LibrePilot Forum
Users => Vehicles - MultiRotors => Topic started by: RayRegan on December 20, 2015, 01:19:19 am
-
When I built my ZMR 250 I used a PDB. I did so to cut down on excess wiring and clean things up. I decided to use female header pins on the PDB so I could simply plug the CC3D into the PDB. When I attached the carbon arms to the chassis I did not use any kind of rubber or silicone sandwiched between the arm and the PDB. Now when I take off, sometimes in a hover, and on landing, I get small oscillations or corrections. I checked to be sure that everything is tight and blades are well balanced. Should I have used something to isolate the CC3D from vibration? Could this be causing the perceived need for corrections that I am encountering? Or is there simply an adjustment that I need to make to reduce the sensitivity of the CC3D so it will ignore this?
-
Small oscillations don't generally come from vibration, but jumps in angle do.
Increase gyro Tau and see if that helps. Gyro Tau is a vibration filter.
Stabilization -> Expert -> Gyro Noise Filtering
I assume you are not running Attitude mode, but if you are you might also try the accel filter at:
Attitude -> Settings -> Filtering
-
I am so frustrated with this thing!!!
I have tried the Gyro filtering you mentioned as well as the accel filter. I did so in a scientific manor (one change at a time) to be sure one thing was not affecting another. I still cannot get rid of this wobble. sometimes it gets bad enough to flip itself over, and sometimes I can get it 2 foot off the ground and out of ground effect and it will calm a bit. It should not be there. I have broken about 6 props and cracked one arm already, I just want to fly this damn thing! I can't really see much difference with any of the changes I have tried. If I go to far it gets worse but then going back the other direction beyond where I started it gets worse again. I seems to over correct again and again. The oscillation gets worse and worse until I either land or it flips.
-
forgive if you already know this
you must set up the quad exactly as described. the CC3D board must be pointing with arrow forward. the CC3D outputs must be connected to the correct ESC/motor. the motor must turn in the correct direction. the correct prop must be on that motor.
some users connect escs to the wrong place in the CC3D and that is one possibility, that you have your esc places off by one.
another possibility is that you are using 50ms esc update rate. you need to use the highest update rate possible. it should be one of: 490 pwm, pwmsync, or oneshot125
check your output page and make sure that all your channels are set to use the same, high update rate
-
My CC3D faces up with the arrow pointing LEFT. In the Attitude tab when the board is plugged, the orientation of the board can be adjusted and saved. This must be redone any time that the board is erased and firmware is reloaded. The Motor outputs can also be adjusted on the Vehicle tab at the bottom. The motor direction and propeller direction are correct per the image for Quad-X on the vehicle tab. The ESC's are setup as RAPID ESC and the motor update rate for the affected outputs is maxed out at 490Hz. These were the first things I checked. I saw others with the same issue online today. Several of them tried to tune this issue away and could not. One guy had his sliders all the way down then 2/3's up and had the same oscillation. Turned out he had his motor update rate set to 50Hz, he changed that setting and it fixed the issue. Another guy said it was his anti-vibration mount causing it. He mounted his board using servo mounting tape instead and says it flies great now. I think I will try to reduce the vibration on my machine somehow to see of that has any effect. It will be a nice break from all this tuning stuff for a few hours. BTW I don't think that adjusting that Gyro Filter on the EXPERT tab has any effect if you are using the BASIC setup. I tried changing these values quite a bit and found no change in flight. Thanks for your help on this stuff
So I started from scratch on this stuff. I erased and reloaded the firmware, went thru the setup wizards again and am using a preloaded config similar to my quad. I have enabled the ADVANCED tab instead of the BASIC tab this time. This should allow me to use the gyro vibration filter seen on the expert tab, as the expert features are also active with the ADVANCED tab. So far the roll oscillation is much less at hover than it was before. I do need to turn something down though. When I make a correction with the stick in ATTITUDE mode then return it to neutral I get a wobble. This must be a sensitivity adjustment. I don't want to fly it like this and have it over compensate and crash. So what do I need to adjust?
-
BTW, as suggested, I could just try my luck with the OPTUNE or easy tune in the TxPID tab. I don't have enough extra channels to setup my radio so I can make this adjustable in flight. When I watched the video for using this feature, the numbers that were typed in as limits for the feature seemed arbitrary and did not seem to have a discernable value in the final PID values. While they did change the values when the control slider was moved, the values had to be checked and changed to represent the ballpark that the pilot wanted to tune in. This worries me because I don't even know what ballpark these values should be in. In the end I want to understand what it is that I am changing and how it affects the flight characteristics of my quad. This way I can troubleshoot my own aircraft in the future.
-
The problem with using someone else's settings is that something fairly small can make some difference.
I always start with stock PIDs.
I think that a good general rule to use with OPTune is to have it adjustable from 1/2X to 2X of stock values if you don't know any different.
Oscillation after a hard bank can actually be several things:
- stick bouncing and settling to neutral. :)
- PI too high
- Needs a little more D term
-
I have attached an image of my current settings for PIDs in ADVANCED tab. I am flying in stabilized mode and am getting this wobble after a stick correction of less than 80%. That means it is still in RATE right? So when you say it needs a little more of this or a little less of that, how much do you mean?
-
What do you mean by "stabilized mode" ?
You should remove TPS at start: why only PD ?
If you are using Attitude as stabilization mode i think you can reduce the values of Attitude/ from 3 / 3.4 to something like 2 / 2.5
-
Sorry, I mean Flight Mode 1= ATTITUDE/ATTITUDE/AXIS LOCK
The attached image is where I am now
-
Raising the Derivative from .00040 to .00060 has fixed most of the wobble. Some still remains, but it is much less aggressive. Not sure if I should keep raising it or quit while I'm ahead. According to the TIP comment in LP when you move your cursor over the adjustment arrows in ADVANCED tab, Derivative controls the speed at which the board will correct in rate mode (inner loop). It seems logical to assume that if the corrections were being made too slowly or that the board was waiting too long that the aircraft could oscillate.
-
Could you guys watch this Video and tell me what you think is happening
https://www.youtube.com/watch?v=WI-iJGjIt7I
-
Since posting the video above I have made some changes. This morning I reduced my PID settings even more as you will see in the image I attached. I also reduced my Accelerometers Filtering from the default of .10 down to .05. This seems to make the aircraft more stable. I will fly again later to see how it is affected in faster flight. I am aware that most of you more senior members advise tuning in RATE mode, however for beginners ATTITUDE is much easier to control when it works as it should.
-
I had vibration issues because of a loud bearing. I mounted the FC with sticky foam with a kind of four corners pillar style approach. Fixed my problems completely. Can't say that's what's going on with you, just my expirence.
-
The video does look like the PIDs are a little too high. If you are using OpTune, just pretend your UOV's are a little lower and ask for a new set of PID's.
If you think oscillation is worse when you use a lot of throttle or that it looses stability when descending vertically, you could try turning on TPS or making the line steeper if you already have it turned on.
-
Thanks for the reply. I have since tried to isolate the FC board even more from any vibration that the motors may be creating. I will do some investigation on the possible motor bearing issue. As you can hear in the video, there is a noise that is not consistent with wind noise but does seem more pronounced with higher throttle. I have since also flipped my CC3D over and will plug the ESC's directly into it to eliminate the connection via female header pins to the PDB. I was worried that this type of pin connection may be to ridged and may also transfer vibration. I will correct the virtual ROLL to reflect the change in board orientation that I have made and give it a test flight. If that does not have a significant affect on the issue, I will have to conclude that I may have a damaged board.
-
I also once had a super thin Chinese POS board that was doing all sorts of weird stuff before it crapped out on me. Weird drifts and occolations. Intermittently. In the end it stopped sending power to the receiver. New board flew sooooo much better but still had a drift. In the end it was a vibration issue fixed with 2$ in double sided stick foam.
-
It seems that it always turns out to me the small things that are the problem. For me right now, most of the hobby is spent on the work bench at night. All I want to do is fly.