Topic: 5770's overheating (Read 1704 times)
Im gonna say it one more time: disable voltage monitoring in Afterburner and make you are not running any other monitoring app like GPU-Z, Everest or whatever. Sudden temperature spikes like that hint you are accidentally overvolting, just running afterburner can cause overvolting. It did on my 5850, randomly my core voltage would spike to 1.65v, its a known issue, at least for Cypress based cards but it sounds like it could be an issue with 5770s too.
thanks for the advice on the thermal paste, I guess it can't hurt to try it your way. I've redone almost every card I own's thermal paste with artic silver so highly doubt that's the problem but will try anyways, as it had the same problem before i messed with it as well.
In case I still don't any luck.... anyone have a 5770 in a dedicated miner, that can give me some settings/software versions I can try copying and hopefully get it working
In case I still don't any luck.... anyone have a 5770 in a dedicated miner, that can give me some settings/software versions I can try copying and hopefully get it working
The very fast heat rise suggests that it's not an airflow problem - the heatsink should be able to sink heat for a good 20-60 seconds with zero fan before you reach 90 degrees. So I think your problem is getting the heat TO the heatsink.
Did you do the thermal paste right? The right way: clean off everything with alcohol; put a single dot of paste the size of a BB (or less!) right on the middle of the chip; squish the heatsink straight on.
Don't spread the paste around - this will guarantee air pockets which ruin the cooling. This is the #1 noob mistake. Don't use any more than a BB size - thermal paste isn't actually a good heat conductor, it's just better than the rough surface of the chip and heat sink, so you want the paste as thin as possible. Don't move the heatsink around: you don't need to smear it over the whole chip. 90% of the heat is right in the center.
Did you do the thermal paste right? The right way: clean off everything with alcohol; put a single dot of paste the size of a BB (or less!) right on the middle of the chip; squish the heatsink straight on.
Don't spread the paste around - this will guarantee air pockets which ruin the cooling. This is the #1 noob mistake. Don't use any more than a BB size - thermal paste isn't actually a good heat conductor, it's just better than the rough surface of the chip and heat sink, so you want the paste as thin as possible. Don't move the heatsink around: you don't need to smear it over the whole chip. 90% of the heat is right in the center.
this is totally accurate good information. I've seen so many people cake that stuff on and use half a tube
If your temperatuers were ok before, and that was after you put on the heatsinks yourself but now they're not i'm thinking you may have damaged the vrm's it's worth removing the cards for a physical inspection looking for discoloration on the pcb
Or he is now monitoring VRM temps with GPU-Z and running some other app, and encountering what I did, the GPU voltage shooting skyhigh randomly because of those apps (or buggy bios or whatever the root cause is that makes this happen). See my post above.
The only reason I even considered it not being a dead card, was that it ran flawlessly under ubuntu for weeks, and within minutes of using it in windows I was cooking my card. Thats the only reason I even googled on it (and found the cause) otherwise I have bet my entire bitcoin wallet on it being a damaged VRM.
If your temperatuers were ok before, and that was after you put on the heatsinks yourself but now they're not i'm thinking you may have damaged the vrm's it's worth removing the cards for a physical inspection looking for discoloration on the pcb
Better in "Hardware" forum
Are you running an app to monitor voltages, like GPU-Z or everest? This can cause the voltage to skyrocket (to 1.65v on my 5850!). At least on 58x0 cards, but maybe 57x0's too. Close all monitoring apps, except Catalyst, and see if the problem persists. More info here:
http://www.battle-fields.com/commscentre/showthread.php?23435-Beware-of-using-Everest-GPU-Z-Afterburner-with-a-radeon-58xx-!
http://www.battle-fields.com/commscentre/showthread.php?23435-Beware-of-using-Everest-GPU-Z-Afterburner-with-a-radeon-58xx-!
The very fast heat rise suggests that it's not an airflow problem - the heatsink should be able to sink heat for a good 20-60 seconds with zero fan before you reach 90 degrees. So I think your problem is getting the heat TO the heatsink.
Did you do the thermal paste right? The right way: clean off everything with alcohol; put a single dot of paste the size of a BB (or less!) right on the middle of the chip; squish the heatsink straight on.
Don't spread the paste around - this will guarantee air pockets which ruin the cooling. This is the #1 noob mistake. Don't use any more than a BB size - thermal paste isn't actually a good heat conductor, it's just better than the rough surface of the chip and heat sink, so you want the paste as thin as possible. Don't move the heatsink around: you don't need to smear it over the whole chip. 90% of the heat is right in the center.
Did you do the thermal paste right? The right way: clean off everything with alcohol; put a single dot of paste the size of a BB (or less!) right on the middle of the chip; squish the heatsink straight on.
Don't spread the paste around - this will guarantee air pockets which ruin the cooling. This is the #1 noob mistake. Don't use any more than a BB size - thermal paste isn't actually a good heat conductor, it's just better than the rough surface of the chip and heat sink, so you want the paste as thin as possible. Don't move the heatsink around: you don't need to smear it over the whole chip. 90% of the heat is right in the center.
I'm using afterburner to monitor temps. Both went to 90 degrees within seconds of turning them on for the first time and yes I have extra fan's even but what could cause this?
what are you using to monitor voltage and temperature, it's possible if youre cooling was insuficent you damaged the vrm's on the cards as well
windows 7, even on stock speeds, caseless, I'm so confused I just ran one of them 36hrs straight without a problem overclocked to 1000mhz now I restart and I can't even run it stock. the only thing I can think of is it somehow being overvolted by some bug... cards are doing the same thing on both my systems driver it looks like is 11.8
clock settings? os? driver version? inside a case? specs of the rest of the system?
I bought two 5770's a sapphire and a powercolor. Both of them almost immediately overheat (less than a minute) when I start mining. I've redone thermal paste. fans at 90%... The first time I redid the paste on the one it was in the mid 70's until I reboot than back to 90 within a minute
Jump to: