R9 390 is dying... D=

Hardware GPU
#1

Hello

After 2 and a half years of good service my CG is nearly dying. With blackscreen everytime I am more then 5 min on my desktop or gaming it freezes with the sound and everything that’s comes with those crashes. I tried litterally EVERYTHING to make it work again (RMA but it actually has no guarentee anymore, VBIOS 2, Power limit from +10 until +50, augmenting the voltage of my ram) the only thing I didn’t tried is to cook it with my oven. But I think that it will be just useless. Of course I searched multiple threads on AMD and other forums to search a solution but there are nearly none (or temporarly solutions but not for long). Has anybody here any suggestions?

Thanks for reading

(p.s. if there is no solution I would like a shop or so where I can send my OOS r9 390. I live in Belgium so international shipping would be nice then or another thing you could do to get some money out of it.)

#2

What are the temps saying in between these crashes?

Are you sure its not just your video drivers? Have you tried the clean uninstall from the Radeon software?

How are you sure its just your GPU? What are the logs in event viewer saying?

There are still things we can try.

#3

ok so the temps are between 35-45 degrees celcius, I clean installed 5 times my drivers and I had the latest one. I just took the GPU out of my rig and now everything works fine 0 crashes since yesterday. I thought at first that my RAM was faulty but it seems I was wrong on that part. EV as I see it just says showes some crashes where the source of the crash is Kernel-Power.

thanks for your help btw.

#4

No problem, glad I can help :slight_smile:

Yeah kernel-power error means that something is not getting enough juice and the kernel sends a SIGTERM and halts the system.

So now, we need to figure out what is doing that; while highly likely to be the graphics card, but not with 100% certainty at this point.

For instance, if your GPU takes a good bit of juice away and something else needs that power, then the GPU is causing the issue, but only because its using the most power in a system that may need more.

So, if you could give us a complete list of the components in your rig, as well as any over clock settings, that will help.

#5

I5 4690k at stock
16 GB of hyperx ram stock 1600MHZ
Z97x-gaming 3 motherboard from gigabyte
stock cooler
Lepa B650 MAXBRON bronze
Phanteks elcips p400s
Crucial mx300 275gb ssd
1TO of storage from Seagate barracuda
R9 390 stock settings no OC (industry OC without backplate)

#6

Yea pushing the PSU to the Limit.
the 12v rail on that is 53A or 636W shared with both CPU and GPU 400 - 450W Power Draw from a 390 isn’t uncommon
@Fouquin do you agree?

Don’t confuse TDP with power Draw.

#7

You mean that that limit occures after 2 and a half years of using intensivelly the gpu? Did everybody tell me that 650w for a R9 390 was enough to power the whole thing? I had like 100 of room for more things and right now what you are saying is that I had to by a 750w powersupply. Even sapphire themselves when i send my full specs didn’t talk about my PSU.

#8

Imo my PSU can go until 636w of power draw on the 12v rail btw.

#9

Power Mosfets do Degrade overtime, Especially if you are running them near there limit and also depends on how hot they are getting.

My 750 Watt packed up with an OCed i7 950 and 390x it ran for about 3 months after i got the GPU and then just refused to work one day.

Meh. Could be a number of factors.