Yesterday I flew my TRex500 for the first time after replacing its flight controller with a Holybro Pixhawk 6C Mini. I had copied the parameters from the previous fc onto the 6C, and then configured all the mandatory hardware. The heli flew fine in Stab, PosHld, AltHld and RTL modes yesterday.
Today I pulled the logs to see if any tuning was required. RATE.Y and RATE.P graphs look normal, with actual and desired graphs following each other closely as single lines, but I got this strange (to me at least) graph for the RATE.R log. What could be the cause of what looks like a high-frequency superimposed on the RATE.R graph, and how do I correct it please? A link to the .bin file is below.
It looks like noise, parameters are embedded in the .bin log file. IIRC it may be D noise, I would try dialing ATC_RAT_D down a bit and running AUTOTUNE on all axis, especially if you replaced non Ardupilot FC with Ardupilot one, parameters don’t transfer from different stacks well, though I have only looked at the screenshot.
Thanks LupusTheCanine, but I can’t find ATC_RAT_D. Do you mean ATC_RAT_RLL_D? It’s currently at the default of zero.
I note that my ATC_RAT_YAW_D and ATC_RAT_PIT_D are 0.001 and 0.003 respectively. Thus far I’ve always used the manual tuning method for my helis, but I guess I should give Autotune a try. I’ll read up on it.
Your vibrations are a bit on high side, and when landed are too high. This may be causing issues so I would try improving the FC mounting.
To me it looks like feed forward is too low and proportional may be too high for available damping so I would try increasing FF a bit and dropping P before running autotune. Though I have limited experience with tuning ( budget 150 class heli I put Ardupilot on)
You didn’t copy over all of the information from your previous controller. The first time you write the parameters, it will enable the features that were enabled in the previous controller but it will not set the underlying parameters for that feature because they are not exposed until the feature is enabled. so you need to do a compare params again and that should write the remaining parameters for the features that you enabled. for instance, both of the notch filters that you were using on your previous controller were enabled but all of the additional parameters that are exposed once the notch filter was enabled are at there default values. That is why your control signals are very noisy.
So I recommend that you do a compare params using mission planner full parameter list page. then click the select all button to deselect all of the parameters changes. then go through the parameters and select your INS_HNTCH and INS_HNTC2 params to write the filter settings that you had on your previous flight controller
LupusTheCanine, my FC is mounted on top of a 6mm 3D-printed spacer block to raise its various ports above the heli’s side frames. Both the spacer block and the FC are mounted using 3M anti-vibe tape. I’ll look to improving that aspect once I’ve got the noise resolved.
Bill Geyer, thanks for the insight. I’ll do what you suggest and then my next flying opportunity will be Sunday. I’ll post the result then.
Bill, seems like I’ve misunderstood your instruction: I changed ALL the INS_HNTCH and INS_HNTC2 parameters that showed up different from my flights with the previous FC. (I note that some INS_HNTCH and INS_HNTC2 parameters are not listed in the compare file because they’re the same in the old and new configurations.)
The heli then wouldn’t arm. On returning to my workshop I saw that MP messages was reporting “Config Error: Throttle notch unavailable with helis”.
When I reverted those parameters back to previous, the heli armed without any error messages. So, which of those ‘New Value’ parameters are causing the problem please?
Google suggests that INS_HNTCH_MODE must be 0 for helis. But I’ve just checked the param file for my other TRex500 heli and it’s 1 – just like this one was before I changed the FC.
Throttle notch doesn’t make sense for helicopters as in helis “throttle” is collective. You want either either ESC telemetry (direct drive or no clutch helis) or rotor RPM sensor to drive your notches.
I believe there was an issue with a previous version of software that allowed heli’s to use the throttle mode. do you know if you upgraded firmware when moving to the new FC?
Please provide me the Notch filter settings (entire list of HNTCH and HNTC2 parameters) from your previous controller. I suspect you intended to use a fixed notch but I want to see the set of parameters to be sure I am guiding you on how to set your new FC parameters.
You’re right Bill, this heli was running on 4.6.x with the old FC and I updated it to 4.7.0 when I installed the new Pixhawk FC. As far as I can see, my other 2 helis still running on 4.6.x and with Matek FCs have similar INS_HNTCH parameters and are flying without any problems.
Here’s a screen shot showing all my HNTCH and HNTC2 parameters for this heli with the old FC. My head speed is 1850rpm governed by the CC Talon 90’s governor mode.
Heli’s arming and flying again, but looks like I’ve messed up
I reset my INS_HNTC2 and INS_HNTCH back to default, then followed the Ardupilot instructions Traditional Heli Static Harmonic Notch instructions to set new values:
A very short hover in my back yard showed the heli flies okay, but the RATE.R, P and Y graphs from the log all have the ‘noise’ superimposed on them, as per the original screenshot.
Here’s my .param and .bin files from today’s tests.
I remember I thought the speed was too high so I used the CC app to re-set it at 1850rpm – or so I thought. That’s why I had the frequency at 31. But the headspeed feels fine to me, and to my ear it sounds steady, so I’ll assume it’s actually 2200rpm, which is what I believe your suggested frequency of 37 corresponds to, and leave it as it is.
I’m downloading your suggested values to my heli now, and I’ll have a few flights on Sunday and report back.
I took off in AltHld mode as normal, and things seemed to be going okay for a few anti-clockwise and clockwise circuits, but then after about a minute the heli started pitching back and forth violently (that’s what I and other observers thought we saw, but the log seems to suggest it was yaw) and the heli started drifting away from me. In my confusion I never thought of switching to PosHold mode, but simply started gently reducing height while trying unsuccessfully to bring it back towards me. It then hit a goalpost and crashed. Does the log give any clue why, please?
Also the .param file in case I did anything wrong:
I’m leaning toward thinking one pitch link could have come off mid flight, or similar issues around blade fork assembly. Can you share more details about the damage, is the rotor head still in one piece to make an assessment?
I can’t see any issues around the flight controller, both IMUs seems to be aligned throughout the whole event, actuators supply seem ok as well.
Thanks for looking at the log Ferrosan. I can see where you’re coming from with collective and roll going wild at the same time. The head is intact and all ball links are still connected except for servo #2 where the servo arm is broken and it and the ball are missing. One link from swash to blade is also badly bent which, with the broken servo arm, makes me think the heli came down inverted (it was too far away for me to be sure) after hitting the goal post at about 20ft above ground.
The only problem with that theory is the heli is quite new and I do check regularly for slop in the links, or anything. But that doesn’t rule it out. Anyway I’ll be replacing the head, mainshaft, links, etc. to be sure
Sure, I was just guesstimating without having the whole picture. I can only suggest to have a deep look at the whole kinematic/ gearbox / bearings. It is scary wild what we are seeing on the log.
After a complete rebuild (main shaft, head, main gear, and collective servos) I managed to get a test flight yesterday. Parameters are unchanged from last time other than the ones which were set by recalibration of sticks, compass, etc. and a minor adjustment to head speed. TRex500 after crash rebuild.param (17.9 KB)
From my perspective the first flight in stabilised mode was uneventful, and the logs appear to me to confirm that roll, pitch and yaw responses are as good as they felt. My second flight was curtailed by rain, but I did manage to get a couple of minutes in AltHld mode without any issues. It seems to me that the high frequency that appeared to be superimposed on my previous RATE traces no longer exists to the same extent as before. The log of the first flight is here in case anyone can see anything I should be concerned about:
I replaced the collective servos because two of them had been impacted during the crash, one having a broken arm and the other a very bent connecting rod. Could a servo fault have caused the noise in my RATE graphs, and then developed further later to cause my crash? They’re Align DS530.
The other thing I found when I disassembled the heli was the one-way bearing was bad. Two needles dropped out when I took it off the shaft. I’ve never had that happen before.
