Gigabyte Aorus 1080ti - card not initialising

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

Examining a working 1080FE, drain is 0.9v until the GPU initialises and then pulls to 0v

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

So I created a new problem probing GPIO pins for voltages. Tried to check pin 39 on uP9511 (VID) which is connected to GPIO0_NVVDD_PWMVID on the core.

Card then stopped outputting any voltages (5v,1.8v etc). Checked the MP1475 buck convertor that generates 5v and its shorted VIN (12v) to gnd.

Not sure why it suddenly failed (unless i shorted some pins with the probe on uP9511, but seems unlikely as they are seem unrelated in that area of the chip (pgood, vid and refadj).

Removed the chip and its shorted (had removed a couple of caps to test between VIN and GND as i thought they were causing it, but no, short was still present)

Now have to wait for some replacement chips from aliexpress to arrive at the end of the month (bought 10)

[Bozog]

Re: Gigabyte Aorus 1080ti - card not initialising

Damn, just when you were getting somewhere, hate waiting for supplies from China, lol

[Traptor]

Re: Gigabyte Aorus 1080ti - card not initialising

Examining a working 1080FE, drain is 0.9v until the GPU initialises and then pulls to 0v

So, the gpio10_fbvref_sel is not needed for working, only to the start?
How did you check this? With oscilloscope?

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

I checked with multimeter, and watched the voltage drop to 0v on the mosfet as the GPIO pin went high. I haven't checked any further to see if this changes with the card in use.

I might hook it up the the oscillopscope and monitor if it changes whilst running MATS on the 1080FE.

It its a constant 0v once initialised then pulling the pin up manually to 1.8v might help get the 1080ti card going again, not sure if the fault is with all GPIO lines or just this one. Won't be able to test that until my delivery from aliexpress arrives.

Will report back.

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

Ok. hooked up the scope to pin 3 of the mosfet (drain)on the 1080FE.

Its high for a second or two as PC switches on, then pulls low. Booted up and ran MATS for a minute or so, pin remains low all the time.

This is looking hopeful that it can be manually tricked by hooking 1.8v to the gate pin (1)

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

Given up waiting for aliexpress (delivery date now showing may 5th!).

Have ordered 2 chips from UK for £16 with delivery, will hopefully be here tomorrow/thursday.

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

Out of interest, here is the scope output for the GPIO pin on the working card (i just used auto ranging on the scope).

It's not a PWM signal as far as I can see as waveform does not change, and neither does the voltage on the MOSFET.

[Bozog]

Re: Gigabyte Aorus 1080ti - card not initialising

Looks a bit dodgy, is your scope isolated from the mains?

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

It's usb powered (using front usb connector on pc). Shows freq of 316khz. Tested the crystal on the card with this set up and it reported a pure 27mhz sine wave (no interference).

[Bozog]

Re: Gigabyte Aorus 1080ti - card not initialising

Strange yeah should be OK if scope has the same ground reference. Still looks like interference, since surely this is a digital signal.

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

Try this one (used USB charger instead of PC) - reported voltage is way better too

[Bozog]

Re: Gigabyte Aorus 1080ti - card not initialising

Yeah that looks more like it, still hasn't helped much with finding out what the problem is tho, hmm...

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

Once i get the replacement chip and get the voltages back on the broken card I can continue testing on that

[Bozog]

Re: Gigabyte Aorus 1080ti - card not initialising

Well good luck with it, hope you find out what's wrong(and it's fixable)

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

I am having a hell of a time finding this 5v short to gnd.

Chips arrived, soldered it in - tested for resistance - 1.1ohms to gnd on pin 3 (SW).

Have removed :

U27
C265
C248
C249
C251
Lifted output leg of L15 choke.

Short is to the right of the choke (as after lifting leg there is no short on u27 or c265 now), so unless it's U6 that's shorted then I am running out of ideas!

R327 is measuring 10.7k

[Bozog]

Re: Gigabyte Aorus 1080ti - card not initialising

Probably the uP9511 unfortunately You can try voltage injection and see what warms up.

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

Yes, my next idea was voltage injection as removing and resoldering parts on this card is a pain

[ss45]

Re: Gigabyte Aorus 1080ti - card not initialising

Maybe sign of GPU failing or bad GPU.
I have had identical 970 in the past. One was cool while idle, the other was 50 c while idle.
Both were repasted

While gaming the bad one initially would play just fine, after a while it would crash, the more I used it for gaming each time the earlier the crash would happen during subsequent game launch. From there it was only going down.

[benwaterson]

Re: Gigabyte Aorus 1080ti - card not initialising

Probably the uP9511 unfortunately You can try voltage injection and see what warms up.

Yup, it's UP9511 and possibly 2 of the DRMOS chips also.

Annoyingly it heats up quicker in the bottom left area where I was trying to probe for GPIO voltages.

Tempted to give up at this point...