Faulty power circuit

[hackerfantastic]

I ordered a Pinebook Pro in August 2019 and my unit arrived sometime in December. I had a US ANSI keyboard model. However over the last few weeks I started noticed weird power behaviour, the laptop would be left on overnight on charge but in the morning I would find the battery completely depleted or the device had randomly powered itself off for no valid reason. I have not taken the device out of my house, and just move it between a few rooms - so I put it down to linux being linux. I would unplug and replug it, power cycle and notice the battery seemed to be getting a charge but that there did appear to be some kind of power related issue but did not think much of it. Until it died.

Now, the pinebook pro will not even charge on the barrel jack. The pbp died completely 3 days ago, when I connect two different PSU's (which I tested at giving 5V out each with a multimeter) on the barrel jack, I got absolutely no red light indicating power, the laptop was completely dead. I left it charging via USB-C which did give me a charge indicator, it took several hours to get enough charge to power on the laptop. Once I booted the laptop I noticed that the battery indicator stated "discharging 0%" and after a few minutes with the device powered via USB-C, it would just switch off anyway (after a good 12 hour charge).

I opened the case, disconnected the battery, connected the bypass wires and then powered from barrel jack - nothing, the barrel jack is completely dead, however the USB-C power input would work but was not able to give it enough power to complete booting from the NVMe. The pbp is hosed, so I looked to see what the fault could be.

I have identified that one of the transistors besides where the "bypass" cables are wired appears to have let out the magic smoke, additionally I am disappointed at the state of the soldering job on those wires - they are basically soldered on in such a way that undoubtedly these bypass cables have moved around over time introducing a short - the metal contacts of the wires runs right across the flex cable battery connector, this I believe is what has caused the transistor to blow and break the charging circuit. The cables are half exposed when connected to the PCB and are soldered on in a blob method right across the battery contacts and alongside where the blown transistor appears. Picture below.



I have ordered a new replacement pinebook pro mainboard as unfortunately I was using this thing for actual coding work at night time and aside from the exploding transistor of death it's been a rather decent little machine! in the interim I will try to swap out the blown transistor - does anyone know what component that is on the board above, its the middle transistor that has clearly cracked open? One thing I would definitely do on my new unit when it arrives is remove those bypass cables unless I actually intend to power it without a battery direct from DC, as undoubtedly I believe they are the reason the board short circuited on the power circuit at some point (at least that is my best guess right now). Oddly, those are both soldered with VCC + GND connected to the same pads, they should be removed if not required. Is anyone else noticing the poor connectivity of these wires? Is this a known fault? Have others experienced it?

The machine was powered on most days, didn't have anything plugged into it aside from USB controllers and an RTL-SDR so nothing else can seemingly explain the transistor of death.

[Arwen]

@hackerfantastic Sorry to hear about your PBP dying.

There are schematics on the Wiki. If you find some details that are useful and do not exist, add it to the Wiki. (Or let someone like me know, I've added other things to the Wiki.)

From reviewing page 11 of the main board schematics, it appears it might be one of these parts:

• Q18, a PL3401 P-Channel MOSFET, in a SOT-23 packge
  
• Q19, a 2SK3019 N-Channel MOSFET, in a SOT-523 package
  

Can't really tell without the board layouts. But perhaps someone with more experience could by the device package, or examination of the board traces.


As for letting the smoke out, Sun Microsystems had a design flaw in their first generation Sun Ray thin client. It had a lower wattage resister than it should have had in one place. Every now and then, too much power would go through it, and pop, the magic smoke was released. Please note their was no fire danger. I actually worked at Sun Microsystems in 2003, (before Oracle bought them). One of my co-workers had that happen to her. We laughed about it, as it was an easy swap out. Take the broken one to the mail room and get a new one.

The point of my telling you this story, is that we probably need more eyes on the schematics of our PBPs to find things that MAY be like that example from Sun. I do have some digital design experience, but certainly not enough for a full review. Plus, we still need the board layouts & BOM, (Bill Of Materials, aka parts list), to be released.

In your case, replacing the transistor with a similar but little bit higher current part may help too.