I was just surprised because I was using an official Raspberry Pi branded power supply (the one for the Pi 3 outputting 2.5A). That's more than good enough to run a Pi 3 with a fair amount of accessories usually. Shows how adapters can be really badly made and introduce a lot more resistance than expected.OK it was power, not a surprise really. On the original Zero, so 1 ARM11 core, I still have some setting in config.txt. Maybe not needed anymore, but in the past there were strange kernel crashes if the VC4 started encoding video, that seem to drop on-chip core voltage too much. I have wires soldered on the Zero, the USB power connector is more a burden than an advantage for a small board like the Zero.After having a closer look at my setup, the power button adapter on the cord was causing an abnormal drop in voltage (found via testing it with a USB volt/amp tester). The losses on it seem to have caused things to misbehave during reboots. Removing the power button adapter caused things to function normally.
You might want to buy an USB power measure device, I have one that cannot log (via Bluetooth) but still very helpful in seeing want voltage and current do. It is USB-C PD compliant, so also can track 10's of Watts and up to 20V. I have also made some custom myself for 12V, with logging 24/7, that also gives a lot of info over time, as for example if and when a backup is running in the background, there is higher power consumption of course.
Now, I have been trying out a new case for this Pi, the Argon40 Pod case. Their LAN-USB module seems to somehow introduce a bug where the ethernet's MAC gets randomized on every boot. Using my own adapter, things work as expected.
Their ethernet uses the RTL8152 fast ethernet chipset while the adapter I have is an RTL8153 gigabit ethernet chipset. That's the only difference, no settings changed. So strange.Statistics: Posted by Randomp — Thu May 08, 2025 9:58 pm