.param (13.5 KB)


(Jakob Schmidt) #31

That was the plan :). Although I must say that halving the values seems drastic? I’ll try it out anyways, hopefully tomorrow.

Thanks for the heads-up on that rotor head. Unfortunately the Protos has an 8mm shaft, so the 550 size wont work and they don’t appear to make a 500 sized any longer. I did find a couple of other 3rd party brands, but it appears that more reading on the scale forums is required.


(Chris Olson) #32

That’s the way Leonard said to do it. And it sure worked here :grinning:

You could do like I did with my 500 stretch. I made my own mainshaft that fits the 500 bearing blocks but steps up to 10mm for the swashplate and head so I could swing the 550 blades on it. If you want to use 500 blades use some spacers in the blade holders. It’s about a one hour project to make a mainshaft on a bench top lathe. I made mine from a piece of chromoly rod that I got from the local steel shop and it’s solid shaft, not hollow.


(Jakob Schmidt) #33

I do have a friend with a lathe :). In that case, I would probably go with this:

One more question regarding tuning; Do you start with all values at 0 (PID + VFF)…or do you have a small value in all fields?


(Chris Olson) #34

That might already have a 8mm shaft being for a 500? And those SAB blades have gotten good reviews too.

I set them all to zero. I started with the VFF at .15 on both, and tuned the tail first. The heli flies fine on VFF .15. Then pull the microSD out and look at the Rate to see how it tracks. You have to make some stick movements get enough info to get a good idea of how the rate is tracking. That will be totally dependent on your mechanical rate too, because rate control is rate control, regardless of whether it is mechanical or electronic. It’s whatever it takes to get the right rate so actual tracks desired.

The heli may get way more responsive than you want. That’s where you use the Accel values to slow it down but 110000 is usually pretty close for a 500.

Then do D gain first, P gain next, and I gain same as VFF. Leonard did not recommend doing it the way I did where I set pitch to very slightly less than roll. That was my own invention. He recommended doing one axis at a time. I tried it on pitch and got to ridiculous values without ever getting the shakes. Since I like a little less damping on the pitch I set it a notch lower than roll. Normally that would be a notch higher than roll. But that’s personal preference because I like 'em to lean in smooth and flare smooth, and just don’t like too much damping on the pitch.

Leonard’s tuning method, tuning the D gain first, seems to allow use of D gain. Most people leave it at zero because the current instructions say to over-tune the P by not turning it down far enough from the oscillation point, so adding D gain puts it into instant shake. The heli actually flies better with some D gain.


(Jakob Schmidt) #35

Unfortunately, the Goblin 500 also uses 10mm main shaft. I was looking at it earlier.


(Chris Olson) #36

Hmmm… then the blade holders might also be for 550/600 blade roots. But it’s no big deal to punch the bushings out of 500 blades and put 550/600 bushings in, and use shim spacers to take up the gap on top and bottom of the blade root to swing 500 blades.


(Jakob Schmidt) #37

Well, it comes with matching blades, but I do think that they’re standard ‘500’ blades.


(Chris Olson) #38

Yeah, 500’s usually have like 425-450’s so they have ~970-1,000mm main rotor diameter. 550’s use 500-550 blades. The Goblin 500 has 500’s with 1,136mm main rotor. So it’s actually a 550 class and that’s why it has the big mainshaft. If you don’t have enough tail rotor clearance to swing 516’s it’s unlikely you’d be able to swing 500’s either. So you’d likely be looking at a mod to put shorter blades on it.


(Jakob Schmidt) #39

Right, but this kit comes with 465mm blades :).

Anyways.
Prepped for tomorrows (pending on how quickly I can get out from works) tuning session.
Zeroed PIDs and then played with the VFF.
I previously had a VFF of 0.03 on both.
I ended up with 0.03 on pitch and 0.15 on roll?!?.
I assume I’m looking for movement rates that more or less matches the stick movement?