It was reported some time ago that is recommend not to put the buzzer very near FC for that reason.
I did do a autotune after setting PWM to oneshot125.
It pushed up all the PID’s.
Old parameter file:
ZMR250 AT Yaw Ajusted 9-8-2017.param (13.3 KB)
New parameter file after ajusting ATC_ANG_PIT_P,15 and ATC_ANG_RLL_P,15 (both was set to 18 by AT). Ajusted ATC_ANG_YAW_P to 9 as well ( set to just over 13 by AT):
ZMR250 AT Oneshot125 All axis(ajusted) 11-8-2017.param (13.3 KB)
I did every axis with a fresh battery.
Roll axis .bin file:
Pitch axis .bin file:
If you could maybe check if all is well and for any recommended changes
I will do some flights tomorrow and report back.
Yes I know but with Pixfalcon it is factory soldered to the FC board. That is why I desoldered it from the FC board and mounted it out side the enclosure as far away as possible.
@henwilsch yes, it generates an IMU mismatch and also some magnetic interference. But this was known before the Pixfalcon was released…
can you give an example on how to determine if rate noise is too high, i.e. where in the log can it be seen. It’s not acc/vibes - or?
I am planing new tests with a smaller copter, which is hard to tune but has a low vibrations (Autotune consistently fails to tune Roll + noisy Pitch and Roll). It is either is a little twitchy or has dangerously low ATC_ACCEL_PR_MAX values. Hence, I want to check for the rate noise you mention before further tests.
I’ll then try a bunch of different setting for autotune and report.
Ok so how do you propose Pixfalcon owners should deal with this problem. I don’t suppose that they are all electronic technicians.
I want to mention, those are exactly my symptoms.
Vibes in the two common ways of checking are low.
Autotune either gives me very low atc_accel, or it’s wobbly… (I wouldn’t call mine twitchy… Mine is always too slow… Although w/ 3.3, it was sometimes twitchy… Since then, it’s been slow)
I mentioned a couple days ago I spent a few min manually tuning, and liked the results more…
I made a couple graphs comparing my roll/pitch vs desired roll/pitch. (I tried to get the scale and flight conditions similar)
The top one is my manual tune w/ Roll P&I = .09 Pitch = .12
The bottom is the auto w/ Rll P&I = .06 and Pit = .08
I haven’t had a chance to auto-tune w/ 3.5 and a higher aggr. I will this weekend, and post results.
I discoverd it quite a while back, when I mounted the buzzer on the same vibration damped platform as the Pixhawk 1. Which was no problem until the tones for the flight mode switch were introduced. There is a longer discussion on the old google groups list about it.
Maybe this helps: http://ardupilot.org/copter/docs/parameters.html#ntf-buzz-enable-buzzer-enable
I just watched the video @Leonardthall posted and the saw the IMU mismatch in Mission Planner. This is definitely the vibrations induced by the buzzer.
My post was aimed at helping people and maybe Leonard to identify the problem with some Pixfalcons. I would rather disable a buzzer instead of a IMU.
I own a Pixhawk as well and believe me the buzzer is nearly mounted in another country, so to speak.
Why do you think this is vibrations induced by the buzzer?
Having a buzzer on the board or in close proximity to the autopilot because it can cause the IMU’s to clip. This problem in the video appears to be a different problem to me. I am happy to be corrected!
I look at the results of Autotune. I have not been able to get a rate noise measure into the logging yet.
You can also do a full rate logging flight and then look at the FFT results in mission planner.
Can one set the buzzer to disabled in the code or not to beep on mode change?
Ok, I must admit that I am not sure it is the vibrations. I just can’t find the discussion(s) about it. I first realized it was related to the buzzer was when I had a battery failsafe and looked at the mag data. But due to the IMU mismatch the devs involved in the discussion pointed to vibrations. As soon as I switched to an older firmware (with no buzzer tones when switching flight modes) or when moving the buzzer away from the damped platform where the autopilot was mounted on, all was ok. So all tests point to the buzzer. But not necessarily vibrations.
I have not tested it but http://ardupilot.org/copter/docs/parameters.html#ntf-buzz-enable-buzzer-enable should do it. You should be able to test it on the desk. If I remember correctly this does not disable the startup sounds - but the flight mode switching sounds.
Do you see it from the final PIDs or the autotune data logged? If it is the autotune data it would be awesome if you could post two screenshots - one with noise and one without - for comparison.
Here is a picture of the buzzers I removed from my Pixfalcon controllers:
The one on the left caused the interference when switching to Alt Hold and the one on the right seemed fine.
I replaced them both with the buzzers I cannibalized from this board and mounted them far from the FC
Hope this helps someone.
Thanks Thorsten, I will investigate.
Here it the link to a discussion on drones-discuss about interference of the buzzer:
Randy mentions that “This is definitely vibration.” But some of my tests (see last message) indicate it could also be something else.
That works as you explained.
Also found this comment under the video in Youtube:
“Cédric DELAIN11 months ago
I’ve solved the “rocket jump” when I switch to Altold or Loiter mode by disabling the internal buzzer of the Pixfalcon. I have a deep look inside my logs and I discovered that accZ was affected at the moment I select one of these mods. I’m currently flying with 3.4 RC2 and I don"t know if this parameter exists in the previous version (cf arducopter wiki). Hope this help.”
I did some flights and I am happy with the performance. A lot more stable in the wind.
Thanks for all your help Leonard I really appreciate it.