Testing – Triflight rebased on 3.2.2 Betaflight

[Kevin_Erik]

@ianlh

If you don’t mind me asking, how are you performing a Tailtune using the BF files? Far as I know of Tailtune isn’t part of that hex.

[lauka]

Here’s the Matek F405 hex.

Attachments:

1. triflight_bf_3.2.2_v2_MATEKF405.hex_-1.zip

[Kevin_Erik]

I’ve 3.2.2 loaded on the Tri-Mini (RCexplorer F3FC)… Will post logs next time I am able to fly.

Note: Still working on the Tri-baby with the Matek F405.

Cheers!

Kevin

[Kevin_Erik]

3.2.2 Betaflight, triflight_bf_3.2.2_v2_RCEXPLORERF3.hex_.zip

Just tried using “tri_unarmed_servo” as it’s supposed to allow the tail servo to be manipulated while the copter is disarmed.

This function appears to be listing twice when the CLI “GET” when it is used.
One set shows “ON” while the other set shows “OFF”.

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

tri_unarmed_servo = OFF
Allowed values: OFF, ON

Any attempt to set the value to ON, does nothing but setting it to OFF shows:
# get tri_unarmed_servo
tri_unarmed_servo = OFF
Allowed values: OFF, ON

tri_unarmed_servo = OFF
Allowed values: OFF, ON

Not sure if this is a typo or if its an error of sorts.

Edit: Just played with this some more, pulled a CLI-Dump and saw this:
set tri_unarmed_servo = ON
set channel_forwarding_start = 4
set tri_unarmed_servo = OFF

Looks like its being enabled then disabled somehow. Any idea if this is deliberate?

[fomoco_44]

@Lauka,

I was able to successfully test the BF3.2 V2 code today, and it looks and feels great!….pretty much the same as the BF3.1.7 V8 code using the same PID’s and hardware.

Here is the log and diff:
https://1drv.ms/t/s!Akrjv9THYPIka2b09pBeo03FxeY

# diff
# Triflight on Betaflight 3.2.2 v2 / RCEXPLORERF3 (REF3) 3.2.2 Dec 29 2017 / 14:29:45 (30f68c8) MSP API: 1.36
name Mini Tri
servo 5 1017 1925 1481 100 -1

feature AIRMODE
feature ANTI_GRAVITY
map TAER1234
serial 1 128 115200 57600 0 250000
serial 2 1 115200 57600 0 115200
aux 0 0 0 1800 2000
aux 1 37 1 1300 1700
set baro_hardware = NONE
set rc_interp_ch = RPYT
set serialrx_provider = SPEK2048
set blackbox_p_ratio = 16
set motor_pwm_protocol = MULTISHOT
set align_board_pitch = 180
set vbat_max_cell_voltage = 44
set current_meter = ADC
set ibata_scale = 350
set tri_tail_motor_thrustfactor = 118
set tri_tail_servo_speed = 340
set deadband = 1
set yaw_deadband = 1
profile 0

set dterm_lowpass_type = PT1
set p_pitch = 40
set i_pitch = 45
set p_roll = 50
set i_roll = 45
set d_roll = 40
set p_yaw = 75
set d_yaw = 60
rateprofile 0

set rc_rate = 130
set rc_rate_yaw = 130
set rc_expo = 20
set rc_expo_yaw = 10
set roll_srate = 75
set pitch_srate = 75
set yaw_srate = 75

I have the video but haven’t had a chance to upload it yet, let me know if you need it or if the log is good enough 🙂

Two quick questions:
1. I’m getting some mild oscillations in both the yaw and pitch axis, and so I’m pretty sure one is driving the other, but I can’t tell which one is the culprit!….do you have any sage advice on how to determine this? I was thinking of raising Pitch P and lowering Yaw P but I would love to hear your advice.

2. BF3.2 has the Dynamic Filter feature which seems to work pretty good on my other quads, do You see any issue with enabling it(and subsequently disabling the notch filters) on the Tricopter?…I’m not sure how it would handle the tail servo feedback routine/code. I’m trying to avoid turning my MiniTri into another flying-blender-of-death!

Great work on the code!!

[Sam Lane]

@fomoco_44

Concerning your coupled oscillation issue, I would suggest taking a look at your blackbox logs, specifically the P-correction values for Pitch and Yaw. Theyll both roughly approximate sine waves. Whichever wave precedes is the axis driving the oscillation.

All the best!
Sam

[fomoco_44]

@Sam Lane

Thanks for that, I’ll go back and have a closer look…..my suspicion is that it’s a yaw oscillation driving the pitch axis.

[Kevin_Erik]

Tricopter-Mini: Just tried doing a quick Tail-Tune & test hover (in-house) as the weather outside is total crap. Seems as if the Tail servo is non responsive. Won’t respond to RC commands nor will it Tail-Tune.

Anyone know whats going on ?

# diff
# Triflight on Betaflight 3.2.2 v2 / RCEXPLORERF3 (REF3) 3.2.2 Dec 29 2017 / 14:29:45 (30f68c8) MSP API: 1.36
name Tricopter Mini
servo 5 1000 2000 1582 100 -1

feature DYNAMIC_FILTER
beeper -ON_USB
map TAER1234
serial 0 1 115200 57600 0 115200
serial 1 64 115200 57600 0 115200
serial 2 128 115200 57600 0 115200
aux 0 0 0 1800 2100
aux 1 1 2 900 1300
aux 2 2 2 1300 1700
aux 3 37 1 1800 2100
set gyro_sync_denom = 4
set gyro_notch1_hz = 0
set gyro_notch2_hz = 0
set min_check = 1010
set max_check = 1990
set rssi_channel = 16
set min_throttle = 1160
set max_throttle = 1900
set align_board_pitch = 180
set bat_capacity = 1800
set current_meter = ADC
set ibata_scale = 336
set servo_lowpass_hz = 100
set tri_tail_motor_thrustfactor = 78
set tri_tail_servo_speed = 203
set pid_process_denom = 1
set dterm_lowpass_type = PT1

[mcstan]

what is the resource dump look like?

[Kevin_Erik]

# resource
resource BEEPER 1 A00
resource MOTOR 1 A08
resource MOTOR 2 B00
resource MOTOR 3 A04
resource MOTOR 4 B01
resource SERVO 1 A07
resource PPM 1 A01
resource SONAR_TRIGGER 1 A06
resource SONAR_ECHO 1 B01
resource LED_STRIP 1 B08
resource SERIAL_TX 1 B06
resource SERIAL_TX 2 A02
resource SERIAL_TX 3 B10
resource SERIAL_RX 1 B07
resource SERIAL_RX 2 A03
resource SERIAL_RX 3 B11
resource I2C_SCL 2 A09
resource I2C_SDA 2 A10
resource LED 1 B04
resource LED 2 B05
resource SPI_SCK 2 B13
resource SPI_MISO 2 B14
resource SPI_MOSI 2 B15
resource ESCSERIAL 1 A08
resource ADC_BATT 1 A05
resource ADC_RSSI 1 A06
resource ADC_CURR 1 B02

[Kevin_Erik]

Just pulled a Resource from Hex version Triflight 0.7 Beta 3 / RCEXPLORERF3 3.1.7 Oct 21 2017 / 13:57:41 (17fa321). So far it looks similar enough but I’ll try powering up to see if the Servo will play.

# resource
resource BEEPER 1 A00
resource MOTOR 1 A08
resource MOTOR 2 B00
resource MOTOR 3 A04
resource MOTOR 4 B01
resource SERVO 1 A07
resource PPM 1 A01
resource SONAR_TRIGGER 1 A06
resource SONAR_ECHO 1 B01
resource LED_STRIP 1 B08
resource SERIAL_TX 1 B06
resource SERIAL_TX 2 A02
resource SERIAL_TX 3 B10
resource SERIAL_RX 1 B07
resource SERIAL_RX 2 A03
resource SERIAL_RX 3 B11

[Kevin_Erik]

So far no luck with either hex loaded. Not sure how but it appears that the servo is dead. I’m at a loss as to how, being that the last few flight went smooth. Is it possible that swapping the hex can cause the servo damage in some way? (doubtful)

Edit: Spent the last 2 hours taking the Tri-mini apart and have yet to find anything physically wrong with it. Wiring looks great and beyond the servo not working, the Trimini should be flyable. (Motors & Sensors test good) Is the failure rate of the servo’s really this bad or am i having a shit run of luck?

If this one is dead, then it would be the third servo in under a years time.

[MadPoet]

Finally sunny and only slightly breezy, I took my Baby Tricopter out for a test, please see below;
# diff
# Triflight on Betaflight 3.2.2 v2 / RCEXPLORERF3 (REF3) 3.2.2 Dec 29 2017 / 14:29:45 (30f68c8) MSP API: 1.36
name Y170
servo 5 1023 1853 1428 100 -1

feature RSSI_ADC
aux 0 0 3 1700 2100
aux 1 2 2 900 1600
aux 2 4 2 1800 2100
aux 3 13 4 1700 2100
aux 4 27 0 900 1300
aux 5 28 2 1300 2100
aux 6 37 1 1700 2100
set baro_hardware = NONE
set min_check = 1010
set max_check = 1910
set min_throttle = 1040
set tri_tail_servo_speed = 312
profile 0

rateprofile 0

# get servo_pwm_rate
servo_pwm_rate = 300
Allowed range: 50 – 498

# get tri_tail_motor_thrustfactor
tri_tail_motor_thrustfactor = 54
Allowed range: 10 – 400

two questions, one as per https://github.com/lkaino/Triflight/blob/master/docs/Triflight%20Checklist.md
“Use the configurator CLI and check tri_tail_motor_thrustfactor. If it’s still 54 (default) it’s likely the tuning failed.”

Should I assume that it failed and keep trying? Previous attempts, I’ve done this dozens of times with the same result.

2nd question: As per https://rcexplorer.se/forums/topic/250hz-for-in-openpilot-cc3d-atom-revo/
indicates that it should be 250, the post is a bit old, I wonder if this still applies.

My tricopter Specs are;
RCExplorer F3FC
BabyPDB – 36x36mm Power Distribution + BEC + Current sensor (REV3)
– BEC set to 6v for servo
BMS210 Servo with feedback wire and stainless steel output shaft
Aikon SEFM 30A BLHeli_S ESC
Emax RS2205 red bottom motors
Dalprops T5045
4s 1400mah

And as a bonus question; it’s time for me to learn how to tune the baby tricopter. Can anyone recommend articles and video to help me tune the baby tri. I see a little tail wag when I come off the throttle and/or when applying yaw. Still new to this, so can’t really explain it. I have no blackbox logs, I don’t know how to add that to my current set up and current FC doesn’t have any RAM for such logging. I have a cheap camera I can hook up and turn to face the tail motor if it would help, but I would need to remove my VTX. I scavenged a camera and VTX from my X220 (Eachine Wizard) when I installed the new gear into it.

@Lauka, thank you for your work on this project – even with the tail wag, should I not figure out how to tune it out, this is a blast to fly! I am so glad I ventured into building a tricopter.

[mcstan]

@kevin ahhh okay.. I didn’t know you had it working before…. servo might be dead then.. I did read somewhere if the timing on the motor outputs pins are the same as the servo output pin and you set the motors to dshot that the servo stop working

which is why I asked for your resource output… but you have not listed the timings

@madpost I found if you set all your I and D down and just mess with the P like you would with a normal quoad you see which one is affecting the tail wag.. the only thing I think when tuning a tri-copter is.. instead of hearing the oscillations like with a Quad you see them in the tail wag… well that’s the case with the Yaw P and roll P

I all so found too much “I” can give you a lot of tail wag

some people may disagree but this is just my point of view.

[Kevin_Erik]

@mcstan

The timings are the default, OneShot125. Part of why is that the F3FC used for the mini can’t handle anything else as its unable to talk to the esc’s. Same is true for the passthrough… While it can be setup, it doesn’t allow you to reprogram or flash the esc’s. (F3 limitation)

What I think happened is that the Tailtune hadn’t yet completed and hence the end limits for the Servo had not yet been fully defined. Thus the servo tried to turn past the physical limits of the motor mount, burning up in the process. Granted i have no real way to know this for sure but from what i saw the gears are fine. Cleaned them and reassembled the whole thing. However not before attempting to spin up the servo with the gearing off. Obviously this didnt spin or even move a little.

Hence there are three possibilities in order of likelihood:
-Bind & burnt
-Over voltage (5-6v is norm)
-Impact damage (not likely as it was working before the hex flash)
-Bad wiring via internal breakage (its possible but unlikely as i have has this same fault before. What are the odds of two exact failures)

Would be nice if the Tail tunes defaults before the tune is performed, were somewhat less than the possible limits. This way it would prevent the servo from being burned after a flash.

[Author]

Posts