• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

PSU can't handle the card?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Krogen

Member
Joined
Aug 29, 2006
Location
Underground
Hello,

I recently bought a x1900xtx graphics card in place of an old 850xt. I put a Zalman VF 900 cu on top of it with some IC Diamond Carat 7 paste. I downloaded and ran ATITool to check the temperatures. They were 'ok' (80C after 10 minutes...). I started overclocking and just after a 50MHz overclock (just GPU... reached 700Mhz) my screen went blank and "no video input" message popped up. I quickly restarted my PC... Everything seemed ok. So I tried again. I upped the voltage to 1.450 thinking the stock voltage wasn't enough. Same thing happened though (crash after about a 50MHz overclock). I tried overclocking the memory with same results.

Other things:
I get weird screen 'tears' right after Window loads. Same thing happens if I try to overclock the card (the screen tears in the middle every few seconds?).

Any suggestions? My PSU and other components are listed in my signature. The card works fine when I play oblivion...
 
Krogen said:
Hello,

I recently bought a x1900xtx graphics card in place of an old 850xt. I put a Zalman VF 900 cu on top of it with some IC Diamond Carat 7 paste. I downloaded and ran ATITool to check the temperatures. They were 'ok' (80C after 10 minutes...). I started overclocking and just after a 50MHz overclock (just GPU... reached 700Mhz) my screen went blank and "no video input" message popped up. I quickly restarted my PC... Everything seemed ok. So I tried again. I upped the voltage to 1.450 thinking the stock voltage wasn't enough. Same thing happened though (crash after about a 50MHz overclock). I tried overclocking the memory with same results.

Other things:
I get weird screen 'tears' right after Window loads. Same thing happens if I try to overclock the card (the screen tears in the middle every few seconds?).

Any suggestions? My PSU and other components are listed in my signature. The card works fine when I play oblivion...

Now this one has got me ROTFLMAO. You were at 80C with stock clocks and Vcore so you tried
bumping the Vcore to get more more core OC with GPU cooling that totally sucks and you wonder
why the card dorks.

Viper
 
80C with an aftermarket cooler?

I'd try reseting that and getting some more pressure. Make sure you don't use too much paste as well :)
 
I also thought the temperatures were really high but when I searched what temps other people get on that card I was a bit relieved ;) I used ATITool so it's also unlikely that any other app will get passed the my 80C mark.

Here's something I found:
http://www.tomshardware.com/2006/08...r_roundup_2006/page9.html#cooling_performance
(Same video card, same cooling tested)

^I actually got a little better temps (I know.... There are other factors involved, not just the VGA cooling, but still)^

Thanks for replies.
 
80C at stock speeds isn't "OK". That's hitting the upper limits on standard cooling, much less aftermarket. Definitely reseat the card with a thin layer of paste and crank the fan up. I get nervous when a videocard gets above 55C on aftermarket aircooling- 80C isn't good at all. Hopefully no permanent damage was done to the card :-/


Additionally, if you're using ATI Tool, you're using the latest beta right? The most-recent stable release had trouble with overclocking x1900 and x1950-series cards IIRC.



:sn:: Not a big fan of Zalman stuff, personally, but even with it you shouldn't be seeing temps that high.
 
Not exactly the same card, but with stock cooling with an ambient temperature of 72f, my cards run at 67c stock, 78c under load. With the clocks at the oc in my sig, they will go to 81c under load. (I get the hottest temps I've ever seen folding @ stock speeds with a gpu client. 84c :mad: )

Just for reference.

fyi: I have heard that the non-pro x1950xtx has better stock cooling than the pros, or the 1900 series. ViperJohn could tell you for sure. When he stops rolling on the floor laughing his @33 off, if you ask real nice, he gives GREAT advice. One does not become the Radeon Mod God without knowing a thing or two about these cards. Now that man knows about cooling vc's!
 
Last edited:
*Scratching my head*

I took out the card, reapplied the paste (as best as I could... almost perfectly), turned on my PC, ran ATITool for 10 minutes. 77C, 4 degrees less. Wouldn't go higher.

I'm not saying 80C is cold. On most cards it's hot. But not on this one. As far as I know most people go well above 80C mark on this card (thermal shutdown is 120C (!!)).

I tried OC'ing again, same thing happened (blank screen). I tried to set the MHz on the core and memory myself, everything fine (before I used the find 'max' option).

*confused*
Why does it crash with "Find Max" option?
 
Krogen said:
*Scratching my head*

I took out the card, reapplied the paste (as best as I could... almost perfectly), turned on my PC, ran ATITool for 10 minutes. 77C, 4 degrees less. Wouldn't go higher.

I'm not saying 80C is cold. On most cards it's hot. But not on this one. As far as I know most people go well above 80C mark on this card (thermal shutdown is 120C (!!)).

I tried OC'ing again, same thing happened (blank screen). I tried to set the MHz on the core and memory myself, everything fine (before I used the find 'max' option).

*confused*
Why does it crash with "Find Max" option?

I have a gigabyte 6600GT. The shutdown temp is 145'C!!!

I idle at 38 and load at max of 55(zalman vf700). Upon overclocking things get shakey and crash at 75'C. Don't listen to the shutdown temp, it is NOT an indication of anything other than the temp at which the computer shuts down to avoid getting molten bits of plastic dripping everywhere.
 
The 'find Max' option is notoriously innaccurate. I think it allowed me to run the clocks at 695/1098... which it would idle just fine at. But as soon as it went under load at all, it would puke.

That temp does still seem high to me for aftermarket cooling... (See my post above.) Although I doubt you damaged the card. Did you clean the surfaces with clean coffee filters, and isoproyl alcohol? Hows the case ventilation? I don't think ocing is safe if you have stock temps at 80c!

If I may give some laymen's advice... Never use the find max option again. Once you get the temp issue sorted out, like at least 10c cooler than it is now, slowly increment the clock higher until you start to get errors, then back it down a bit. Then slowly increase the mem until you get errors.

But until you get the temps sorted out, you're asking for a Kentucky Fried VC.
 
Well, I tried again. I got to 690,810 (650,775 stock). My PC locked up and, bam, blank screen. "No video input"

Reboot.

Why did it not give me any artifacts???

I'm done for now. For some reason I'm thinking (though I'm probably wrong) it's the PSU. A PSU calculator (in one of the sticky PSU threads) gives me a total of 480W while my PSU is 450W with not-so-high amperages (close to 30a total?).

So... I'll try to grab a new PSU and see if that helps. Does anyone have any recommendations? I'm looking for something for $60-$70, not much more.

Gotta go on a diet to save some money... :(
 
When I used the card it was on water and temps were very low... idle around 32C, load I think was like 46 maybe? Anyways, I would try to address the cooling first(maybe subpar case airflow, etc) and maybe PSU second. But, the PSU calcs aren't very accurate. So, don't put much stock in them. Can you do a case side blow hole? That would give you near ambient air directly to the card area. I bet that helps a good bit. If you would like to test this: Open your case and put a small fan pointing toward your card area. That will give you some indication of how much you could stand to lose temp wise from doing that mod.

Also, take the fan for the vf900 and plug it into the motherboard so you don't have to worry about turning up the fan speeds with ati tool. That will give you 100% constantly if you have it setup right in the bios.


Hope that helps!! That was a GREAT card for me!
 
Krogen my 1900xt idles at 43c with a Zalman FC-ZV9, after running loops of 3DMark06 it might hit 70-75c depending on ambient temps. You have something seriously wrong with your setup....its not your power supply. To little voltage or current would not cause over heating, the card would just be unstable.
Try keeping the side panel off your computer and seeing what the temps are like.
 
OK. So I replaced my old PSU with a 700 watt OCZ GameXStream, problem not solved.

I forgot to mention my idle graphics card temps - they hover around 42-44C.

ATITool states that my graphics card ambient temperatures are 48-50C (on load). Low 30s on idle (definitely not ambient but not-that-hot either).

Any other suggestions? Thanks everyone.
 
Last edited:
Back