- May 2, 2017
- 7,398 (3.73/day)
- Back in Norway
|Processor||AMD Ryzen 7 5800X, 142/95/120, PBO +150Mhz, CO -7,-7,-20(x6),|
|Motherboard||ASRock Phantom Gaming B550 ITX/ax|
|Cooling||LOBO + Laing DDC 1T Plus PWM + Corsair XR5 280mm + 2x Arctic P14|
|Memory||32GB G.Skill FlareX 3200c14 @3800c15|
|Video Card(s)||PowerColor Radeon 6900XT Liquid Devil Ultimate, UC@2250MHz max @~200W|
|Storage||2TB Adata SX8200 Pro|
|Display(s)||Dell U2711 main, AOC 24P2C secondary|
|Audio Device(s)||Optoma Nuforce μDAC 3|
|Power Supply||Corsair SF750 Platinum|
|Keyboard||Cooler Master MasterKeys Pro M w/DSA profile caps|
|Software||Windows 10 Pro|
First a disclaimer: I barely know what I'm talking about here. So take everything with a grain of salt - I'm just learning about these things. Still, none of this should have any potential to damage anything on the board. Your multimeter will input some voltage during continuity testing, but nothing that would come close to damaging any of the components present.
First and simplest test: check for continuity across the EPS12V connector.
Second, slightly less simple: check for shorts on any of the capacitors surrounding the VRM. If the larger caps are through-hole, measure the legs on the back side, with the black probe on a separate ground point (screw hole, rear I/O casing, EPS12V, whatever). One side of the cap should be short to ground, but not both. You can do this for the inductors/coils as well - they shouldn't be shorted to ground at all. While you're at it, SMD caps around the VRM (if there are any - doesn't look like there are many here, but there may be on the back) are similarly worth checking.
The MOSFETs look like QFN-8 pacakages, which means their pin layout is this:
(The connections on the side of the package are also drains, like you can see on the back of chips like this one)
The dot on the package (clearly visible on your photo, less so on the illustration above) indicates pin 1. The gate is pin four, so same side, opposite end of the row of pins. Drains are on the opposite side.
The source is the voltage input, the gate is the input for switching the MOSFET on or off, the drain is the output (into the vCore power plane through the inductor). In a VRM with discrete MOSFETs like yours (as opposed to not integrated power stages) you'll have high side MOSFETs where the source is +12V, and low side MOSFETs where the source is ground. I don't know which are the high side and low side MOSFETs on your board, but in either case, only three pins on one of each pair of MOSFETs should be shorted to ground.
Given that the upper and lower rows of MOSFETs are different models (the upper ones seem to be lower current ones, ~40A vs. ~70A for the lower row?), it seems logical that one row is the high side and one row is the low side. Which is which? No idea. I would guess the upper row is high, but that's just guesswork based on them being generally oriented towards the 12V EPS connector.
What I would do for MOSFETs, especially given how small and hard to reach those pins can be, is just try and probe everything. Only three pins on one side of one of the MOSFETs in one of the rows should show a direct connection to ground.
If there is a short somewhere: congrats, you've confirmed that the fault is somewhere around the CPU VRM. Unfortunately there's still work to be done. It might be a failed MOSFET, it might be a dead capacitor, and it might even (but much more rare) be a failed inductor. Identifying which it is, and which one (or several) are failed is not trivial.
The easiest way to tell is by injecting voltage and seeing what gets hot - which requires a current-limited low voltage power supply (if it isn't current limited you'll just blow something up - either the power supply or whatever is shorted - or both!) and some way of detecting heat, either a thermal camera (expensive), rosin vaporiser (cheap and relatively simple to use, but essentially repurposed drug paraphernalia, FWIW), or touch (free and easy, but very unreliable). The "screw it, I have no idea" high effort/low cost solution is to just desolder successive components until the short goes away. (Of course, you'll most likely need a decent quality hot air soldering station to desolder SMD components on a heavy power/ground rail like in a motherboard.)
If you're interested in learning more about VRM repair, I'd strongly recommend Adamant IT's youtube channel. Lots of repair videos (lots of laptops with shot VRMs!), but also some great intro videos covering VRM layouts and components, board repair basics, and more. And IMO a lot more down-to-earth and approachable than someone like Northridgefix or Louis Rossmann. They're just about at the opposite end of the UK from you (if you're in Hull as mentioned above), but they might still be able to fix the board for you if you want.