Palit GTX 460 Sonic (Loud? Hot? Rubbish?)

RiGoRmOrTiS · 13 Sep 2010 at 19:05

hi all; I'm at a loss; my patience is wearing thin and I honestly wish I'd never even cosidered upgrading my PC. But maybe someone can offer me some advice.

I recently ordered 2 GTX-460's from overclockers. It was the OcUK Value 1GB cards. When I first received them they were 2 Sparkle Cards. Both ran very well; were exhaust coolers and could maintain a good temperature of 78c-80c under 100% load for 13 hours straight in FurMark. They were not loud and fan speeds barely (if ever) hit 60%.

Unfortunately one of them was faulty and I had to return it for a replacement. This was done quickly; however I received a Palit Sonic GTX 460 as the replacement (appears brand new and sealed). At first this didn't concern me until I opened the box and I noticed it has a cooler that vents internally. I was of course disappointed and confused; but this didn't prepare me for what was to come next..

This thing runs 10c higher idle than my sparkle card and when under load this thing needs to throw the fans into take off speeds (80-90%) just to maintain a temperature of 80-83c; all the while the sparkle is comfortably sitting at 77c with the fans only going 59%. at these speeds the noise is totally unbearable and its making playing any game or sitting near the PC during Furmark stress tests totally maddening. If I'm being honest this card looks, sounds and feels like a dog. am I missing something here or have I been handed a cheap dud as a replacement.

When I ordered the OcUK Value card it showed a picture of a card with an exhaust cooler and didn’t say anything; anywhere that the card may come with a different kind of cooler. Indeed the first cards I received did have the correct kind of cooler as expected. What can I do guys?

varkanoid · 13 Sep 2010 at 19:10

I think you may have been sent the wrong card. The value ones have been usually Inno3D or Sparkle. The Sonic is a listed card on their website. They are reportedly one of the loudest under load too. With the value ones you usually get what you get. The one OC have bought in bulk cheap can be different brands.

I imagine they were either out of stock or low on stock and sent you an alternative arguably slightly better than the "own brand" make.

Why dont you just ring them and explain the situation they will probably swap it you anyway they don't bite !

mmj_uk · 13 Sep 2010 at 19:11

Return it via Distance Selling Act if it's within 7 working days and you aren't happy, then perhaps order a Sparkle separately from elsewhere if OCUK can't specifically sell you one.

vorticalspace · 13 Sep 2010 at 19:14

RiGoRmOrTiS said:
could maintain a good temperature of 78c-80c under 100% load for 13 hours straight in FurMark.

Yep, stop running furmark

Why would you want to run furmark for 1 Hr, never mind 13 :confused:

By all means, run a few loops of Vantage etc and test with your favourite games for stability, but IMO, 13 Hrs of furmark is bonkers.

If your games don't crash, whats the problem. :confused:

Regards

RiGoRmOrTiS · 13 Sep 2010 at 19:15

varkanoid said:
I think you may have been sent the wrong card. The value ones have been usually Inno3D or Sparkle. The Sonic is a listed card on their website. With the value ones you usually get what you get. The one OC have bought in bulk cheap can be different brands.

I imagine they were either out of stock or low on stock and sent you an alternative arguably slightly better than the "own brand" make.

Why dont you just ring them and explain the situation they will probably swap it you anyway they don't bite !

Hi, I've left a reply on my original RMA thread on the forum. No reply yet but it was only an hour ago so not really given them a chance.

They've been good so far and I'm certainly not bad mouthing them; I've been buying stuff from OC's for years. I'm just hoping they won't wash their hands of it and say "working as intended".

Even with the side of the case off this thing goes into rotation speeds that'll cause the fan to travel through time. So it certainly isn't an airflow problem; its simply what the card is like under load.

RiGoRmOrTiS · 13 Sep 2010 at 19:17

vorticalspace said:
Yep, stop running furmark
Why would you want to run furmark for 1 Hr, never mind 13

By all means, run a few loops of Vantage etc and test with your favourite games for stability, but IMO, 13 Hrs of furmark is bonkers.

If your games don't crash, whats the problem.

Regards

I was originally doing that when I started having problems and thats how I worked out which of the cards was faulty.

This time I'm trying games and the Palit card is just too damn loud. I then ran FurMark & afterburner to look at temperatures and fan speeds etc..

RiGoRmOrTiS · 13 Sep 2010 at 19:17

mmj_uk said:
Return it via Distance Selling Act if it's within 7 working days and you aren't happy, then perhaps order a Sparkle separately from elsewhere if OCUK can't specifically sell you one.

I think thats only if it hasn't been opened.

mmj_uk · 13 Sep 2010 at 19:18

It's no more bonkers than running CPU stress tests for 13hrs like most overclockers do, the main difference is GPU manufacturers push the boundaries a lot more and as a result too many duff cards are slipping through the net and into retail.

mmj_uk · 13 Sep 2010 at 19:19

RiGoRmOrTiS said:
I think thats only if it hasn't been opened.

Nope it's unconditional as long as he has taken good care of it and returns it as it was sent with all the bits and bobs.

BigBruiserAl · 13 Sep 2010 at 19:19

rig has a bit of a problem with stress testing, see his other thread, he loves to nuke his hardware

vorticalspace · 13 Sep 2010 at 19:21

As stated, mis-representation of goods.
You bought on the strength of the displayed cooler, your replacement is of a different design -return, end of!!

You are obviously clued up enough to know the difference, therefore it is not unreasonable to assume that this knowledge was a factor during your pre-purchase deliberations

RiGoRmOrTiS · 13 Sep 2010 at 19:22

BigBruiserAl said:
rig has a bit of a problem with stress testing, see his other thread, he loves to nuke his hardware

lol; don't muddy the water; I've not personally stressed anything to death.

vorticalspace · 13 Sep 2010 at 19:30

RiGoRmOrTiS said:
lol; don't muddy the water; I've not personally stressed anything to death.

Hi
I didn't mean anything personal, but after reading a recent rant by Drunkenmaster on these forums about Furmark, Kombuster et al, I have to say that I agree with his comments.

These stress testing programs serve no purpose other than to heat your room in winter!!

maybe a quick burst if you must, but in my experience, if its going to throw an error/BSOD, it happens pretty quickly or not at all.

What if you get no errors after 12 Hrs 59 Mins and then suddenly......
In reality what does this prove, what have you learnt?
Would you ever stress your card that much in a life time of gaming :confused:

Regards

JTrickle · 13 Sep 2010 at 19:40

RiGoRmOrTiS said:
I think thats only if it hasn't been opened.

The idea behind the law is to give people the chance to view goods in the same way you can in the showroom. Should people include putting it in their PC and testing its max overclocks?* No, but phone/email support staff will never ask direct questions.

In this case its a warranty replacement. Long distance change of mind doesnt apply. The usual reject repair/replacement on what basis factors are whats going on here.

*point aimed at other topics

drunkenmaster · 13 Sep 2010 at 19:41

mmj_uk said:
It's no more bonkers than running CPU stress tests for 13hrs like most overclockers do, the main difference is GPU manufacturers push the boundaries a lot more and as a result too many duff cards are slipping through the net and into retail.

I post this in every thread, a CPU is VERY easy to load in ANY situation, cpu coolers, mobo's, cpu's and everything involved are designed to cope with their max output WITH ease.

Gpu's are incredibly hard to fully load EVERY part of the gpu at any one time, its completely unnatural, no benchmarks, no games, no cuda apps, no opencl stuff, NOTHING loads it fully except Furmark, the described by many reviewers when Furmark would crash 4870 cards, as a power virus.

The program is designed to do nothing more than fully load a gpu to a level NO other software can, gpu's are designed to handle the load they've been designed to handle, they have no been designed to run 24/7 in Furmark.

Ok, lets complain, Furmark is a program, it does, well nothing, its boring, its ugly, its pointless but its there, so they make cards that can run it without dying/crashing, but then someone finds a way to bug out cards further, and so they increase the cost to make something run that program 24/7, yay, or, we could all just not run furmark.

Furmark is 100% NOT required to be run to check if your gpu is faulty, if its faulty, basically any game will show it.

If its only faulty in Furmark after 13 hours(he didn't say that, I'm using an example) you could send it back as faulty even though it worked 100% required in EVERYTHING else, for 24 hours a day 365 days a week, for a decade.

ON the same level I could take my pc to the Sahara, get a really really long extension cable and try running my x980 under full load, it would crash due to heat instability, would it be fine to RMA that?

Sensible use, common sense, I've never once seen Furmark give an accurate stability indication for overclocking a GPU and its completely hopeless in crossfire/sli(most of the time it rarely loads the second gpu properly).

Either way, to the OP, unless you have a faulty second card it really shouldn't be running at those temps, then again the temps on the first pair seem high, what case have you got, and case cooling, remember with an ultra quiet psu, you'll probably have less exhausting power than with a standard psu, depends on the case and other fans though.

If you have a pair of hot cards, and a hot case, then run furmark, you're asking for trouble.

AS for those suggesting DSR, its a returned/replaced product.

I had a quick check of your other threads, which seem to centre around, overheating, and then using a pointlessly excessive load on the gpu's for days on end in total is just asking for problems and VRM's to die.

Let us know what case you have, what cooling, and maybe a pic of the case.

EDIT:- with regards to sending it back as not advertised, OCUK made a balls up, in general when you show a picture with an OEM(or own brand) product like that, you usually put a little line in, picture not representative of card you will receive/for illustration purposes only/ etc etc etc.

As they haven't done that, you can certainly argue you reasonably expected to be provided with the card shown in the picture and as such, want to return it. Its also worth pointing out though that the standard cooler on a 460gtx isn't exhausting only, if you put the fan in the middle with heatsink on both sides of the fan, the air doesn't magically go in, go through the fins towards the power plugs and end up going out the back of the case, it dumps a lot of its air all around the sides of the heatsink and both ends, its fair from a proper solid/sealed and exhausting design.

RiGoRmOrTiS · 13 Sep 2010 at 19:45

vorticalspace said:
Hi
I didn't mean anything personal, but after reading a recent rant by Drunkenmaster on these forums about Furmark, Kombuster et al, I have to say that I agree with his comments.

These stress testing programs serve no purpose other than to heat your room in winter!!

maybe a quick burst if you must, but in my experience, if its going to throw an error/BSOD, it happens pretty quickly or not at all.

What if you get no errors after 12 Hrs 59 Mins and then suddenly......
In reality what does this prove, what have you learnt?
Would you ever stress your card that much in a life time of gaming

Regards

well the final 13 hour furmark may have been overkill but it was the final test to make sure my other card was ok

and it was. Thus confirming it wasn't system instability caused by other components.

Normally my max test time per GPU (individually) is 6 hours since thats long enough to consider the cards stable (imho).

the faulty card would lock anywhere from 8 mins to just over 3 hours... it didn't always happen straight away as you said; so not to sound personal either; but it sounds like you don't have much personal experience in testing GPU's

of course that faulty card has now been replaced with this Palit "Sonic Hoover".

Just tried playing Final Fantasy XIV and Crysis on it; Vroom Vroom! I'm almost certain its changing local weather patterns.

Joking aside though; the noise is just unbearable.

vorticalspace · 13 Sep 2010 at 20:04

Touche

Listen, if your not happy with the card and it isn't as represented, therefore not what you would have chosen to buy given the choice, speak/email Gibbo and RMA it.

I am sure he will be understanding of your situation.

As they say, A happy customer is a spending customer

RiGoRmOrTiS · 13 Sep 2010 at 20:50

vorticalspace said:
Touche

Listen, if your not happy with the card and it isn't as represented, therefore not what you would have chosen to buy given the choice, speak/email Gibbo and RMA it.

I am sure he will be understanding of your situation.

As they say, A happy customer is a spending customer

yeah I've left a message on the RMA message board

so hopefully everything should be ok.

MooMoo444 · 13 Sep 2010 at 21:11

I dont think they'd swap it just because its a few degrees warmer than a reference design, but doesnt hurt to try I guess

Good luck with it

RiGoRmOrTiS · 13 Sep 2010 at 21:16

MooMoo444 said:
I dont think they'd swap it just because its a few degrees warmer than a reference design, but doesnt hurt to try I guess

Good luck with it

i take it you skipped over most of what has been said...

1) it internally vents heat (not as advertised)
2) it reaches unacceptable DB noise levels
3) it runs far hotter and is turning my case into a cooker.

I can live with all of them; except for the noise. but thanks for your input all the same..