iNav for KakuteF4V2 based Tricopters

[yumemi5k]

Has anyone flown and tuned yet? I tried Triflight0p7 on inav 2.2.1, bench and flying tailtune completed successfully, with following results.

# get tri_
tri_unarmed_servo = ON
Allowed values: OFF, ON

tri_motor_acc_yaw_correction = 6
Allowed range: 0 - 200

tri_motor_acceleration = 18
Allowed range: 1 - 100

tri_servo_angle_at_max = 400
Allowed range: 0 - 400

tri_servo_feedback = RSSI
Allowed values: VIRTUAL, RSSI, CURRENT

tri_servo_max_adc = 3145
Allowed range: 0 - 65535

tri_servo_mid_adc = 2281
Allowed range: 0 - 65535

tri_servo_min_adc = 1423
Allowed range: 0 - 65535

tri_tail_motor_index = 0
Allowed range: 0 - 7

tri_tail_motor_thrustfactor = 64
Allowed range: 10 - 400

tri_tail_servo_speed = 272
Allowed range: 0 - 1000

tri_yaw_boost = 100
Allowed range: 50 - 700

The copter punches up somewhat more straight than before, but I still can’t get tail to feel locked-in: It either oscillates at higher I term, or stays smooth but can yaw on itself for up to tens of degrees if a lower I term is used during hover. I’m currently on 85/10/30 on yaw P/I/D, which I assume I’m doing something very wrong. It would be helpful if someone has a better tune and could share.

[yumemi5k]

I have since adjusted yaw PIDs to 90/18/10. The lower I term wobbles much less, but takes quite a while to kick in, in extreme cases (e.g. strafing left-right) allowing up to around 45 degrees of yaw drift before stabilizing.

I am wondering if it’s how tricopters are meant to be, though: The tail provides some degree of intrinsic stability in forward flight, and a loose I term allows the craft to “ease” into a turn, like david has mentioned. The response could be perceived as less robotic and more “natural” compared to a perfectly locked-in quad. Heading hold mode can be activated during LOS landing, when a steady heading is more important than steady footage.

Anyway I’d say I’m now more or less happy with how TriNAV flies now. I’ve also tested nav poshold and it works nicely.

[PeterPankras]

Maybe a noob question,

But where can I find this TriNAV firmware if i want to give it a try?

@yumemi5k do you use a GPS with or without compass?

[yumemi5k]

TriNAV is a nickname I gave @jihlein’s iNAV branch with Triflight routine :D, sorry for the confusion!

The git repository is here: https://github.com/jihlein/inav ,branch in question is triflight0p7. I believe you need to set up development environment and build it yourself for the time being. If you are not familiar with these I can build one for you – no warranties whatsoever though!

PS: I use a BN-880 gps with compass unit. Orientation in inav is CW270FLIP when connector is aft of GPS ceramic antenna.

[Wally]

Hello:

What’s the status of this branch with respect to the iNav devs from the first post?
https://github.com/iNavFlight/inav/pull/4861

Looks promising, but I’m not a coder & not familiar with the environment & verbage.

Thanks

[PeterPankras]

Did anyone ever fly a tri with Inav stable?

I see they have a Tricopter setup in the options.

What is the big difference between basic Inav and TriNav?

What happens if I try basic iNav?
Does it have no tri code or something? Will the tricopter be harder to handle?

just like to know as I’m looking for something stable with all gps features enabled.
What about Arducopter?

[jihlein]

Not sure what you are referring to by TriNav, and by iNav stable I take that to mean the stock build from the repository without any modifications, that is basic iNav.

I’ve flown iNav both with and without the triflight algorithms on the Tri LR. I didn’t spend a lot of time tuning it, but it appeared to fly just fine either way. The issue is with the KakuteF4V2, you have to use the modified KakuteF4V2Tri target, the basic target will not work. This is due to the timer allocation in the basic target not being suitable for tricopters, since you can’t choose a separate output protocol for the motors and servo.

The big issue for tri’s in general is just because the firmware (iNav/BetaFlight/CleanFlight, etc) supports the tricopter mixing (not triflight algorithms), it doesn’t mean the flight control board can be used due to poor timer allocation.

Depending on the FC chosen, BetaFlight may be able to get around this with resource mapping. For example on the KakuteF4V2 resource mapping can be used to swap the LED pin and M4 pin, to obtain favorable timer mapping. But no GPS support here, other than GPS rescue…….

Arducopter, again, will need target file modifications to work on the KakuteF4V2, for all the same reasons mentioned above. I have attempted the modifications, and did have them running on a spare KakuteF4V2, but have never tried it with servos and ESCs, so no idea if it could actually drive the outputs correctly. It did connect to Mission Planner for what that was worth. It was not straight forward to get the firmware loaded on the KakuteF4V2 the first time.

At this point in time I think the KakuteF4V2Tri target on iNav is the best path forward for the Tri LR and KakuteF4V2. But there unfortunately appears little enthusiasm by the iNav developers for including it in the mainline repository.

Lastly, with dRonin development stopped, and an apparent problem with the attitude estimator, that no longer is a viable option. If development was continuing, I’d take a stab at fixing the attitude estimation drift, but at this juncture, it doesn’t seem to be worth the effort.

[PeterPankras]

@jihlein Yes with Trinav I mean your branch 🙂 or the iNav version for the Kakutev2

It sucks that iNav does not have the recource mapping from betaflight, would make it so much easier.
I don’t really understand why they don’t include the KakuteV2 in the main build.

So if you use a board supported by iNav you are basicly good to go? most of the supported boards are Matek I see. If you use a Matek board with iNav on a Tricopter do you have servo feedback etc?

I see 2 paths for my LR Tri.
1- KakuteF4V2Tri target on iNav.
2- Switching to a supported board in basic iNav for example a Matek f722std or something.

I’m going to try the first path first. I have a spare KakuteF4 board so would like to flash the KakuteF4v2Tri target on it and do some testing first before putting it on the Tri itself.
I want to make sure everything works, gps, crossfire, etc.
Does anyone have a hex file I can flash on it without me having to try and build the target myself?
Never build anything like that, and although I might be able to pull it off in the long run.. it’s much faster I someone can provide me with a ready to flash version.

I will also see if I can get a reply from iNav devs why they cannot include it in the main build.

[PeterPankras]

I tried building the tri target but I cannot do it.
@jihlein or @yumemi5k can you get me the build hex file so I can try flashing it?

[jihlein]

This is back from the September time frame, but it’s what I was experimenting with.

Attachments:

1. inav_2.2.0_KAKUTEF4V2TRI.zip

[BLeeK]

That file of yours jihlein….I would be able to convert/re-assign/re-map it to run on Matek F405Wing when I finally finish my ‘LR Tri’ build???
Will first be setting it up on dRonin & getting everything working properly before experimenting with ‘TriNav’…
Cheers for advice! 😉

[PeterPankras]

@jihlein thanx, I will check it out.

Is this triflight0p7? the same as @yumemi5k tested? I like to be sure I have the last one 🙂

[PeterPankras]

@BLeeK

You might be fine with basic iNav with the Matek.
iNav supports the Matek boards.

Not sure if you need a modified target for a tricopter when using a matek board. I don’t think so, but cannot check.
I only have a Matek 765 here.

[jihlein]

The build is basic iNav with the pin swap for the servo on the LED pad and LED on the M4 pad. Triflight 0p7 is still giving me some problems, and I made a big mistake trying to both fix that and move to iNav 2.3 at the same time. Right now nothing on that branch builds……

I don’t know anything about the Matek Wing boards, I imagine they can be made to work, but I don’t know if the stock iNav build will work or not. You’ll have to give it a try. Unfortunately there is no resource mapping in iNav, so hopefully the standard pin mapping will work.

[PeterPankras]

@jihlein Ok, thanx.

I got an error trying to build your branch.. but that could explain it.
Did not have time to see what I did wrong.
I can count the times I build something from github on 2 fingers 🙂

I want to see if I can get that working so I can at least build it myself.

[Author]

Posts