RK3588 Reform not booting on battery

Sorry in advance about a mess of a post, but I received my RK3588 Reform Classic about a month ago (from the MNT shop) and recently noticed a curious problem: it doesn’t seem to power-up when not connected to power. This may have a problem since I got it (or soon thereafter, since I had trouble booting Void Linux as discussed below) but did not recognize the cause until recently, since I typically use it while it is connected to power.

Generally, when I start the computer and nothing is plugged into the power slot, the MNT logo briefly appears on the small keyboard LCD display, the keyboard lights up and the screen backlight turns on and nothing happens. If I connect the USB-C to my computer to read the serial messages after waiting a while, it seems to be stalled on the U-Boot prompt.

If I instead boot it when connected to power or when connected to my computer to read the serial output (a USB 3.0 port without power delivery, so the power draw from the batteries is still very positive), the computer boots normally.

On a few occasions, when I start it without power, it barfed up a bunch of messages onto the screen waited for several seconds, then attempted to reboot. Unfortunately, it scrolled too fast to see the actual error message and was just countless instances of (reproduced from a slightly-blurry photo I took):

sysrq: HELP : loglevel(0-9) reboot(b) crash(c) terminate-all-tasks(e) memory-full-oom-kill(f) kill-all-tasks(i) thaw-filesystems(j) sak(k) show-backtrace-all-active-cpus(l) show-memory-usage(m) nice-all-RT-tasks(n) poweroff(o) show-registers(p) show-all-timers(q) unraw(r) sync(s) show-tasks-states(t) unmount(u) force-fb(v) show-blocked-tasks(w) dump-ftrace-buffer(z) replay-kernel-logs(R)

which looks like a list of options, although I did not realize it at the time. On one such instance, after I plugged it into my other computer to get serial output, it rebooted correctly but the battery level in linux kept jumping around (even though the battery display on the LCD was stable and at 79%).

I have also noticed that, on occasion when running, the battery display on the little LCD glitches out a few times, briefly showing 9.9V for every battery, with the battery percentage and “On” disappearing. This happened several times when I was running on batteries today.

It is worth noting that I did open up the reform and briefly removed all 8 batteries (I couldn’t disconnect the battery connectors at the time; I reinserted them all as they were, before closing the computer up) to enable the serial port early on in its life. This was done in an attempt to debug a problem booting Void Linux (from an SD card, I never installed any boot stuff on the NVMe nor did I touch any boot-related stuff on the eMMC, except to add “console=ttyS2,1500000” as the last kernel parameter on the default Reform Debian install). It is possible that this was the same problem, as I never did figure out what the problem was, but I am not certain of this. A subsequent install with a different kernel version worked.

I can currently reproduce the problem on both the default Debian install (running from eMMC) and Void Linux (running off of my SD card). I should also mention that bootloader was the latest (2026-01-28) version, but I did reflash it in case that made a difference (it did not).

Can you reproduce it with a vanilla MNT Debian system image from SD card?

And if yes, can you paste us the serial output from an unsuccessful boot of MNT Debian from SD card when AC power is not connected? Thank you!

Yes, the same thing happens. If I boot without AC power, nothing ever comes up on the screen. If I boot with AC it spends some time doing initial boot things (resizing the SD card’s inode table or something; subsequent boots skip this), then boots to the setup wizard.

This may be a bit tricky. Connecting the USB-C port (on motherboard 3.0) to my computer, which is how I’ve been accessing the serial output, is sufficient to cause it to boot normally.

If the problem is really about whether power is supplied from outside or not (and not some other adjacent property of your system) then this sounds like a classical Heisen-bug. Attaching the USB-C cable to observe your system via serial will also mean that you supply a little bit of power over the USB-C cable to your system. So once you look, the bug is gone…

Maybe you want to go to support.mnt.re and file a ticket there?

Yeah, that’s what I figured I’d have to do, but it doesn’t hurt to ask on here first. Thanks for all the help!

1 Like

USB power blocker :slight_smile:

Unfortunately, I don’t think that would work. If I’m reading the schematic correctly, the USB-Serial chip is (reasonably) powered from the USB-C port, and is therefore powered off if there is no power provided to the USB-C port.