I really need to pick up the pace with my 9 series cards! I am really starting to like the 9 series. The 970, 980 and 980 TI were all very fine cards in their time and can still be great cards for HD gaming. They improve significantly over the 7 series both in terms of power and efficiency.
Common Issues
This is mainly what I have seen, but it may also be that I favour picking up cards with circuit/power related issues.
- VRM shorts – mainly GTX 970/960
- Crashing e.g. running for short periods, crashes on driver load – This seems relatively common for all 9 series
- VCore voltage drops e.g. to 0.2V, likely PWM fault condition detected in VRM
- BGA e.g. core and VRAM – only 980/TI so far
- PCB damage – The EVGA GTX 980 TI seems to be classic for this, very common to see faulty ones for sale.
Less Common Issues
- INA 3221 e.g. Core frequency stuck at 135Mhz – One 970 like this
GTX 960
- MSI GTX 960 2Gb Armor (Crashes under load)
- EVGA GTX 960 2Gb (Fans twitch at intervals, 0.22V on VCore, also had a shorted 4C10N)
- EVGA GTX 960 SC ACX 2.0 (May have a VRM issue, crashing to blue screen after a while)
- EVGA GTX 960 4Gb SC ACX 2.0 (PARTIALLY FIXED – Had a blown h/s vcore MOSFET, burned 0hm resistors across the 12v power planes, waiting for 4C10N)
GTX 970
The 970 is seeming like the more reliable used/faulty buy for me so far, seemingly less likely to have core-related issues. It still games well enough for me too, nice cards! Shorts seem to be the most common problem I have seen so far. Also, I’ve seen no BGA-type faults so far, unlike with the GTX 980 and 980 TI.
GALAX HOF
This is my absolute favourite (just got it, not repaired yet) and one the most impressive card designs I’ve seen so far! Makes me want more HOF cards!
- GALAX GTX 970 HOF (Just got it, crashes under load, possibly vcore VRM related, as VCore drops to 0.2V)
MSI Gaming
I have worked on a few of this type now. 2/3 had VCore VRM shorts and I am finding that the PCB temperature on the back of the VRM is typically 100°C+ when running sustained. I am still trying to work out if this is a trending issue or whether it is just normal for the design.
- MSI GTX 970 GAMING 4Gb Card A (VCore short, short removed, needs root cause analysis)
- MSI GTX 970 GAMING Card B – (WORKING, TESTING, more investigation needed to check thermals)
- MSI GTX 970 GAMING Card C (FIXED, TESTING – 12V VCore VRM short, replaced DRMOS and checking card thermal performance)
ASUS Strix
This type actually seems rather nice so far.
- ASUS STRIX GTX 970 DirectCU II OC (Card A) (FIXED – Not detected due to missing PCI capacitor on lane 1, 5V short caused by MOSFET driver, missing 2/6 phases, faulty VDRV resistor)
- ASUS STRIX GTX 970 DirectCU II OC (Card B) (SCRAP – B GRADE – Suspected faulty VMem PWM, PEX also likely missing as a result. Actually shorted memory/core)
- ASUS STRIX GTX 970 DirectCU II OC (Card C) (Has picture, blank screen shortly after driver load and multiple stats missing in GPU-Z as a 2nd card. MATS passes, try MODS, suspect possible power delivery issue)
- ASUS STRIX GTX 970 DirectCU II OC (Card D) (Crashes on driver load, check with MATS, if OK check VRM function)
Others
- EVGA GTX 970 SC GAMING ACX 2.0 (PARTIALLY FIXED – Missing voltage rails, there was a partial short and fried MOSFET, plus some corrosion. Waiting for 4C10N)
- Zotac GTX 970 (Seller Note “Short. Can see 3.3v PCI pin damaged” – Actually, has a bad VCore VRM short and PCB damage under the removed MOSFET)
- PNY GTX 970 (FIXED, TESTING – listed as stuck at 135Ghz , INA 3221 issue – damage to nearby resistors, failed repair attempt by seller perhaps)
- MSI GeForce GTX 970 4GD5T OC (SCRAP – C GRADE – Has 12V pci-e short, badly damaged PCB due to burns and corrosion)
GTX 980
I have only looked at these two, but the 980 seems a bit harder to get hold of. I will be lucky if I can get a fix out of these two, both appear to have been reflowed!
- GTX 980 GAMING 4G (Has green line artefacts, MATS seems to crash, need to find some way to test the VRAM, but could be a core issue)
- ASUS STRIX GTX 980 DirectCU II OC (Card A) (SCRAP – A GRADE – Artefacts on BIOS screen, failures in MATS on multiple chips. Suspected core faults, memory chips possibly good)
- ASUS STRIX GTX 980 DirectCU II OC (Card B) (Artefacts on BIOS screen, failures in MATS both B channel chips and all bits, damaged VRAM components near channel B chips)
GTX 980 TI
Like the GTX 980, I have found the GTX 980 TI seems to be more the danger zone for repairs! This is mainly due to core related issues, and often savage PCB burns in the case of the EVGA GTX 980 TI. The EVGA 980 TI is a tempting repair buy in this respect, because it seems the VRMs fail before the core gets a chance. However, their repair doesn’t always seem particularly easy e.g. partial repair with missing phase/damage is often a much easier end state. They are very nice cards when they work though.
- Inno3D iChill GTX 980 Ti Black (Overheating, AIO needs fixing or replacing)
- ASUS STRIX GTX 980 Ti DirectCU III OC (WORKING – Lovely card! So far all it seems to have needed was a thorough clean. Initially, any load at all would sky-rocket the GPU temperature and cause thermal overload! It was very dirty. I would also like to change the VRM thermal pad, do further cleaning and check the VRM health with an oscilloscope. Very capable, but hot and juicy, 1440p gaming, GPU temp typically high 70s and hotspot around 90)
- EVGA GTX 980 Ti 6GB 06G-P4-4995-KR (Card A) (Had quite a savage VCore VRM short damaging the PCB, works to windows without stress)
- EVGA GTX 980 Ti SC GAMING ACX 2.0+ (Card B) (SEMI WORKING for desktop use, but will crash to desktop under full load)
- Gigabyte Windforce Extreme GTX 980ti Card A (Has artefacts, coil whine – MATS failure, don’t expect it will be a simple as replacing the faulty chip)
- Gigabyte Windforce Extreme GTX 980ti Card B (this one is seemingly more interesting. Was concerned at slightly lower vcore of 2.9 and very low vmem resistance of only 16 ohms. When powered on there is no voltage to vcore or vmem. Not many minor voltages at all. The pci 12v fuse has 12v on one side and 0.79v on the other, but I would imagine there is more wrong than this and a deeper investigation is required – Fuse replaced, now has artefacts, 2 chips fail in MATS)
- Gigabyte Windforce Extreme GTX 980ti Card C (Has artefacts, coil whine – Vert similar to card A, MATS failure, don’t expect it will be a simple as replacing the faulty chip)