On Fri, May 22, 2020 at 10:05:05PM +0200, Mac Goever wrote: > Hi Sean, > > thanks for the analysis. I was a little unspecific about the dmesg: > The disconnects are explainable by me unplugging the cable. This means > all the disconnects are explainable. The resets are not, but I > suspected them to be normal. My Ubuntu Laptop does not do the reset.
The resets are not normal. I checked the logs on a number of RPis I have that use USB devices, and none of them exhibit such resets. > The Power supply is a modded 450W ATX supply I modded. I had the > experience with wall warts, so I chose the big cannon for testing ;) > > I will do more testing on the power side anyway. One slightly counter-intuitive thing to watch out for with ATX power supplies is that they may not behave ideally with very lightly loaded. Especially if it's an older unit meant for PCs that drew large amounts of power on their 5V busses, a Pi's draw may be too low for it to keep a stable voltage. Some of the answers here go into more detail: https://raspberrypi.stackexchange.com/questions/29190/can-i-use-a-standard-pc-power-supply-to-power-a-raspberry-pi If you happen to have an oscilloscope handy, it might be worthwhile to see if there's any oscillation happening on that rail. It would also be informative to get a trace of the voltage sag when the scanner first initializes. But this is probably getting off-topic for the sane-devel list. --Sean
