Debugging the tricopter mini racer

[alexandervdr]

Hi Guys (are there ladies?)
I am new to this forum, but have been playing around with self designed tricopters since 3 years as a sidekick to my very intensive experimentation with RC helicopters.
Not sure if it is relevant, and maybe it has been mentioned before, but noticing the servo specs used for the rcexplorer tricopter it’s a very slow one. Especially in ‘heading hold’ PID systems this has a big impact on accuracy because it induces delays in the system loop (the more that the ‘yaw’ function has more mass inertia than roll/elevator making things worse). Mounting a high speed servo makes a huge improvement in control system performance (talking .06sec or less for 60°, versus double for the BlueBird servo recommended). In RC helicopters a high speed servo makes all the difference for the tail/Yaw control (100% use them). And on my own tricopter it does too. My 2C…

[lauka]

Triflight 0.4 Beta 2: https://github.com/lkaino/Triflight/releases/tag/0.4-beta2

Changes compared to 0.4 Beta 1:
– Critical bugfix for TAILTUNE mode on bigger tricopters.
– Improved TAILTUNE speed on bigger tricopters.
– Luxfloat integrator disable limit increased from 30 dps to 40 dps to avoid drifting on very windy conditions.
– Increased default I gains for pitch (0.4), roll (0.4) and yaw (0.8). Now more suitable for non-acro and LOS flight.

NOTE for bigger tricopter users! If you are using Triflight 0.4 Beta 1 on a V3 or other bigger tricopter you must upgrade to Beta 2 before you can use TAILTUNE.

The default PIDs work well enough on V3.5 as well. Enter following to CLI to calm the controls and increase the swooshiness :).

set rc_expo = 70
set roll_rate = 11
set pitch_rate = 12
save

I have a feeling that @David will post setup video very soon :).

[Patrick]

Hi, had someone a stl for the Servo setup tool – or could give me the dimensions than i will create one. Thanks!

[lauka]

Triflight 0.4: https://github.com/lkaino/Triflight/releases/tag/0.4

I’m liking the performance so much that decided to make 0.4 Final now. I get zero drifting with the default I gains. They are a bit too high for my liking, I term is fighting back a little in some aggressive freestyle/acro maneuvers now, but at least it is not drifting. I suggest lowering the I gains at least to half when ripping balls :D.

Noticed that the default servo PWM rate was 50 Hz, changed it to 250 Hz. This could be a reason for poor performance for some users. Didn’t notice it before as I haven’t done full chip erase for my copters in a long time.

Changes compared to 0.4 Beta 2:
– Changed default servo PWM rate to 250 Hz.

Tune for BMS-22HV 8V:
Finalized my tune for the super fast 22HV servo on the mini tri. I had to enable the servo low pass filter with a very low cut-off frequency. My current assumption is that the fast servo throws the unbalanced mass around so fast that the PID must fight itself when making corrections. This kind of system is prone to oscillations and makes the yaw very challenging to tune.

Paste this to CLI if you’re using the BMS-22HV servo with 8V supply:

set tri_tail_servo_speed = 1000
set servo_lowpass_freq = 18
set servo_lowpass_enable = 1
set p_yawf =  2.300

[LeoTheHuman]

I would warn everyone using analog servos here. They should change the default servo PWM rate back to 50 Hz before powering the servo. There might be damage to the servo otherwise. It is not very likely on tricopters, but it might happen.

[mcla118800]

Hello all,
Currently rebuilding my tricopter and was wondering if there was a way to port the algorithms for multiWii, as i have a multi Wii board and the tricopter support is not fantastic.
thanks
amazing work this far

[SMJ]

Hi Guys,

I’ve made my own tricopter and I’m loving all the excellent work being done here by you guys.
I’ve got it flying pretty well in LOS so today I took it for an FPV.

Seemed to go pretty well at the start but after a while the yaw started acting weird.
I had a look in black box and the I term seems to be really winding up?

I’m now at a bit of a loss of what to adjust next?
Can some one please have a look at the attached and point me in the right direction of which values to tweak.. 🙂

I’m using a Hitech HS-81 servo.

Thanks Guys!

Blackbox log here…
https://www.dropbox.com/s/5l26ltr92teb6df/blackbox_log_2016-01-02_152045.TXT?dl=0

Attachments:

1. CLI-Backup-tricopter-20160102.txt

[lauka]

@mcla118800: MultiWii is 8-bit IIRC, it would be huge work to port the code there.

@SMJ: Thanks for the log! It seems that your yaw rate is quite high and you might be asking too much from your copter. That might be the reason why the I term is winding up as there is constant error when the copter can’t achieve the rate you’re asking for. Or I might be completely wrong :). Was the default yaw_rate of 50 too slow for your copter? Please try with the default rate or at least a bit lower rate.

What tricopter frame are you flying?

If that doesn’t help I could make you a special build for testing without the integrator disable function.

[Patrick]

http://www.thingiverse.com/thing:1239154 Servo Setup Tool – hope it fit … For the guys that have an old V3 Kit …

[Patrick]

Sorry i miss the original stl in the post here http://rcexplorer.se/blog/2015/12/triflight-setup-guide/

[mcla118800]

thanks lauka,
i might work on something to optimize multiwii for tri
thanks for your help

[LeoTheHuman]

It might not be the best way to spend your time. Naze32 is only ~$20. (:

[SMJ]

@lauka:

Thanks for the tip on Yaw Rate!
I’ll give it a go today and report back. 🙂

I’m flying a frame I designed myself.
Here is a picture but this is David’s thread for his awesome mini tri so I’m not going to go into mine here.

I’m loving the fantastic development going on here guys. well done!

Cheers

Attachments:

1. 

   180-Tri.jpg
   

[mcla118800]

yeah i know, but in some circumstances multi Wii is cheaper and flies better.
(the flip 1.5 is only 12$)
aidan

[SMJ]

Hi Lauka,

I tried dropping the yaw rate to 50 with no luck.
I also dropped the i term to .005 and it was still the same just to a much lower level.

I like the rates up pretty high to make it snappy (and fun). 🙂

The pitch and roll i term seem to be really good and seems locked in which is good.
The yaw i term starts OK but seems to build up over time. (After a couple of laps of the race track)
I’ll have to do some more research as to what can cause this (Other factors)

Cheers!

[Author]

Posts