Tricopter V3 guidance needed!

[NHS77]

Today I finished my build of the Tricopter V3 with integrated Naze32 from RCExplorer. Beautiful machine. Buttery video, lovely to fly... except: the yaw drifting. Now David mentions this in his setup video:

Set I and D gains to zero, go fly it and compensate the yaw drift with a trim until it's gone. Go back to cleanflight and transfer the offset value from 1500 in the yaw channel to the midpoint in the servo tab. Re-enter the I/D values, delete the trim on the tx.

This I did. Problem is: To get rid of the yaw my offset value is almost 200! The motor on the tilt is angled at about 25 degrees. Is that normal?

Of course I put the servo in it's neutral position before mounting it. In doing so it may have been turned by 2-3 degrees but by far not that much.

This being my first Tri I'm not sure whether this normal or not. David does mention that if the values are extreme the servo may have to be re-positioned. But if the copter only stays straight with the back motor angled that much, what difference would that make?

Furthermore: I don't if it's related but CF calls for all 3 motors spinning CCW. David suggests to have the front left to spin CW. Did this as well...

Any help would appreciated! Thanks!

 

[Craftydan]

25 degree tilt for a tricopter isn't unusual -- it's got a lot of torque to cancel. Even with the front two motors counter balancing each other, that back motor is spinning like mad and the prop is generating a lot of force. Keep in mind, a 25 degree tilt equates to only a 10% loss in downward thrust (1-Cos(25)).

As far as 200 . . . think of it this way -- full swing typically runs from 1100-1900 ms, over 90 degrees throw. That's a sweep of 800ms total from min to max. 200 is 25% of 800, and 25% of 90 degrees is 22.5 degrees . . . so what you're seeing pretty much checks out.

One think I'd mention as an aside: re-center your servo so the servo naturally centers (1500) at that angle. That way you've got an equal yaw throw in either direction. Centering upright and then trimming to neutral is good for getting a feel of where the servo should sit when it's done, but like any airframe you should do your best to mechanically trim as much as you can, then trim the leftover in your radio.

 

[PHugger]

I had to do exactly what Dan suggested about re-centering the server. On my Fortis, it's a bit of a pain as I had the servo tapped down. It can drive you crazy - you trim it and it doesn't get enough movement to arm!
After re-centering and just a bit of End Point adjustment, my servo is pretty much level when hovering and arms without issues.
I think that CleanFlight has some strange issues with Yaw. This is why David has his own custom version. I'm running the Standalone copy with good success, but CleanFlight did act a bit funny at first.


Best regards,
PCH

 

[PHugger]

Another thing to check is balance. If you COG is off it will act very strange and your motors will not be balanced. They may be very different temps as well.


Best regards,
PCH

 

[NHS77]

Thank you guys!
Glad hear that this is supposed to be like that. Will do the re-centering, that sounds good.

The only thing wonder now though: In David's video he needs something like 40ms to get it right. What's different in his setup? The only difference I can make out is that he seems to have 9" props - I'm running 10". Would the differnece in torque from a slightly smaller prop make up for this?

 

[Craftydan]

Prop size can have some effect -- 10's will spin a bit slower in hover, but they'll have a stronger bite into the air . . . WOT is easy to say what direction the torque will change with prop length, but in a hover? Not so much.

I'd lean toward checking balance as Phugger says -- if the CG is too far forward, then the rear motor will be under-loaded, and the lower RPM may cause it to tilt more to get the requisite counter-thrust.

AUW will also have an effect since the heavier it is the higher the hover throttle will be. If he's flying his power system at the lower end of it's throttle, the motor will tilt less. Load the same airframe down more, it'll tilt more.

I'd also say David has made more than his share of tricopters (He's a great designer, I kinda wish he'd branch out more as of late ). I wouldn't be surprised at all if he baised the servo in the direction he knew it was going to go -- haven't watched the videos , so that's unsubstantiated speculation.

 

[makattack]

I haven't built a RCETricopter V3 yet, but I built by electrohub / toughtilt tricopter based on David's FT video on scratch built tricopters, and in that video, he had suggested a biased servo/tilt setup so I've done that and it's a great flying steady machine. I repeated the same with a HK Trifecta mini tri and am also happy with that. All my tricopters also have props/motors that rotate CCW.

 

[NHS77]

I'd lean toward checking balance as Phugger says -- if the CG is too far forward, then the rear motor will be under-loaded, and the lower RPM may cause it to tilt more to get the requisite counter-thrust.

Thanks for getting back to me on this guys!

The CG Thing sounds intereseting and plausible. So far I've balanced it so COG sits at the cross section of the booms, which should make the load equal on all 3 motors. I'll see what changes if I move it back a bit.

Although, in the meantime I noticed that I also had a distinct drift to the right (not just yaw to the right). Can't really be CoG issue since the battery tray sits straight in the center and there's no chance to move the battery sideways. Not sure where it comes from but I trimmed it out in the accelerometer roll adjustment in CF. Since I've done that I was able to reduce the tilt to about 130ms / 15°. It flies like a dream now, flips on a dime. The best flying experince I've ever had. Really happy.

Only bothers me a bit that I had to do so much in software to get it right, especially the Acc roll trim.