SPCR

Thank you, a great review of a great product for the silence hungry gaming crowd.

I would also like to draw attention to Asus' 670 GTX offerings which have a just as low noise and efficient cooling solution. For some reason it isn't mentioned in the price discussion where Radeon HD 7970 is mentioned, but the ASUS GeForce GTX 670 Direct CU II TOP edition is almost as fast as the 680 GTX (reference clock) at a much better price!

See eg the TPU review: http://www.techpowerup.com/reviews/ASUS ... ect_Cu_II/

Damn you making it very hard to resist the upgrade, i been more than happy with my Galaxy GTX580, but the GTX680 is wonderful card, i was just a little worried about noise, but with your review, there seems to be no issues..... my wallet is going to cry.

Thanks for the review,

Thanks for the review. From the pics, I understand why the VRM temps get so high at the reduced fan speed. The air has to blow through the gpu heatsink first before it hits the VRM heatsink. Lots of resistance. I have to think some other designs that use thermal tape between the VRM and the backplate of the main heatsink would be better for lower rpm cooling (plus providing some cooling for the RAM). On the other hand, I was surprised the 112C temp only decreased the efficiency by a watt (figuring the GPU VRM is where most of the loss is).

Kudos to Nvidia for their low idle/2D and 2D multimonitor power use on Kepler.

errata: page 3, Estimating DC Power - The Kingwin PSU table has the Seasonic label (but has the correct Kingwin data).

I recently switched from an Asus 580 DirectCUII to this cards little brother GTX670-DirectCUII TOP, and while they use a very similar type cooler, the difference in fan-noise is just staggering. For me the upgrade was worth it because of the decrease in fan-noise alone. Add in GTX680 reference card performance at a lower price and you have yourself a winner. I'm very happy with this new generation cards from Asus.

Great review but one has to wonder if the GTX660TI wouldnt have been a better choice for silentpcreview.com

Just to make sure - is 840 RPM on both fans the lowest the card selected itself, or would it be the lowest you could force manually?

Not to derail the thread, but could some owners of similar Asus cards (670, 660 Ti) share their fan speeds? At least just for idle? These two latter cards have 2-slot coolers so there may be some differences in performance under load, but I guess in idle all these should behave similarly and so should their fans. Come to think of it, the fans on 670 and 660 Ti Asus look a little smaller, could be they're not 100 mm in diameter?

Would also love to see 660 Ti tested, especially Asus and GB Windfart versions. Both look v promising for noise conscious users.

AckeDman wrote:Great review but one has to wonder if the GTX660TI wouldnt have been a better choice for silentpcreview.com

It probably is, being more compact and arguably more efficient (though the difference is not big, the 660 Ti is rather cost-effective) for 1920x1200 and under, but the review makes mention of how the 680 is overkill. It is also a plain fact that SPCR works with the material and resources they have; if you, on the other hand, have an extra $150, I'm sure the 660 Ti review can be arranged.

It's good to see the DirectCU cooling system get an SPCR level review, even though it is only one - the beefiest - iteration, and Asus hasn't seen fit to designate the different versions, just the "generation", which feels more like marketing than anything else.

AckeDman wrote:Great review but one has to wonder if the GTX660TI wouldnt have been a better choice for silentpcreview.com

You realize this is probably the video card on hand from the AVADirect PC that was then put into SPCR's video card reference PC for testing... I call it serendipity - two reviews from one delivered product.

How come GPU heatsinks go directly on the die, whereas CPUs have a IHS ??

Rebellious wrote:How come GPU heatsinks go directly on the die, whereas CPUs have a IHS ??

Because GPU cooler come premounted. CPU coolers have to be user mounted.

CA_Steve wrote:

AckeDman wrote:Great review but one has to wonder if the GTX660TI wouldnt have been a better choice for silentpcreview.com

You realize this is probably the video card on hand from the AVADirect PC that was then put into SPCR's video card reference PC for testing... I call it serendipity - two reviews from one delivered product.

I did realize this. I was just thinking it would have been nice to know how low in SpL(dbA@1m) the GTX660TTI would have been at load. Either way other reviews puts the GTX660TI under the GTX680 so i will probably just buy that version.

Thanks to SPCR i now feel safe buying the DCUII versions of GTX660TI or GTX670.

AckeDman wrote:I did realize this. I was just thinking it would have been nice to know how low in SpL(dbA@1m) the GTX660TTI would have been at load. Either way other reviews puts the GTX660TI under the GTX680 so i will probably just buy that version

Ah. I've been tracking GTX 660 Ti reviews here. Take a look at the HT4U review of the Asus card. You can compare it's performance vs the GTX 670. Not quite the TDP of the 680, but you can see the delta.

CA_Steve wrote:

AckeDman wrote:I did realize this. I was just thinking it would have been nice to know how low in SpL(dbA@1m) the GTX660TTI would have been at load. Either way other reviews puts the GTX660TI under the GTX680 so i will probably just buy that version

Ah. I've been tracking GTX 660 Ti reviews here. Take a look at the HT4U review of the Asus card. You can compare it's performance vs the GTX 670. Not quite the TDP of the 680, but you can see the delta.

Yes actually did see that review in that specific thread and i was amazed. A lot of the times when sites have dbA or SpL of 30 as quiet im very sceptical so it was nice to see that the GTX 660Ti is actually very quiet considering they measure it at about 16-18dbA at gaming or full load. The fact that they have the GTX670 at 30dbA att full load means the accurucy in their equipment must be high.

If techpowerup's measurements are correct, then the 660 and 670 versions are even quieter (they use a dual-slot cooler).

GTX 660ti / GTX 670
http://www.techpowerup.com/reviews/ASUS ... II/27.html
http://www.techpowerup.com/reviews/ASUS ... II/27.html

Idle: 24 dBA
Load: 25 dBA

GTX 680
http://www.techpowerup.com/reviews/ASUS ... II/27.html

Idle: 27 dBA
Load: 35 dBA



Sadly I don't see SPCR getting a review sample of either of those cards.

My Asus GTX660Ti have 2 fans and the lowest they will go is 10% at 1020rpm. However at this point i cant seem to hear them because the vibration from 1 of my hardmounted Scythe GT 800rpm are louder because of vibration, also the 180mm fan as intake is problably louder at 400rpm.

/The swede

My friend also got the 670 model recently (ASUS GTX 670 DirectCU II) and he too says approx. 1040 RPM is the lowest both fans will go, either manually or automatically. Given the similarity between the 660Ti and 670 there's no wonder even the cooler is the same. Perhaps many of us would really prefer 800-900 RPM in idle, but I guess a little over 1000 is good enough too. Especially that the fans seem good quality, smooth. No real need to use an aftermarket cooler on these cards.

I recently upgraded to this card from MSi GTX570 Twin Frozr II and this card by comparison is VERY noisy. There is a some kind of mid-frequency bearing noise that is transmitted directly into my computer case and has made my system considerably more noisy.

I was originally going to buy the 670 DirectCU Overclock but my preorder fell through - it seems the 670 would have been a much better card for silent computing. At this point I think I will need to look into an aftermarket cooler as I can clearly hear my computer from across the room.

drez wrote:I recently upgraded to this card from MSi GTX570 Twin Frozr II and this card by comparison is VERY noisy. There is a some kind of mid-frequency bearing noise that is transmitted directly into my computer case and has made my system considerably more noisy.

I was originally going to buy the 670 DirectCU Overclock but my preorder fell through - it seems the 670 would have been a much better card for silent computing. At this point I think I will need to look into an aftermarket cooler as I can clearly hear my computer from across the room.

I recently installed a second GTX 660 Ti in very quiet test system (15 to 19dB inside). Unfortunately, when I added a second GTX 660 Ti, some new source of noise kicked up. I thought it was the second video card but it turned out to be the motherboard. A faint high pitched EMF tone starts up when running in SLI mode. If I remove a card, the system is silent.

To further test the issue, I installed a single ZOTAC GTX 680 and ASUS GTX 680 DirectCU II which created the same noise. I swapped the card out with a GIGABYTE GTX 670 OC and the EMF tone is less prevalent. That has me wondering if its your card, or the motherboard.