SDK • Re: Receiving data from Pico on Linux

But in my experience the data is often corrupted ...

Code:

pi@Pi4B:/tmp $ cat /dev/ttyACM0 > /tmp/grabbed.txt^Cpi@Pi4B:/tmp $ cat /tmp/grabbed.txt1609478190160916094781921609478193

Is this random, or just near the beginning of transmission? Related to cable length? Doesn't USB have error correction?

Corruption appears throughout a run, not necessarily at the beginning. I would colloquially call it "random" but am not convinced it truly is as the explanation for the corruption may have a more concrete explanation - such as buffer overflow, data being dropped somewhere.

It's notable here that data was truncated - I should have wrapped the data with start and end characters to make that more obvious ... print("<{}>".format(time.time())) ...I was using vanilla MicroPython on my Pico W for my test and corruption is always present. The weird thing is that when using 'minicom' I see no corruption at all. When I have in the past written my own serial receiver in Python I see none there either. This is what has me excluding cable length and transmission issues, has me believing it's something with using 'cat' on the USB serial device the RP2 presents, the Pi, OS, kernel or drivers, even though I am using the vanilla versions.

I haven't investigated deeply nor attempted to qualify the issue (*) as I've never needed to do things this way. You usually need something more complicated to handle Pico reboots, disappearance, reappearance, potentially non-reappearance, of a USB serial device as you have found.

Though I did notice, when producing my latest dump as above, that running in the '/tmp' directory -Shows far less corruption, an often none other than in the first line, than when run from my home directory -Go figure!

Statistics: Posted by hippy — Wed Oct 23, 2024 3:13 pm

---