Open IanSB opened 2 years ago
A voltage scaling bug that only affected Pi Zeros was introduced in May 2021 and fixed on 14th November. Although it reports a voltage of 1.8V, the hardware actually caps this to just under 1.4V - not ideal, but not catastrophic.
Firmwares built since 14th November, include the current stable
and oldstable
branches includes this patch.
Firmwares built since 14th November, include the current stable and oldstable branches includes this patch.
Thanks, I'd been using the files from the most recent raspberry Pi OS image I had which were dated 29/10/21 and I'd been getting some reports of lockups with my latest beta releases which is why I started looking at it. I thought I'd also tested the most recent firmware build but I guess not.
the hardware actually caps this to just under 1.4V - not ideal, but not catastrophic.
I noticed that vcgencmd returned 1.4V
@pelwell Is there a Github issue for the voltage scaling bug?
I think you are reading it.
@pelwell
the hardware actually caps this to just under 1.4V - not ideal, but not catastrophic.
Has this set the over voltage warranty bit? (The project uses force_turbo=1 but doesn't use overvoltage settings)
Has this set the over voltage warranty bit? (The project uses force_turbo=1 but doesn't use overvoltage settings)
No, you need an explicit over_voltage
settings (plus force_turbo=1
) to set the warranty bit.
I'm working on this bare metal project for the Pi zero: https://github.com/IanSB/RGBtoHDMI/wiki
I recently started using newer firmware (bootcode.bin, fixup_cd.dat, start_cd.elf) but the default core voltage has increased from 1.35V to 1.80V on the Pi zero or zeroW on all firmware from the 14-jul-2021 commit. The most recent one that works for me is the commit on 21-apr-2021
Other pi models (pi1, 2 ,3 zero2) have sensible voltages between 1.20 and 1.35 volts with recent firmware
My software doesn't attempt to change the voltage either via the mailbox or using config.txt
21-apr-2021 commit:
14-jul-2021 commit and all later versions: