12 May, 2017 at 18:37 #37613
@jukkaaakko: you’re right. The fact that it only corrects for decelation causes drift to one direction.
With the fixed filtering the default value for the correction is high and increases the drifting.
The correction was originally based on throttle, but it didn’t take into account the motor speed changes from pitch commands.
I personally think we don’t even need the correction anymore with the increased low throttle authority.
Need to do some testing when I get home. Have been traveling and still am for the next week.12 May, 2017 at 20:24 #37614
In the FWIW department:
I’ve been flying the 0.5 Triflight algorithms quite effectively in my Dronin branch for a while. I’m not the worlds best pilot, but it has performed real well. The one change I did make is I did not implement the gyro correction, as I could not wrap my head around the reason for it.
Then I upgraded everything to the 0.7 algorithms, including the gyro correction, and the flight performance seriously degraded. To make a long story short, I ended up going back to 0.5 algorithms (without gyro correction) with the addition of the 0.7 yaw boost function. This is flying very well so far, but I don’t have a whole lot of time on it yet, and it probably could stand from a little more parameter tuning. The ability to adjust the dynamic yaw at low and high throttle settings with individual gains seems to work better for me, and it plays nicer with the Dronin autotune feature.
My primary test vehicle is an Armattan mini tricopter, similar in size to the RCE Mini. Mine it’s probably not as powerful as the RCE Mini, but it’s significantly lighter, so performance might be somewhat equivalent between the two. I’ve not done any testing on my full size tricopter with this version yet.12 May, 2017 at 20:39 #37616
I’ve been flying the dRonin version @jihlein has on my original Baby Tricopter with the F3FC Racing FC. His original version based on 0.5 was flying great the last time I had it out. (best it has ever flown) Just started working on testing his latest version after a few weeks off from just about any flying due to wind/rain etc… Need to finish up my 2nd Baby so I can do side by side comparisons to TriFlight 0.7 Beta1.13 May, 2017 at 03:34 #37621
I’d like to try the new version on my Mini Tricopter (latest electronics package and integrated FC).
Are instructions posted anywhere or maybe a least a dump of the configuration for a tricopter like mine?
Larry13 May, 2017 at 15:04 #37625
Hi guys I would like to second LarryTru’s request for a mini tri pid settings dump for those of us who cannot tune as yet?16 May, 2017 at 00:22 #37642
I had the same issue, but then read the instructions for V4, mini and baby and realised that the firmware default PIDs are for the mini, while the PIDs listed in the instructions for baby and V4 are for those models respectively.
I run the RCE electronics pack so the default PIDs on my mini is pretty much spot on, for my noon eye, at least.16 May, 2017 at 21:10 #37657
Hi biggestRCEfan, in post 37274 of this thread, Lauka states that the default PIDS for this Betaflight release are for the baby tri.16 May, 2017 at 21:50 #37658
I can affirm the .7b1 betaflight based triflight default PIDs will get a v3/4 RCExplorer tricopter (based on the rcexplorerF3 board) in the air. The I is a bit too high for my liking for pitch and roll, but it works. P seems ok. Might even be able to go higher if I drop D more. I was lazy and curious how his default tune would work on a larger tri.
I am seeing right yaw on throttle up, so that bug would be nice to have quashed. I flew it enough to just get an in-air tail tune done.16 May, 2017 at 22:48 #37660
Bassmaniac, you are right. This is not consistent with the build instructions. I am running 0.5 from cleanflight though.18 May, 2017 at 23:46 #37674
So ive updated my RCexplorer nazeboard with this update. (mini tricopter)
And after my first couple of test-hovers, i noticed these problems:
Firstly, the tail was really irratic, so i tested to set tri_motor_acc_yaw_correction to 0. That took care of that problem.
I also have quite large amount of drift to the right (roll) and maybe some pitch to front. But the roll is really bad.
Were should i begin to look for a solution for that?21 May, 2017 at 21:19 #37702
@db1, drift on roll? my first guess would be vibrations in the frame? Can you post a BB log? so we can look at the gyro’s.
here is mine, baby tri, flies good except some drift on yaw in turns.
@lkaino, what do you think of these logs, better he?22 May, 2017 at 17:48 #37723
Yes, and i can confirm that this is only in angle-mode. Not rate.
And its quite substantial, so i dont think i could be so off in my calibration.
Sorry, no BB right now. However, iam thinking of installing one. But im not sure if i can have both serial-reciever and a BB at the same time with Naze32 RCexplorer board..22 May, 2017 at 18:58 #37725
If only in angle mode, then I suspect the autolevel pid settings? or a combination, did you lower the auto level term?22 May, 2017 at 19:35 #37726
@db1: In most cases the drift on angle mode is caused by too much vibrations (noise) on the accelerometer. Please check your motors without props and feel by hand if any of the motors are producing more vibrations than the others. You can also try with a fresh set of props.
@bhuism: That looks pretty good. You’ve got some noise on the D-term on lower throttles. You could try lowering the cutoff for the D-term lowpass to 90-95.
What kind of drift are you having? Would it be possible to shoot a video?
I thought that I could do some flying on the weekend but I need to go help saw some logs to planks. I’ll take the copter with me if I would have some spare time for flying.
Would you be up for testing reversing the front motors? Left motor CW and right motor CCW. I have a feeling that might help with some of the issues I faced before.23 May, 2017 at 09:58 #37735
I’m just going to build my Baby Tricopter.
I’ll use the RCExplorer servo with feedback wire along with my AirbotF4 target FC. I’m using HOTT receiver with serial SUMD and HOTT telemetry. Current and voltage measurement are made by RCTimer PDB. Nothing else connected (ie. OSD or GPS, etc.).
Could somebody sive me instructions on how to configure board for 3 motor 1 servo 1 feedback wire configuration? Also, how to connect wires?
Many thanks in advance.
You must be logged in to reply to this topic.