IMU not working properly on X12s #2786
Comments
Kevltan
added
bug 🪲
Kevltan
changed the title
|
Raphael has commented in the dev chat that there is a clash between touch screen controller task and imu task, so this may be related, and as such is on the radar to be fixed, but that will most likely be a 2.8.1 fix. |
|
*************************************************************************************************************************************
*************************************** SOLUTION TO THIS ISSUE IN LAST COMMENT **************************************
*************************************************************************************************************************************
Nice one, glad to here this Peter. Can live with at the moment but look
forward to the fix in 2.8.1 🙂👍
…On Thu, 24 Nov 2022, 09:09 Peter Feerick, ***@***.***> wrote:
Raphael has commented in the dev chat that there is a clash between touch
screen controller task and imu task, so this may be related, and as such is
on the radar to be fixed, but that will most likely be a 2.8.1 fix.
—
Reply to this email directly, view it on GitHub
<#2786 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ANPFAW6EJZKGCKZWHBQC2Y3WJ4WELANCNFSM6AAAAAASJLLUUQ>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
I am having no problems with IMU since changing to 2.8.0 "Flying Dutchman" now working everytime. Has something been changed? I have noticed that there is no call for IMU_LSM6DS33 in System/information/version/OPTS:, which I called for in |
|
That is quite possible... Unless you're using the LSM6DS33 and not on the X12S,
In this particular snippet it should still correctly use the X12S IMU config, but perhaps it was going wrong somewhere else. Or you've just been lucky so far :) |
|
There are still problems with the IMU freezing when going through various model/system pages, but it can be resolved with a power re-cycle now. |
|
A big problem, because IMU and display interfere. Raphael already has a great idea and I hope he solved this bug in 2.8.1 as well. |
|
I temporarily disconnected P8 and I2c wiring from the touch panel to eliminate all hardware additions. Tried IMU functionality and still failing with the same shortcomings as decribed earlier, this isolates the touch panel itself from the equation. So believe this is definately a firmware and/or SD card software conflict somewhere and NOT hardware related. |
|
Some interesting information I found which might help the developers pin point the problem. I have 2 SD cards one 8GB and one 16GB, both formated and named the same with EXACTLY the same software copied on them. The IMU does not work with the 8GB but does with the 16GB when TX booted, but they both load with everything else working normal. This is 100% repeatable.....Any ideas?? |
|
Still experiencing same problems, IMU sometimes works and sometimes doesnt when powering up or during editing it freezes. Always works when initiating boot loader and immediatley exiting using the exit prompt but after this can still freeze whilst editing pages. This is seems to be NOT related to a hardware probem as is still repeatable with the touch panel disconnected. Is the IMU problem "historical" and has not been picked up before because of its limited use?? |
|
Only someone with access to the hardware is likely to be able to fix it. @raphaelcoeffic is one such person, and as was commented on earlier, has some idea as to what could be the issue ;) As far as whether there is a historical problem, that is anyone's guess atm, as since you point it has had limited use. However, I'm inclined to think it is still related to the TP support, regardless of if one is connected or not, as that is fairly new and an unusual combination. |
|
OK that's great, so thinking it is a clash/conflict in the firmware coding. But
finding it strange that it works everytime from the bootloader and exit🤔
…On Thu, 29 Dec 2022, 00:20 Peter Feerick, ***@***.***> wrote:
Only someone with access to the hardware is likely to be able to fix it.
@raphaelcoeffic <https://github.com/raphaelcoeffic> is one such person,
and as was commented on earlier, has some idea as to what could be the
issue ;) As far as whether there is a historical problem, that is anyone's
guess atm, as since you point it has had limited use. However, I'm inclined
to think it is still related to the TP support, regardless of if one is
connected or not, as that is fairly new and an unusual combination.
—
Reply to this email directly, view it on GitHub
<#2786 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ANPFAW6O3XVO4EMHGKWMQLLWPTKMXANCNFSM6AAAAAASJLLUUQ>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
All problems with IMU not working on start up caused by having a splash.png file too large in the System/Images file on the SD card. The size was 220kb this is too large and seems to upset the IMU on start up. I used an online .png file compressor to reduce it to 89kb without changing pixel count of 480 X 272 https://tiny-img.com/png-compressor/. I have tried this on cards from 256Mb - 32Gb standard and SDHC format. All formatted to Fat32 default allocation. IMU has not failed since! So if your running your own splash screen at start up it can't be much over 100kb in size. The ones I created online in https://www.air-rc.com/ or https://www.skyraccoon.com/ were too big and needed reducing. |
|
THIS NOW WORKS FINE AFTER THE ABOVE ISSUE ADDRESSED |
|
@Kevltan the issue with picture size / time to load is now obvious. The sensor initialisation is done in the UI task, while it is queried in the mixer task. This can never work and basically amounts to a race condition as which one of the two happens first:
|
|
Ok so would the answer be to limit the image file size of the Splash.png
and if it is to big, it is thrown out before startup?
…On Fri, 10 Feb 2023 at 17:07, Raphael Coeffic ***@***.***> wrote:
@Kevltan <https://github.com/Kevltan> the issue with picture size / time
to load is now obvious. The sensor initialisation is done in the UI task,
while it is queried in the mixer task. This can never work and basically
amounts to a race condition as which one of the two happens first:
- error counter in mixer task is incremented up to 100, at which point
the sensor will not be asked again,
- and when the sensor is actually initialised (hopefully before the
error counter reaches 100)...
—
Reply to this email directly, view it on GitHub
<#2786 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ANPFAWZ4ALUNZJC7C6XCODDWWZYVBANCNFSM6AAAAAASJLLUUQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
No, you don't take aspirin when you have a bullet in your head: you go to surgery. Here it means, we fix it properly (see 828b5c0). |
Is there an existing issue for this problem?
What part of EdgeTX is the focus of this bug?
Transmitter firmware
Current Behavior
The IMU problem is back again with touch panel modification!
I am re-opening this issue because I believe there is still something wrong here. What I have noticed now is IMU won't work all of the time on power up, but if I hold the T1, T4 trims in and and power on as though I was going to do a firmware update in bootloader and then exit instead, it works every time! Anybody any ideas? I really am giving up here because there seems to be no logic in it. I think I've fixed and then I haven't! Still thinking there must be something in the startup/powerup routine that is different to the boot loader and exit routine - works everytime from the bootloader🤔.
Expected Behavior
IMU not functioning every time radio is switched
Steps To Reproduce
Version
Nightly (Please give date/commit below)
Transmitter
FrSky X12
Anything else?
2.8.0 selfbuild RC4
The text was updated successfully, but these errors were encountered: