Jump to content


Member Since 13 Jan 2011
Offline Last Active May 25 2018 07:31 AM

Topics I've Started

Killed my GPU. Can I recover it?

22 May 2018 - 01:25 AM

So this just happened. First some backstory

R9 fury got scratched pcb shorting the fan-control, making the fans go 100 percent all the time. Bought a gpu pwm to motherboard pwm converter and ran the fans using speedfan, regulated by gpu temp

Insert little brother who plays with headset and doesn't notice when things go wrong (eg artifacts and the like).

Insert Asus game booster

Maybe you've already guessed what happened, the game booster disabled speedfan while my bro was playing pubg and according to him the pc suddenly went to sleep mode, but I recon the card just stopped giving output since the pc was still on when I came to take a look (was downstairs studying)

What are the chances I can recover the card? I read that I could use a heatgun / oven to melt the solder, possibly re-seating the chip on the pcb

Extra info:
The heat exhaust at the top of the case was a lot higher than usual when I shut the pc down after gpu failure
Card did not show aging symptoms prior to failure
According to him he only played one and a half rounds, it should be noted that pubg keeps the gpu pinned at 100 percent even in the menu
Led for bios toggle on card works
Pc starts with the card installed but gives no display output
Power supply is a Corsair 650 watt 80 plus gold, pc works fine on the integrated graphics
Device manager detects an unknown pci-e device but can't detect what drivers it uses nor any info about it

Any input is welcome, with the gpu prices I can't afford to buy a new one right now. Thank you

I have no Idea what I'm doing

18 November 2017 - 04:52 AM

Soo I fucked up. Badly.


So it was building a ryzen system in a corsair carbide air 240, and tried putting a r9 fury nitro from sapphire in there. It worked but was a very tight fit, and the window panel wouldn't

close - planned on buying low profile pci-e connectors. Thing is, I never got a signal on my monitor even though I knew the card worked.


Tried using different graphic cards, but because of the space limitation I had to unscrew the motherboard to remove/add the card. It was quite frustrating, because everything worked when the mobo was out of the case.

At some point I didn't give a shit anymore and tried getting the card out without removing the mobo - aaand as you may guess this is where things went south.


The heatsink of motherboard scraped the pcb of the card, short circuiting 2 pcb lanes. Since then the fans go full on blast whenever I give power to the machine (had another system available, same story). Only one displayport

gives a boot screen, but when the pc is fully booted all of them work. Tried manually changing the fan settings in afterburner, it changes nothing. What I did notice is that the system shows the fans being off (fan stop), so I figure

the commands of the card don't reach it's own fan controller.


So now to my question - is it possible to connect the fans of the Graphic card directly to the motherboard? I might buy a fan controller for the dvd bay with temperature sensors, if the voltage/connectors are compatible.

Other "solution" would be to water-cool the card.



Oh and you know why the system didn't work? The screw holding the hdd bay in place was creating a short-circuit on the motherboard -.-)

Vega question

06 August 2017 - 04:12 AM

This might be a bit soon, but I takes time to sell used equipment for a decent price.


I sold my rx 480 during the mining craze, and now bought a 770 for 60 eur to use in the meantime.

My monitor is a ultrawide 1080p freesync monitor, and I'm planning to upgrade to ultrawide wqhd.


Vega should be able to handle this res quite easily, but I'm concerned about the miners. Should I get a vega card as quickly as possible or wait for aftermarket cards? Time isn't really an issue, but if the prices skyrocket again...


Seeing that I'm going to upgrade my monitor anyway, I can jump to the green team. A 100hz uwqhd g-sync displqy costs 200 eur more, but might be more rewarding when volta launches. Same can be said for amd, as they will probably try to use infinity fabric on their gpu's.


Main concerns are Heat/Noise, overall cost and power draw, in that order. I require it to fit in a mini itx case, and be supplied by a 560 watt psu


What are your recommendations? I'm leaning towards buying a vega 64 and upgrading the screen later when they are a bit more affordable (maybe getting a ultrawide 4k). But with miners and the simple fact that amd will only get further behind on nvidia, it may be best to invest in the surest future


?Edit: The site registers enter as a question marks, with pale moon as browser

Patriot Viper 1066/single ch only or crash

18 July 2017 - 10:06 AM

After having a fair bit of trouble with motherboards, I now have some issues with ram.

i got a triple pair of patriot viper, cl7, 1333mhz for 20 euro's, paired with an 2500k and a asrock extreme 4 gen3. 


One stick works fine, two only boot up when I put them in A1 and A2 slots, no other config is stable. i can't get

the third stick to work, but 4 gigs is enough, at least for me. 


Problem is, as soon as I change the clock to anything higher than 1066, the pc either boots up and shuts down,

or crashes when loggin in windows. XMP just crashes the pc immediately after post. 



Now with both stick at 1066 the system is stable, but according to cpu-z they operate in singlechannel mode.


Is there a way to get them to work in dual channel and or make them work at their advertised speed? It probabily 

is a compatibility issue, the mobo is up to date but that didn't do anything.




I'm going to give up building after this, too much of a headache :/

No post, asus P67deluxge, q-code 14-19-20

30 June 2017 - 02:36 AM

Sooo I messed up.


I posted earlier about this board, and how I could only use the bottom pci-e slot.

My solution was to switch motherboards with the one of my brother's pc, 

as his case doesn't hinder the bottom slot to be used.


The board worked fine before with a 2600k, and I updated to the most recent bios

After I inserted a 2500k, I got no post.


So when I boot up, without even a graphic card nor display connected, one piece of 

ram (doesn't matter which slot), no sata drives, it goes in a loop, from 14 to 19 to 20.


Sometimes I get 00, but one of the ram sticks I used might be bad. When I press clear cmos

button on the back of the board, I get F3 (recovery image found).

Tried ram that worked -before the switch. 

Before swapping the parts It gets stuck on 20, this happened often.


Without ram I get error 14, static.


Things I tried: clear cmos, taking the board out and looking for shorts, reseating the cpu, there are

no bend pins on the mobo, trying different kinds of ram on all slots, putting the graphic card in all 3 slots. 

Power supply works fine.


Only variable left is the cpu, but I'm not really inclined to open them up again, I have only so much thermal paste.

Do you think I botched the board? Or maybe microcode error from the cpu? The cpu worked before the switch, 

and I took both out at the same time and put them directly in the motherboards.




At this point I'm seriously considering buying a new mobo and ditching this horrible motherboard