[billyd]

Yes it still flies. Although it’s been gathering dust as I’ve got a lot of projects.

If you read the rcgroups post I linked above it details the setup in detail.

Good luck with the build!

[billyd]

You can download the files at the bottom of this post: https://www.rcgroups.com/forums/showpost.php?p=38845780&postcount=12

[billyd]

Yes but I figured the board might censor that name

[billyd]

I strongly suggest we change to another forum. This one is too unreliable and obviously will not be here forever. Is the one with groups in its name ok for everyone? I will open a thread in the tricopter drones heading. Please consider moving there.

[billyd]

I’ve been doing alot of reading on tuning and also using blackbox and setting filters. As a result I have greatly improved my tune as of 1/29/20. See post #64837 for my latest pids and other settings I have in INAV to make the LR fly great.

I ended up increasing tri_dynamic_yaw_minthrottle to 500 to improve manual tail authority. tri_dynamic_yaw_max_throttle is already at 100 which is the maximum.

[billyd]

Yes absolutely. The other thing is the kakute f4 v2 has the gyros sitting on a soft vibration absorbing pad which most likely is not perfectly parallel to the board. Perhaps this is the cause of the slight discrepancy? Not sure that’s right though because I don’t think it should matter. Just trying to reason out why ANGLE level is not level after successful calibration.

I am not sure why INAV doesn’t just read perfectly level in the first position of accel calibration (assuming the user takes the time to put the craft in perfect level). Other flight software does this. It seems like INAV doesn’t do it, or doesn’t do it well.

[billyd]

@peterpankras Yes it is for precisely that, getting the angle mode the last tiny bit more perfect. It has a big impact on Nav modes quality of behavior.

Sorry I didn’t mean to insinuate you didn’t know I worded that too strongly.

In 2.4 those that develop INAV say specifically that diff all is fine for this upgrade. I guess there aren’t any format changes or changes of user entered settings (outside of the new throttle_idle replacing min_throttle).

[billyd]

@peterpankras ps compass helps with orientation in yaw not in pitch and roll. And the calibration process obviously doesn’t work well enough. The first thing I did was try to recalibrate. I must have run it 10 times. I even put the triLR in a perfectly level position in every orientation. It still wouldn’t fly level in angle mode. This happened with my previous kakute f4 v2 as well. So it’s either INAV or the kakute. Plus like I said why would they provide tiny adjustment capability to the fc position if all you had to do was calibration.

I watch vids online and people aren’t even trying to level the fc in calibration and it still is accepted. I tried that too, same result. I had to use the align fc adjustment.

[billyd]

@peterpankras Obviously if all you had to do was calibration, they wouldn’t provide tenths of a degree adjustment capability on roll and pitch (and yaw). My calibration was done per the instructions and accepted but my triLR would tilt left and fly toward me (tail in) in angle mode. After adjusting alignment of the FC by 1.8 degrees in roll and .8 degrees in pitch it now flies perfectly still in a hands off hover. Also all of the nav modes work 10x better because it’s level when executing the nav command. This is true for pos hold and rth, and most likely true for wp (I haven’t tried them yet)

On another note I think dshot min throttle has been replaced in 2.4 anyway with throttle_idle. From github:

“Upgrading from 2.3.0
Download and install the new configurator
Save to a file the current diff all from the CLI.
Upgrade to INAV 2.4 using the Full Erase option in the configurator.
Upload your OSD font of choice from the OSD tab.
Find on your diff all result the min_throttle line, and replace it by throttle_idle, setting the percentage of the idle throttle. The default is 15.
Go the CLI again and paste the contents on the file you previously create and write save , press ENTER and done!
You should be ready, explore new 2.4 features and enjoy!”

[billyd]

Just flew on 2.4. Everything works great. Best it’s ever been.

One thing that’s very important and I don’t hear much talk about it. But if you want the nav modes to work properly, you must go through board alignment in angle mode. Now you can use the tx for this if you use the adjustments tab, but to me it’s a little complex and not really necessary. I just flew los and kept a laptop nearby, and adjust pitch and roll in the configuration tab to align the FC until the LR flew in a hands off still hover in angle mode. Basically if the craft tilts to the left with the tail pointing at you, rotate roll to the left. If it pitches toward you with tail toward you, rotate pitch back. Keep it up through trial and error until it’s perfect. Took me about 3 tries and it was done. Now all the nav modes behave really well. Do this even if you never use angle mode. The calibration is not enough to get the craft flying perfectly in angle mode. You have to go through this procedure.

[billyd]

Looks good. I hovered it and it flies even better. Navigation throttle issues fixed. I think that will take care of yumemi5k’s problem too.

Noticed a couple of parameters were rejected. Yaw jump prevention and one other. And they added a throttle_idle command defaults to 15. Kind of high I may lower it some.

ps Finally fed up with them, I pulled the emax motors off of my TriLR. Not happy with them. Bearings are garbage. Replaced them with hobbywing 2306 1600kv’s. Much quieter and are better for my triLR’s intended flight profile. The emax motors have a pretty bad rep online fwiw. I think they kind of stink frankly. I’m pretty sure they were the culprit behind the triLR’s vibration complaints.

So far I have revised the motors, tail servo and pdb as the stock stuff all had issues.

Thanks for the 2.4 hex awesome work as usual jihlein!

[billyd]

Yes absolutely I’ll take it. I am also seeing some strange behavior with position hold and the throttle. Hadn’t noticed it before. But I’ve been working on so many parameters it’s easy to miss. 2.4 is supposed to correct this I believe.

ps I increased P and I for yaw even further as I noticed some drift in today’s flying. I edited the post above.

I also lowered I by one point to 14 on the velocity xy in the GPS section. That helped with some overcorrection I was getting in position hold leading to constant small slow movement. Not toilet bowl because it wasn’t increasing. Just a constant little circle not getting bigger or smaller. Lowering the I seemed to help it quite a bit. I might go to 13 and test again.

[billyd]

Back by popular demand, ok maybe just because it got deleted by the duplicate post filter…..

Here are my INAV pids before and after triflight is enabled.

Couple of caveats:
My tricopter LR is heavy. I have a full nav setup and I use a 3sx6200mah lipo. The AUW is 864.3 grams.
Also I am using a 22HV tail servo on 6v. It’s an 8v servo and alot faster than the 210 even when only driven at 6v. I have also removed the stock emax motors and switched to hobbywing 2306 1600kv motors.

Finally this copter is designed for long range flight not freestyle racing and tuned as such.

So with that in mind here are the PIDs that worked well on my tricopter prior to enabling triflight in INAV

set mc_p_pitch = 26
set mc_i_pitch = 15
set mc_d_pitch = 15
set mc_p_roll = 18
set mc_i_roll = 12
set mc_d_roll = 10
set mc_p_yaw = 100
set mc_i_yaw = 23
set mc_d_yaw = 1

After triflight was enabled and after I ran airborne autotune, I had to increase the P in all three axes. Here is my post tailtune/triflight enabled INAV pids and other settings: (revised 1/29/20)

set mc_p_pitch = 35
set mc_i_pitch = 50
set mc_d_pitch = 40
set mc_p_roll = 30
set mc_i_roll = 60
set mc_d_roll = 35
set mc_p_yaw = 195
set mc_i_yaw = 90
set mc_d_yaw = 1
set max_angle_inclination_rll = 700
set max_angle_inclination_pit = 700
set yaw_lpf_hz = 30
set dterm_notch_hz = 125
set dterm_notch_cutoff = 90
set nav_mc_vel_xy_i = 14
set nav_mc_heading_p = 65
set tpa_rate = 20
set tpa_breakpoint = 1650
set roll_rate = 30
set pitch_rate = 30
set yaw_rate = 30

My tailtune results: (revised tri_dynamic_yaw_minthrottle 1/29/20)

set tri_dynamic_yaw_minthrottle = 500
set tri_dynamic_yaw_hoverthrottle = 1405
set tri_tail_motor_index = 2
set tri_servo_max_adc = 2751
set tri_servo_mid_adc = 1918
set tri_servo_min_adc = 1100
set tri_tail_motor_thrustfactor = 112
set tri_tail_servo_speed = 634

Hopefully these will work well as a starting point, but again if you tricopter is much lighter it might not help much.

Attachments:

1. 

   tricopterLR_INAV_p5Triflight-1.jpg
   

[billyd]

FYI I just saw on youtube INAV 2.4 was released today and one of the things they changed/improved was low throttle on navigation. So that might be directly related to the issue you had…

[billyd]

@yumemi5k Go over your advanced tuning settings carefully. You may have something set wrong in there.

[Author]

Posts