EVGA GTX 295 Backplate Performance Tested - 20C Temp Drop!!

[AuDioFreaK39]

*Final Update*

I ran the test three more times, with and without the backplate. The official temperature difference is 10C. Not quite as high as originally stated, but nevertheless impressive. Apologies for the confusion.

NO BACKPLATE - stock speed - 100% fan - Furmark Xtreme Burning Mode:

78C/68C




WITH BACKPLATE - stock speed - 100% fan - Furmark Xtreme Burning Mode:

68/61C



================================================== ==============

I have some great news for you guys.

So I got my EVGA GTX 295 Plus cards today from Micro Center for $479 each. The backplates also arrived from EVGA in the mail, and I thought I'd give you guys the rundown on what kind of temperature differences you can expect with and without the backplate installed.


Without further adue, pics:



Hmm, wonder what's in the sunlit bag.







Here are some shots you don't normally see. For those who don't own an EVGA GTX 295, the shell is made out of hard rubber:



Inside of the heatsink:



Backplates:









As you can tell, the aluminum really is only 1mm thin as mentioned in the specs:





And now for the grand finale:


Idle Temperature WITHOUT backplate (stock speed) - 48/51C




Load Temperature WITHOUT backplate (stock speed) - 91/93C




Load Temperature WITH backplate (stock speed) - 70/63C



^ A note on the time length of the above GPU stress test:

I initially ran it for 1000 seconds and then stopped it for about 10 seconds as I couldn't beleive my eyes. I then ran it again for 173 seconds, and as I saw the temperatures did not fluctuate, I decided to stop it.


add to the fact that my Thermalright TRUE Copper is idling my Core i7 965 Extreme at 23C and this is a dream come true!




And finally, installed pics.

This is an Apevia X-Discovery case that I've had since February 2006. It's in dire need of replacement. My hardware has "outgrown" it, i need more room.

[breathemetal]

Hmm...I may rethink NOT getting this card. How loud is it?

[AuDioFreaK39]

Hmm...I may rethink NOT getting this card. How loud is it?

I'd say 25% louder than a GTX 260 or GTX 280 on both idle (~42% fan speed) and 100% load.

[breathemetal]

How much louder than a 8800GTX?
And stability? Driver support?

[AuDioFreaK39]

How much louder than a 8800GTX?
And stability? Driver support?

Probably 50% louder than an 8800GTX. Stability is rock solid. Driver support...I'm going to figure out over the next few days.

[Marvin54]

will be compatible with a XFX ?

[Vapor]

That's pretty sweet!

295s are pretty loud though...watercooling is a must for my ears with them I wonder if the backplate would help at all with watercooling anyway...

[AuDioFreaK39]

That's pretty sweet!

295s are pretty loud though...watercooling is a must for my ears with them I wonder if the backplate would help at all with watercooling anyway...

Hmm, that's an interesting idea.

[eligray]

It doesn't surprise me that there's such a drop. I don't think people realize how good of a heatsink the PCB is. When i shut off my WC rad fans across the room, and let core temps increase by 20c, my memory and PWM temps went up ~20c also. This is with 2 fans directly on the card! the GPU block had been sucking heat out of the entire card, and these results have been replicated many times by myself

[largon]

IMO, the correct conclusion would be:
The backplate just is not worth it.
Better put two 120mm fans blowing on the backside and get a larger temp drop. And save some $$.

[Xello]

Load Temperature WITHOUT backplate (stock speed) - 91/93C

That's 16-18c higher than the highest mine have been under load (and i have two so the bottom one is slightly hotter than usual)

No idea why yours is running this hot but seems backplate is not a bad idea for you

In my experience the cooler is excellent at keeping the gpu temps in check, the way it does it though is by venting some of the heat out the rear and some of the heat out the side as you can see the fins in OP's pic and here:



Which should go out the rear side mesh mostly, but always some of the heat is going to be transferred to the case. This is different from ATI's approach - my 4870x2's exhausted all the heat out the rear so there was less put into case, though this gave me idle/load temps of 80-95c, so each has its merits.

[Demo]

Wow 91c, how much temps did nvidia rate the GTX295 for safe 24/7 operation

[zlojack]

Something is weird about these results.

You ran Furmark for a much shorter time with the backplate on. Perhaps the cards just didn't heat up enough?

You should try it with both cards for 500 seconds and then we can see what the results look like.

[Slovnaft]

Wow 91c, how much temps did nvidia rate the GTX295 for safe 24/7 operation

gpus are generally more robust than cpus. they can run up to 120C before being damaged, so heat is more of a performance thing rather than actual danger.

[AuDioFreaK39]

Something is weird about these results.

You ran Furmark for a much shorter time with the backplate on. Perhaps the cards just didn't heat up enough?

You should try it with both cards for 500 seconds and then we can see what the results look like.

Here is my clarification on running the Load w/ backplate test:


I initially ran it for 1000 seconds and then stopped it for about 10 seconds as I couldn't beleive my eyes. I then ran it again for 173 seconds, and as I saw the temperatures did not fluctuate, I decided to stop it.

[SNiiPE_DoGG]

gpus are generally more robust than cpus. they can run up to 120C before being damaged, so heat is more of a performance thing rather than actual danger.

and efficiency.... the card draws much more power when its running at 90C as opposed to 40c

[eligray]

and efficiency.... the card draws much more power when its running at 90C as opposed to 40c

indeed. adding a fan to a decent cooler actually lower power consumption compared to having the fan off.

[AuDioFreaK39]

Alright guys, I know you've been waiting for my follow up on this. Here it is:

I ran the Load w/ backplate test again for 1134 seconds, Furmark Xtreme Burning Test, 100% fan speed, and I still got 68/61C

[HuffPCair]

Thats weird I just dont get how a backplate does that much i would think it would hold in more heat then help get rid of it. but the results i guess are right there. Looks good atleast

[FragZero]

My guess is that the backplate increases the pressure on the GPU and that the same result can be achieved without the backplate by adding some washers !

[HuffPCair]

My guess is that the backplate increases the pressure on the GPU and that the some result can be achieved without the backplate by adding some washers !

ahh yeah i can see that. but the plate does make it look nice if you are a person that cares about that stuff.

[FragZero]

ahh yeah i can see that. but the plate does make it look nice if you are a person that cares about that stuff.

I looks nice, this should have been include in the reference design !

[nullface]

Eh... are you running 2 cards in SLI one with back plate and one without? I would say take on out test it, the take it out an place the second card in the same slot, airflow might be bad around one of the cards compared to the other = giving a incorrekt result.
I just cant belive there would be a 20C difference backplate vs. no backplate.

[LennyRhys]

He has only one GPU installed. I mentioned this on another forum: I think the other GTX 295 should be tested in exactly the same manner.

[tiro_uspsss]

I still dont understand how contact is made with the v.card?

do some testing with & without fan blowing on v.card with & without backplate

edit: my bad, I see the thermal pad, placed on the 'opposite' of the GPU - thats an awesome idea.. I bet TR is gonna make backplates now for heaps of diff cards.. the GT2XX series, the ATi 48XX series etc