Triflight 0.6 RC1

[swissfreek]

I did rerun the tailtune on the bench. It was a brand new FC so I was starting from scratch no matter what.

I am going to put my OpenLogger on this afternoon to get you more data (and video hopefully). Will see if I can get mine to play nice with 0.6.

[swissfreek]

Found some real-estate to mount an OpenLogger to the baby tricopter (forgot the bottoms of the arms are fair game since the battery elevates everything off the ground!), and I can confirm that I have the same issue as @fomoco_44. On TriFlight 0.6RC1, a log file is created, but there is no data in it. Flashing back to 0.5 confirms that the logger does in fact work, and log files are created with data as expected. So something has changed in 0.6RC1 to not work, or at least not work as it used to. Baud rate is currently set to 250000. Will try setting it lower and see if any data populates (if the processor were overloaded, at some point it should be possible to lower the data rate sufficiently that at least *some* packets make it through, no?).

UPDATE: Nope, went all the way down to 19200. Nothing in the log file.

[fomoco_44]

@swissfreek Thanks for confirming that, I thought I was going mad!!

[swissfreek]

Haha well if anything it’s both of us.

[swissfreek]

I also tried loading vanilla CleanFlight 1.14 as well as 1.14.1. I didn’t go deep into the setup other than confirming that it was a tricopter in settings, so it wouldn’t arm, but again, the OpenLogger created blank files. So it sounds like maybe this is an issue with 1.14 rather than TriFlight. Perhaps they’ve changed something about logging that requires different settings or new firmware for the logger (please no!) or something of the sort…

[LitterBug]

@lauka I had noticed higher than normal CPU utilization too. Thought about adjusting the looptimes, but the latest cleanflight configurator has moved things around to the point where it wasn’t obvious where to go. Wanted to get in the air this weekend and didn’t spend a whole lot of time on it. Everything was fine when I went back to 0.5 and flew a few packs through yesterday. Fortunately doesn’t look like there was any permanent damage to the motors. Not kidding, these things were smokin after 30 seconds on 0.6. Sounds like adding logging isn’t going to be much help at this point, so I will hold off on 0.6 for now unless you have any suggestions or updates.

Cheers!
LitterBug

[LitterBug]

With the in-air tailtune, what is being measured, and after what changes to the vehicle would it need to be rerun? Are the results stored in a parameter that can be viewed from CLI, or are they still hidden in flash?

Cheers!
LitterBug

[lauka]

I will wait until the CPU load issue is resolved in CF and release a new version for testing. Dominic hasn’t been very active lately.

Result from the in-air tail tune is the thrust to torque factor which is available in CLI. Nothing is hidden in flash.

Bench tail tune stores the ADC values in flash only. Those need to be recalibrated if flash is cleared.

[LitterBug]

Great info Lauka!

Thanks!
LitterBug

[LitterBug]

I’ve been doing a bunch of Quad flying on BetaFlight running stock PIDs and it has been rock solid for my variety of quads. I’ve been doing much the same leaving PIDs stock on the Baby Tri running TriFlight 0.5. I have a good grasp on tuning quads, but the tail tuning throws a new set of variables into the mix for the Baby Tri. I’d like to dive into more actively tuning the Baby, as it does have a few characteristics that I think should be able to be tuned out. Waggle foremost, and the twitches that happen during rapid throttle change or during some acrobatic maneuvers. Are there any good guides geared more on Tricopter tuning? I’d like to get a handle on the tuning before swapping the rather mellow 2205 2100kv motors on 3 and 4S out for some insane 2207 2450s on 3 thru 5S. I’ll be adding an OpenLager to the F3FC on uart3. I’m sure that will help me understand what is going on too.

Cheers!
LitterBug

EDIT: I have read through the triflight setup guide HERE I’m looking for more info on the additional CLI commands that TriFlight introduces.

[Terje]

I saw Joshua Bardwell had a BabyTri on his wall. Maybe you can persuade him to do a PID tuning session with it? 🙂

[swissfreek]

By the way @LitterBug, my Baby Tricopter doesn’t exhibit the motor heat issues you have. Mine is stock, it sounds like yours might not be.

@dkdarkness, based on the comments, he was sent it by a viewer to get sorted out and he’s having trouble as well. Maybe one of us should send him some Blackbox data to take a look at (though obviously it would be for 0.5).

[LitterBug]

I noticed the baby tri on Josh’s wall too. LOL

Cheers!
LitterBug

[LitterBug]

@swissfreek not sure what you mean by “stock”. Stock firmware, hardware, CleanFlight settings? I should be able to give 0.6 another try sometime this week. I didn’t want to spend much time on it last weekend since the weather was good and we aren’t going to have many more great flying days like we had. Mine on 0.6 was definitely pegged on CPU. The 5 second countdown beeper was distorted when starting the in-flight tail-tune in addition to the motors being very gritty. I should have more time to poke around, grab status and tasks, etc… I may have inadvertently botched up some setting(s). Looking at a diff between default 0.6 settings for example, I somehow had min_check at 1000 and max_check at 2000. That’s not stock or my normal settings and I usually leave those stock. I have the results of the diff of the 0.6 defaults and my final dump beore reverting back to 0.5 shown below.

feature -BLACKBOX
feature RX_SERIAL
map TAER1234
serial 1 64 115200 57600 0 115200
set min_check = 1000
set max_check = 2000
set serialrx_provider = SPEK2048
set min_throttle = 1050
set disarm_kill_switch = OFF
set failsafe_procedure = 1
aux 0 0 1 1450 2050
aux 1 1 0 1450 1550
aux 2 2 0 1950 2050
aux 3 12 2 1450 1550
aux 4 27 1 1450 2050
aux 5 28 2 1950 2050
servo 5 1035 1887 1447 100 -1
set p_level = 50
set i_level = 50

My build:
Stock Components:
RCExplorer Baby Tricopter
RCExplorer F3FC Racing FC
RCExplorer Baby PDB
BMS-210DMH Servo
Not quite stock but certainly adequate:
DYS XS30A ESCs running BLH_S 16.4
EDGE Racing 2206 2100kv motors

I’ve been re-reading the whole “Triflight Setup Guide” thread as well as the “Triflight Checklist.md” on GitHub. Going to start again from scratch with 0.6. I’m game for exploring and open for suggestions. If I burn up the current motors, I have new better ones arriving tomorrow anyway. I’ll download the firmware again and make sure the files match. Maybe I got a corrupt file.

Cheers!
LitterBug

EDIT: Diff of Triflight 0.5 default settings vs what I have been flying:

feature -BLACKBOX
feature RX_SERIAL
feature CURRENT_METER
map TAER1234
serial 1 64 115200 57600 0 115200
set serialrx_provider = SPEK2048
set min_throttle = 1040
set disarm_kill_switch = OFF
set tri_tail_motor_thrustfactor = 46
set tri_tail_servo_speed = 268
set failsafe_procedure = 1
aux 0 0 1 1475 2025
aux 1 1 0 1450 1550
aux 2 2 0 1950 2050
aux 3 12 2 1450 1550
aux 4 27 1 1450 2025
aux 5 28 2 1950 2050
servo 5 1105 1975 1539 40 40 100 -1
set p_level = 50
set i_level = 50
set rc_expo = 83
set roll_rate = 70
set pitch_rate = 70
set yaw_rate = 80

[LitterBug]

Just looked at CPU Load in CleanFlight Conigurator for my Baby Tri in it’s current 0.5 config. 0%.
It was floating in 40-50% when I had 0.6 running on it.

Cheers!
LitterBug

[Author]

Posts