Switch OLED no boot. No shorts / apparent water damage - weird power behavior

Hi there!

Got a OLED switch here that will not boot. Here’s what I have tested so far:

Plugging in a USB charger with an inline power meter gives 15V, 0.45A at first, which jumps to around 0.75A after a couple seconds (second stage boot?), then will either drop back down to 0.45A, and alternate between the two current levels or turn off completely. Then the process repeats.
Leaving the switch plugged in DOES charge the battery over time.

On a visual inspection, no apparent water damage, no missing components. No shorts when checking capacitors around M92, PI3, BQ, MAX77812, fuel gauge. MAX77620 shows shorts but I believe these are low impedance caps (see next section).

On injecting 4.2V in VSYS (battery and screen detached), switch usually gets stuck in first stage boot. current draw between 100mA and 145mA. Measuring MAX77620 Caps. 1.0V, 1.1V, 1.8V line present. 1.35V line absent (I think this is normal?). 3.3V and 1.8V lines present.

On injecting 4.2V in VBAT, switch goes from around 150mA to around 350mA after a few seconds (second stage boot?), but will fluctuate between 275mA and 350mA. Eventually power draw goes down to 0mA after 20-30 seconds. The process may or may not repeat on its own without me pressing the power button.

When power draw is at 275mA to 350mA, 1.35V rail is present, however 1.0V rail seems to drop to 0.6V. Is this normal?

Any suggestion or help is appreciated! :slight_smile:

Okay so minor update. Replaced the BQ chip on a whim and the power behavior seems to have changed somewhat. On USB power I’m getting a consistent 15V 0.7A (current dropping as I assume the battery charges up). Power no longer drops to 0.45A or turns off, so potentially I’ve fixed one issue?

On the other hand, now injecting VSYS or VBAT now remains stuck in first stage boot at 100-120mA. Weird. All rails (besides 1.35V) are still present. I’m following the retrosix guide (which I understand is for the original switch models, but I believe the power circuitry is quite similar on the oled switch), and it’s pointing me towards a bad fuel gauge perhaps?

Again, any pointers are appreciated!

Replaced the fuel gauge and no dice. I imagine it is possible I messed up the install, but there’s been no change to the behavior

I’m inclined to hope that it might be a bad emmc. I’ve ordered a PicoFly and will attempt to boot into Hekate. Is that a reasonable course of action? Is there something else I should try first?

what label/ brand emmc on it?

The emmc is a Samsung KLMCG4JETD

Installed a Picofly, which flashes a single yellow led (which i believe indicates a successful glitching), but nothing appears on the screen. May be a broken LED driver or busted screen, but this is seeming more and more like a No Fix.

Probing the I2C lines for the MAX77620 and MAX77812, indicated the SDA line being stuck high. I removed the MAX77812 IC and SDA seems to have returned. Power draw is stuck at 135-140mA which is consistent with a bad second stage PMIC according to RetroFix (makes sense). Will replace this IC but I’m not optimistic.

Okay so probing around the LED driver IC, none of the three output voltage rails are being generated. I understand the IC is a TPS65633 which is an OEM part. And that is S2DOS04 is a an equivalent part that appears on some of the OLEDs.

Would anyone know if the TPS65632 is a drop-in replacement (since these are in stock on Digikey)?

If I’m correct, these three test points correspond to the three output voltage rails from the LED driver. In theory I suppose I could try injecting voltage into these points to test whether the display is functional? Does anyone know what these voltages should be? I do not want to rely on the TPS65632 datasheet, in-case it isn’t a drop-in replacement.