I’m sorry to report that I’ve experienced this issue again: plugged in my Pocket Reform to (a brand new) Anker A2343, smelled a sudden faint smell of burning, and the system controller became totally unresponsive. On inspecting the charging board under the rear case I can see a bubble in one of the ICs. I’ve already reached out to MNT support, but I’m reporting this here because I have a suspicion that it’s related to the specific interactions between the Pocket Recorm and the charger, and I would recommend avoiding the Anker A2343 it for now. I ordered a UGREEN charger instead.
N.b. my A2343 is brand-new because I replaced it after the higher-power usb-c port stopped working. I do use other devices with it, and it could be a coincidence, but I suspect this is also related.
I’ll join in here with my experience of “doesn’t boot”. I’m still looking for a solution, but here’s what I’ve tried.
A) First occurrence
After not using it for about 4 months, the laptop no longer boots. The batteries were apparently charged (display in OLED), but after about 3 hours no response from anything. I charged with the same power supply unit as at the beginning: 102-SWI60C-V-EB-I40-ND, AC/DC WALL MNT ADAPT 5V-20V 60W.
B) Test with only power via USB-PD
Power supply only via USB cable with jumper set on Charger PIN 2-3 → OLED display on for 2-3 seconds but nothing else. Red LED flashes approx. 2sec on and 8sec off
C) Power behavior
Add a USB-PD power meter for reading out the current USB-PD mode and currently drawn power
LED not lit for 8s, then lit for 2-3s
USB-PD power meter 0.5s nothing, then 7.5s 0.5V 5A 2.5W, then PD switch to 20V / 9V, LED starts to light up, 2-3s later everything off, 0V 0A; rinse and repeat
from ~2s mini-OLED display can be switched on, switches off as soon as 0V 0A
with a bit of luck, the laptop had 3s instead of 2s USB-PD, with which the laptop could be “switched on” for 1-2s (keyboard LEDs lit up shortly) → here short UART output; however, full termination at 0V 0A
I would recommend inspecting your charging board to see if it seems to be the same issue as mine. Both times for me the MP2650 (U1) was visibly deformed. If this is the case your best bet would be a warranty replacement of this board.
Good advice, although the boot problem also occurs without the charger board connected (USB-PD powered only), but I inspected the charger board and in particular the U1/MP2650 chip more closely.
I would say the pictures show that the chip is not running to specification, right? The one in the photo, top left of the chip looks like a cracked bubble.
Yes, that’s almost exactly how mine looked, down to the placement of the little bubble. Contact MNT support for a replacement. Out of caution I might recommend using a different charger, too.
The problem you’re describing where the system controller powers off after a few seconds was indeed fixed in firmware a while back. It’s apparently possible to update the firmware without a working charging board but I didn’t try it — I updated after receiving the new charging board. You might find this path easier.
I was able to desolder this IC pretty easily but I haven’t tried to solder a new one on yet. It’s like a $5 part but the fine pitch will be tricky. That might be another option.
I have now also sent the picture to mnt-support. It looks like it is a warranty claim.
Because the laptop won’t stay on for 4 seconds, I can’t update the firmware. Let’s see what the cause is.
Desoldering and resoldering the IC would be an option for an old device where it is clear that the chip has died due to old age, but this is certainly not the case here.
We’ll contact MPS (makers of this chip) next week to get their opinion on how this failure mode can happen. We also offer free replacements of this PCB to anyone who runs into this issue via warranty. Just mail your order number, order date, postal address (incl a phone number for UPS if possible) to support@mntre.com.
A few people are looking into this right now and there’s also an improved PCB layout in progress, but i need some opinions from MPS to make sure that our mitigations will stop this from happening.
minute, thank you for taking this issue seriously and looking for improvements. I would buy a reform again at any time, the focus on quality and open source is inspiring.
Thanks to mnt support and the people behind it. My Pocket Reform is working again. In fact, it was the rare problem of bubbling on the charge controller chip. With a new charging board and firmware update, it’s working again. So I’m writing this text on the Pocket Reform.
I’m still in the debugging process with MPS but one hint I can already pass here is that their theory is that the charging chip has 21V max input, which is pretty close to the 20V we request from USB-C PD. They have a theory and some simulations that using a long cable (maybe 2m cables?) can cause ringing and push the voltage over the 21V max for a moment, potentially killing the chip. So it might be safer to use a short (1m) cable for charging until we know for sure / have an updated charger board. We could also limit the charging voltage to something below 20V in the firmware in the meantime to make extra sure. (cc @zeha)
Just a quick update while compiling a report for MPS who have agreed to look at actual broken boards for failure analysis:
I know of 10 confirmed cases out of a total of 458 shipped Pocket Reforms, so the failure rate is around 2.2% so far. Still, we’re working on trying to understand the root cause. I’ve done some tests with overvolting the chip on purpose, but couldn’t break it so far even at 32V input (the theory was that there could be voltage spikes on hot plug with long cables). I’ll keep you posted.
I’ve not opened mine to investigate, but suspect I’ve got the same/similar issue. The Pocket just froze while watching a YouTube and the display gradually faded out. I promptly unplugged, but around the keyboard smelled slightly of that horribly familiar electronics smell (thankfully no “magic smoke”). Don’t plan on attempting to switch on or charge, and have opened a support ticket too…
