Hiya, I have an HIS 5770, both Everest and GPU-Z are showing the following temps: Idle DispIO: 35C MemIO: 42C Shader: 34C Load (Furmark) DispIO: 62C MemIO: 107C Shader: 59C The MemIO temp under load is too high, what temp should it be? What would cause the MemIO temp to be that high but the DispIO & Shader temps to be okay? Edit: got temps sorted but pulled a capacitor off, see below
Having had a quick look at some pictures it's probably the VRAM on the back of the card that's heating up. You could try some RAMsinks, the Enzotech ones are pretty good. If you want some simple aluminium ones I'm certain I have some spare I can send you. Additionally directing some cooling on the back of the card should help, either a side panel fan or something like a Antec Spotcool. EDIT: Probably a quick way to check is to remove the side panel and direct a desk fan onto the back of the card.
On closer inspection, I think the thermal paste needs redoing. The h/s & fan isn't actually tight on the card, it moves around a little. I have read elsewhere that the MemIO diode is a particular corner of the juniper chip and if it is not seating correctly at that point, it can cause this problem. I am going to re-do the thermal paste tonight and see what happens. What paste is the best to use on a GPU chip? I have some AS5 and some MX-2 - would one of those do?
How long have you had the card? If it's any less than 6 months, I'd be sending it back as a defect. I'd still take my chances and get in touch with HIS after a year, tbh. Don't want to lose your warranty over it. As for thermal pastes, yeah, AS5 would be spot-on.
Card is from the market place but has warranty left if it comes to that. Hopefully the seller would be nice and help with that if need be. I'll use MX-2 paste I think.
Yeah, my bad. Sleep deprivation does that to the boy. AS5 should be spot-on if used REALLY sparingly!
I'd be surprised if a loose heatsink was causing the VRAM to overheat, particularly as there's no cooling on the VRAM at the back of the card. Mind you, you never know.
Like I posted earlier, I don't think it is the VRAM overheating, I think it is a MemIO diode on the GPU chip. I'll reseat it tonight and let you know how it goes.
Okay, the good news is that the reapplication of thermal paste (I used MX-3) has resolved the temp issue, max temp on any of the three sensors is now 62C after 10 minutes running Furmark. Fan is only running at 61% as well but the ambient is only 18C here at the moment. The bad news is that I pulled a capacitor off the card whilst dismantling it - bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger, bugger! However, the card still works and is still running at full load on Furmark as I type this (it's on another PC). This is the cap that has come off (pic from web, not my actual card): Given that it is beside the power connector, I assume that it is to smooth out the power from the PSU? I can source an Aluminium Electrolytic cap from Farnells and I am happy to solder it on. Can anyone help me identify the value of the cap? Written on the cap is "j9 100 16Z". What value does that represent? Is it a 100µF 16v cap? Can anyone confirm?
Now, I really don't want to say it, but... well, you get the idea. Capacitor markings normally show voltage and capacitance on the side, as indicated here.
I know, I know but I couldn't realistically RMA it because of the temp issue (which is resolved) and now I can't because the capacitor falling off is accidental damage, not a fault or failure of the product. Even if I glued the cap on I would have no cause to RMA it as it is still working fine as far as I can tell. If I can identify the capacitor value, I can repair it and it will be fine in my opinion. I got it second hand in the marketplace here for a good price so I think it is worth repairing. So can anyone help me identify the capacitor rating?
Well if it turns out that they're similar to: C 89N 270 16v I can supply a few from a 4870 card I trashed fitting a crap waterblock. I think it's the GPU that fried not the capacitors but not 100% sure.
Cheers for that mate. Unfortunately I do not know enough about capacitors to determine if yours are the same. It doesn't help that the labelling of them is all over the place due to size etc. Still looking for help to determine what rating cap I need...
Bumpty bump... Can anyone help or point me in the right direction to identify the correct rating of this capacitor? I have tried numerous informational websites but it is all very confusing. I have emailed the manufacturer asking, no response yet.
According to here, you were right first time. J9 may just be a code, but always worth checking as many places as possible, I suppose!
Cheers, the confusion for me is the "100" bit. I can't assume it is 100µF as different manufacturers shift the decimal place around. It could also be 1000µF...I just don't know. Edit: I think the cap is a Sacon FZ but I am not sure.
as long as the cap is big enough.. if it's rated at 14v and 100 microfarad- you could stick on a 20 or even 50v, 150 microfarad capacitor without issue if under rated voltage wise it could blow.. probably 12v so a 14 should do it