MCU memory corruption when sending data from MPU via SPi

[hmarthens]

#17305

Hello,
I wanted to kindly ask what could be the reason behind this bug that we are sometimes facing when sending data to MCU from MPU over SPI. Sometimes, when sending data packets from Raspberry Pi to MCU, we see that the data is reset on MCU, and this can be seen through the free-master variables which also change to initial values. What could be the reason behind this, and have you experienced such bug ?

Thank you.

• This topic was modified 10 months ago by hmarthens.

[Leo]

#17310

Hi Hmarthens,

No, I haven’t experienced such a bug. Can you help with the following questions to help further tackle this issue?
1. What steps have you taken that can reproduce this bug?
2. What’s your SPI baud rate config
3. When u say it reset on MCU. U mean the value changes back to the init value of SPIFromPi datastore or it’s been overwritten by the SPI commservice as we discussed previously?

[Leo]

#17309

Hi Hmarthens,

No, I haven’t experienced such a bug. Can you help with the following questions to help further tackle this issue?
1. What steps have you taken that can reproduce this bug?
2. What’s your SPI baud rate config
3. When u say it reset on MCU. U mean the value changes back to the init value of SPIFromPi datastore or it’s been overwritten by the SPI commservice as we discussed previously?

[hmarthens]

#17311

Hi Leo,

Yes, I was able to reproduce the error sometimes when sending data from Pi, please see the script that I use attached above.

1. Sometimes when sending it occurs, so it is not something I can reproduce with certainty
2. From the python file, the speed is 1 MHz
3. When I say it reset, I mean all parameter values on MCU are re-initialized…like all other variables in program not related to SPIFromPi….there could be memory corruption I suspect with arrays writing…kindly also see the document on how I unpack the data from SPI, I hope it is helpful.

Thank you very much!

• This reply was modified 10 months ago by hmarthens.

Attachments:

You must be logged in to view attached files.

[hmarthens]

#17406

Hello,

I am following up on this request to see if anything has been found, or if there is any suggestions ?

Thank you,

[Leo]

#17418

Hey Hmarthens,

Sorry for the late response. I tried to reproduce this by replacing your midpoints.csv and flashed the sample project EmptyProject.slx to my Flexbench. I had SPI communication running for quite a bit. But I wasn’t able to reproduce this issue.

1. Can you try flashing this EmptyProject.slx attached below. See if you can reproduce this issue.
2. Can you run ps -ef | grep ‘python’ and post the response here? I wanna make sure there is no other python process that is currently running.

Leo

• This reply was modified 9 months, 1 week ago by Leo.

Attachments:

You must be logged in to view attached files.

[hmarthens]

#17593

Hi Leo,

Thanks for the response. I saw it a while back, and actually I changed the program and the way of sending data over SPI, now sending much less data…the issue above never occurred again(after this change). I suspect it was related to the sending of a lot of data via SPI.

[Author]

Posts