Bad Mobo VRM - Symptoms?

[Cableaddict]

This will be a long post, but someone please, PLEASE try to answer my questions at the end. I can’t find this info anywhere:


I suspect my mobo’s vrm circuit is wonky. (details, below) However, I don’t yet have the knowledge to test it directly. REGARDLESS, I’d like to know if the following symptoms do indeed suggest a bad vrm, or if my problems likely lie elsewhere.
———————————

Background:

EVGA mATX X99 mobo, 2 years old.
Win10-64.
Used for live music performance, and the system had been rock solid for about 18 months. - Then it got rained on just before an important gig. I had no choice but to run it anyway, holding my breath. It booted & ran, but blue-screened about 15 minutes later.
————————————————————————

The symptoms:

Since the above debacle, I get bluescreens, apps crashes, and restarts maybe 2X a day, if it’s run continuously.
The ram, drives, etc all test fine.
Nothing has changed with my software.
I’ve deleted & replaced most drivers, and with the same driver versions.

There also doesn't appear to be any micro cracks on the mobo, which is common with portable rigs. No problems when rocking or shaking the thing.
—————————————————————————

WHAT I’VE TESTED:

Naturally, I suspected a bad psu or a bad mobo.

### I checked the voltages with CPUID and Speedfan, and the 12v rail read very low. ~ 8.9v on one app, and 9.5 on the other.

Well, lots of threads on the internet indicate that you can't really trust these apps to give accurate voltage measurements. It's said that some mobos "report the voltages differently" and so the readings are off. I find that hard to believe, but OK it's possible, so I tested my Corsair HX-series psu with a multimeter, under load:

12.25v, and steady as a rock.

Regardless, since I don’t have an oscilliscope to test for ripple, I bought a new Seasonic Prime 750w psu last week and swapped it in. Afterwards, I got the SAME READINGS IN THE SOFTWARE, and I’m still getting occasional bluescreens & reboots.

### It’s important to note that the readings in Speedfan & CPUID are NOT JUST LOW, BUT MUCH LESS STABLE than the readings directly off the PSU. The software readings don’t fluctuate by much, but they DO fluctuate continuously. (with both psu’s.)

And so:
—————————————————————————

MY QUESTIONS:


1: Is it really true that some (expensive) mobos have an architecture that gives radically-false 12v readings to both Speed fan and CPUID? (I find this hard to believe. One would think the software would then have multiple settings, to compensate.)

2: Is it true (as I’ve read) that some mobos take 12v into the vrm, and convert it? - Do they also pass some 12v through, for use with some hardware?

3: Where exactly does CPUID, Speedfan, etc read the various voltages FROM?

## 4: Does the fact that the software voltage readings are fluctuating slightly indicated a problem?

### 5: The bottom line: Does it sound like this mobo should have it’s vrm circuit tested, or should I be looking elsewhere for my problem?


6: Where can I find a VISUAL guide to testing the vrm’s output voltages?

-------------------------------------------

Thanks in advance !

[kc8adu]

Re: Bad Mobo VRM - Symptoms?

i would tear it down and wash the board and ram.simple green is good for this.
pull battery and wash it in the sink.
scrub with toothbrush and final rinse with alcohol.
saved many laptop spill victims like this.
after final rinse blow off with compressed air.
preheat oven to 250f.
and bake it dry about 1 hour.
most vrm blow up grandly if the water hit it.
ignore the voltages in the bios.measure directly with a good meter.
measure at the row of caps along cpu.

[eccerr0r]

Re: Bad Mobo VRM - Symptoms?

As hardware monitoring has always been a Linux issue, I suspect windows has the same issues. Usually the software that reads the voltages are getting raw data from the hardware monitoring chip, usually through i2c. However that's the problem - it's raw data. Sometimes it can be assumed that the data is correct as is without preprocessing as the reference design says it should be, but that only can be an assumption, and we know that assumptions can be the source of all fsckups. Board design will vary from manufacturer to manufacturer, model to model, and you can only trust whatever software the manufacturer supplies. Usually the reported BIOS voltages should be much more like what reality is.

Really need the scope/external meter to be sure.

[RJARRRPCGP]

Re: Bad Mobo VRM - Symptoms?

Those unrealistically low voltage readings are typical of HW Monitor and SpeedFan...

[ChaosLegionnaire]

Re: Bad Mobo VRM - Symptoms?

whoa... x99... top end mobo... here's a lesson: always take good care of your stuff esp. high-end stuff and never run electronics when wet. most likely the board suffered water damage which will be nigh on impossible to pinpoint exactly what shorted and failed.

since u've had it for 2 years and mobos typically have 3 years warranty, just rma the mobo for a replacement. have u tested the cpu in another system as well? hope the cpu didnt suffer water/short-circuit damage also. its possible an adjacent pin with a higher voltage shorted another with a lower voltage. this can cause cpu damage.