Home › Forums › Everything about everything else › New type of yaw control
Tagged: EAGLE TREE VECTOR, GAINS, PID, RCEXPLORE TRICOPTER V4, TUNINING
- This topic has 224 replies, 20 voices, and was last updated 5 years, 11 months ago by mcstan.
-
AuthorPosts
-
31 July, 2017 at 18:18 #38877swissfreekParticipant
You mean like with an oscilloscope? No I don’t have one of those. Yet.
I followed the image in the post I linked earlier. Confirmed that my servo works, but coper won’t arm (I’m assuming it’s not arming since the servo doesn’t become active, though I can do the unarmed tail-tune just fine).
31 July, 2017 at 18:21 #38879laukaParticipantThere is the new arming protection which prevents arming if the feedback reports servo angle outside the range.
Is your servo in the middle position when you’re arming? Did your servo speed calibration finish successfully?
31 July, 2017 at 18:34 #38880swissfreekParticipantIt’s pretty much centered. I’ll do the unarmed tail-tune again tonight to confirm. I don’t have a beeper yet but I assume it completed successfully since I watched it go through it’s entire routine. I’ll attach a beeper tonight to cofirm via beep and check the yaw settings in CLI to see if they changed from default.
1 August, 2017 at 03:18 #38881swissfreekParticipantOK, confirmed that I can run the motors (manually, using the motor tab) when I set the ESCs to multishot in the configuration tab. I can run the unarmed tail-tune and it does appear to succeed as the diff reads that my servo speed is 289 (so it’s not the default). Tail is dead-center. Receiver tab shows all channels functioning. But it won’t arm. Again, here’s my diff:
# diff
# Triflight 0.7 Beta 2 – new control test v6 / BETAFLIGHTF3 3.1.7 Jun 23 2017 / 14:31:15 (0110b62)
name SwissFreek
resource MOTOR 1 B08
resource MOTOR 3 B09
resource MOTOR 4 A06
resource MOTOR 5 NONE
resource MOTOR 6 NONE
resource MOTOR 7 NONE
resource MOTOR 8 NONE
resource MOTOR 9 NONE
resource SERVO 1 B07
resource PPM 1 NONE
servo 5 1155 1899 1528 40 40 100 -1feature AIRMODE
serial 0 64 115200 57600 0 115200
serial 1 0 115200 57600 0 115200
serial 2 2048 115200 57600 0 115200
aux 0 0 0 900 1100
aux 1 13 2 900 1800
aux 2 31 3 900 1100
set motor_pwm_protocol = MULTISHOT
set serialrx_provider = CRSF
set tri_tail_servo_speed = 289
set tri_servo_feedback = EXT1
profile 0rateprofile 0
rateprofile 0#
1 August, 2017 at 23:50 #38900laukaParticipantYou could make sure it’s not because of it if you change the the servo feedback to virtual.
Only other thing I can think of is too high CPU load. Don’t see anything else than the crossfire protocol that could cause load. You could also try increasing the PID loop time denominator.
2 August, 2017 at 01:00 #38901swissfreekParticipantAha! Setting the tail servo to virtual solved the problem. So, tell me more. How do I troubleshoot whatever issue I’m having with the tail servo?
2 August, 2017 at 10:48 #38906laukaParticipantYou could
set debug_mode = TRI
and check the value of debug[3]. The value is the calculated servo angle from the feedback.3 August, 2017 at 06:13 #38918fomoco_44Participant@lauka,
Have you or anyone else tried faster servo’s on their Tricopters?
I was contemplating using one of the following servos on either 6v or 8v, which all have faster response times for similar or better torque values:https://hobbyking.com/en_us/bms-21hv-high-voltage-micro-servo-metal-gear-3-2kg-09sec-15-2g.html
https://hobbyking.com/en_us/turnigytm-tgy-313c-helicopter-cyclic-ds-mg-w-alloy-case-servo-3-5kg-0-08sec-20g.html
https://hobbyking.com/en_us/pdi-2004mg.html
https://hobbyking.com/en_us/goteck-hc2123t-hv-digital-mg-metal-cased-mini-servo-3-7kg-0-05sec-20g.htmlThe PDI-2004 looks like it has the best performance, but would need the mounting tabs machined off to fit.
The Gotek HC2123T and Turnigy TGY-313C(Gotek copy??) should drop straight in hopefully.
The BMS-21HV should also drop in easily, however there is a slight loss of torque @ 7.4v compared to the others, and I have heard reports that running this at 8.0v might burn them out??.Obviously they would all need the extra feedback wire connected, and I was able to do this with a Savox SH-0255MG in my Trifecta frame without too much hassle, so hopefully it shouldn’t be too hard!!
3 August, 2017 at 06:54 #38919sbole04ParticipantJust wondering is ghere a big diff from v6 to v7? Im finally gusting my machine it is like it is on rails, I have fought and fought with this thing using the gruapner fc.
All i did once i set tri_motor_acc_yaw_correction to 0 reduced the d on pitch and roll, the thing is a different machine now, so excited… thx lauka for the time and effort you have put into this
7 August, 2017 at 10:48 #38982noir522ParticipantSWISSFREAK.
I noticed you had your baby tricopter initially set up with a colibri powercube. If you still have acces to the cli “resource list” could you tell me which pin the servo is assigned to? The colibri has a dedicated servo output pin on the side of the bourd but I cant figure out the pin’s address for betaflight cli resource list (ie “resource servo 1 ????”) I have tried A03 , B15, A08 and a number of others but nada.
Attachments:
7 August, 2017 at 16:22 #38988LitterBugParticipant@fomoco_44,
I recently had to rebuild the tail on one of my baby tricopters after destroying the servo. Was trying to get it back together before FliteFest so ordered some alternate servos in case a replacement RCE servo didn’t make it in time. Ended up putting a TGY-211DMH on it which is slightly faster but with slightly less torque. Added the feedback wire myself and am running it at 6Volts. Servo speed comes out to 384-402ish with this servo on the ground tailtune which is better than the 290-315 that I get from the 210DMH servos. Have not done much flying with it, but the servo is noticeably faster. Had issues with rapid tail oscillations trying to get an autotune running dronin and have not yet managed to get that worked out. Have flown it a handful of times running TriFlight 0.7 beta2, but have not tested TriFlight0.7 Beta2 New yaw control V6. I should be able to get it out to the field(s) this week.
I also have a real fast high torque BlueBird BMS-A207 to test at some point. First one had a bad potentiometer or I would have used that one instead of the TGY-211DMH. Replacement is good. Just need to add feedback wire and work on connectors to easily swap servos out on my Baby and Mini tricopters. I have a spare tail pivot that I can use for bench speed testing too.Cheers!
LitterBug7 August, 2017 at 18:16 #38989laukaParticipant@LitterBug: did you notice the v7 build?
7 August, 2017 at 20:03 #38991LitterBugParticipant@lkaino Somehow I missed that. Better grab it and give it a spin!
Cheers!
LitterBugEDIT: I had seen the post and watched the video, but somehow missed the links at the bottom. It’s calm but wet here. Should be able to at least get it loaded up and tail tuned in the back yard. Will try to get some FPV later this week.
7 August, 2017 at 20:10 #38993swissfreekParticipantI have v7 loaded on my baby Tricopter, and with feedback set to virtual I managed to get it off the ground. Wasn’t able to really fly it because the RunCam Split (this is my theory at least) got upset that the servo started competing for power from the PDB when I armed, and it froze, so my OSD worked but the picture still had me on the ground! Very disorienting. Anyway, I have a BEC on order so I can run the Split directly off the battery leads. Have still not solved the issue with the servo feedback not working, but at least there’s hope that I will get this thing running again. Looking forward to it!
Oh, and @noir522, I never set the Colibri up because I couldn’t figure out the servo feedback (doesnt look like colibri has a pin ex[osed for it) so I switched over Thea BFF3.
7 August, 2017 at 21:17 #38994LitterBugParticipant@lkaino,
Speaking of servo feedback… I’m looking at using the AirDTF DTFc FC (doge) so I can have native dRonin support for tricopters. There is no RSSI pin, but @jihlein discovered the the LED strip output could be used on A06 and we have successfully bench tested dRonin using that pin for feedback. He also created a copy of TriFlight0.7 Beta2 V6 that had the pin mapped to EXT1. Everything seemed to be working up to the speed calculation and beeped successful, but the results never got calculated and/or placed in to tri_tail_servo_speed. Was wondering how hard it would be for you to build an image for the “doge” with the LED unmapped from A06, and the RSSI mapped to A06?Here’s the mapping that I had working for everything else.
Servo on PWM 1
Tail Motor PWM 3
RF Motor PWM 4
LF Motor PWM 5diff
# Triflight 0.7 Beta 2 – new control test v6 / DOGE 3.1.7 Jul 7 2017 / 18:03:36 (fc9f38d)
resource MOTOR 1 A10
resource MOTOR 2 A09
resource MOTOR 3 A00
resource MOTOR 4 NONE
resource MOTOR 5 NONE
resource MOTOR 6 NONE
resource MOTOR 7 NONE
resource MOTOR 8 NONE
resource SERVO 1 B08
servo 5 900 2100 1464 40 40 100 -1feature -RX_PPM
feature RX_SERIAL
map TAER1234
serial 2 64 115200 57600 0 115200
aux 0 0 1 1450 2050
aux 1 28 0 1450 2050
aux 2 1 0 950 1050
aux 3 2 0 1450 1550
aux 4 13 2 1950 2050
aux 5 31 2 1450 1550
set serialrx_provider = SPEK2048
set tri_tail_servo_speed = -1
set tri_servo_feedback = EXT1 -
AuthorPosts
- The forum ‘Everything about everything else’ is closed to new topics and replies.