VGA Cooler Roundup: A Thermalright, two Zalmans, and an Arct
-
- SPCR Reviewer
- Posts: 1850
- Joined: Sun Sep 21, 2003 11:23 am
- Location: Vancouver, BC, Canada
VGA Cooler Roundup: A Thermalright, two Zalmans, and an Arct
VGA Cooler Roundup: A Thermalright, two Zalmans, and an Arctic Cooling.
Sometimes, you DON'T get what you pay for...
Sometimes, you DON'T get what you pay for...
-
- Moderator
- Posts: 7681
- Joined: Mon Dec 09, 2002 7:11 pm
- Location: Maynard, MA, Eaarth
- Contact:
-
- Posts: 1386
- Joined: Fri Jun 18, 2004 6:53 pm
-
- Posts: 54
- Joined: Sun Feb 13, 2005 5:09 pm
- Location: Vancouver
- Contact:
I can vouch for the VF900. With some slight modding, you can also attach the VF900 fan onto your video card's fan connector (rather than motherboard) and take advantage of any thermally controlled features your card may have.
I did this with my X800XL AGP, and having the fan thermally controlled via ATI Tool.
I did this with my X800XL AGP, and having the fan thermally controlled via ATI Tool.
-
- Posts: 198
- Joined: Thu May 19, 2005 12:48 pm
- Location: Shi-Khan: Vulcan or MosEisley Tattonnie
-
- Posts: 274
- Joined: Sun Aug 14, 2005 10:12 pm
- Location: Sydney, Australia
I recently had the chance to put both a NV5.3 and an Accelero X1 on my 6800 GT (actually an AOpen Aeolus too ). Being quite a bit of a newbie, I still found them very easy to put on (even if the instructions weren't very helpful).
One thing I wasn't sure of was how to affix the aluminium backing plate for my non-existant rear RAM chips. How did you stick it on? (I noticed you didn't use screws).
Even the NV5.3 was reasonably quiet, but the X1 was difficult to hear
I haven't actually put the card in my case yet so I can't comment on that, but if you thought the Silencer series were big, wait until you see the Accelero! It has a new-fandangled FDB fan too
One thing I wasn't sure of was how to affix the aluminium backing plate for my non-existant rear RAM chips. How did you stick it on? (I noticed you didn't use screws).
Even the NV5.3 was reasonably quiet, but the X1 was difficult to hear
I haven't actually put the card in my case yet so I can't comment on that, but if you thought the Silencer series were big, wait until you see the Accelero! It has a new-fandangled FDB fan too
At the bottom of page 2 devon recommends I do so anywayThe last step is optional but highly recommended: A separate aluminum plate provides additional cooling for any RAM chips on the back side of the card. It also draws some heat through the PCB, so it is useful even on cards without rear-mounted RAM chips.
Better later than never!
Madshrimps' article showed similar results, but they also tested the Accelero...
IMO Zalman could improve on the VF-900, since it's weight is only 185 g.
There is room the extend the fins by about 1 - 1.5 cm sideways.
Of course, the fins wouldn't have the same length anymore, which would slightly increase manufacturing costs.
Such a modified VF-900 could handle larger loads, upto 90 W IMO, which the current design is not really up to.
Madshrimps' article showed similar results, but they also tested the Accelero...
IMO Zalman could improve on the VF-900, since it's weight is only 185 g.
There is room the extend the fins by about 1 - 1.5 cm sideways.
Of course, the fins wouldn't have the same length anymore, which would slightly increase manufacturing costs.
Such a modified VF-900 could handle larger loads, upto 90 W IMO, which the current design is not really up to.
-
- Posts: 46
- Joined: Thu Jun 08, 2006 3:32 am
- Location: IL
CPU cooler?
I realise I'm being a little crazy:
could the zalman VF-900 be used as a low profile cpu cooler?
I assume that the die size is a major factor. but I couldn't tell how much smaller it is from say the 9500 or the actual cpu area of a standard CPU.
maybe Zalman alreay has that in development
could the zalman VF-900 be used as a low profile cpu cooler?
I assume that the die size is a major factor. but I couldn't tell how much smaller it is from say the 9500 or the actual cpu area of a standard CPU.
maybe Zalman alreay has that in development
Great review!
Just wanted to mention that Thermalright's compatibility list is pretty strict on the V-1/V-1 Ultra. I have a V-1 (non-ultra) on an x800pro (should be pretty much the same as x800xt) and it works just fine. Better than stock and very quiet. Also cut the connector off the stock heatsink and wired it to the fan for the V-1 so I could plug it into the card and control the fan with ATITool.
Setup a buddy with a vf900 on his 7900gt and I was definitely impressed, wish I could fit one in my case.
Just wanted to mention that Thermalright's compatibility list is pretty strict on the V-1/V-1 Ultra. I have a V-1 (non-ultra) on an x800pro (should be pretty much the same as x800xt) and it works just fine. Better than stock and very quiet. Also cut the connector off the stock heatsink and wired it to the fan for the V-1 so I could plug it into the card and control the fan with ATITool.
Setup a buddy with a vf900 on his 7900gt and I was definitely impressed, wish I could fit one in my case.
wow, was waiting for a SPCR review of these. I've got the VF700 on my 6600GT-AGP hooked to a zalman fanmate2 @ lowest possible setting.
i found it very easy to install. aside from the fact that Zalman states it does not work w/ the 6600gt cards, it does but there are a few catches.
1. The HSI Bridge chip needs its own cooling, some 6600gt agp's come w/ a seperate heatsink on it from the oem, like mine. (eVGA)
2. ram sinks. You can cut them up and bend them to fit under the vf700, as the HSI chip moves the gpu further up the card causing the ram sinks to get very close to the base. I just didnt use them, since my oem cooler just hovered over the mem anyways.
I find it quite good. recently w/ the warm weather in Toronto i've had ambient temps in the room of 27c (maybe up to 30c, darn A/C doesnt work too good upstairs). and it's idling @ 51c loading at 65ish.
the loudest part in my system is the s478 cooler (oem) but that will be fixed w/ an XP-120 on the way
As for the modification to allow this vf700 to mount to the cards fan plug all you need to do is pry the female plug on the fan off. it's just pressure fit on the pins, then the regular fan header should fit.
you can use RivaTuner to quiet down the fan during 2d use, and ramp it up to 85% during 3d use. works good.
i found it very easy to install. aside from the fact that Zalman states it does not work w/ the 6600gt cards, it does but there are a few catches.
1. The HSI Bridge chip needs its own cooling, some 6600gt agp's come w/ a seperate heatsink on it from the oem, like mine. (eVGA)
2. ram sinks. You can cut them up and bend them to fit under the vf700, as the HSI chip moves the gpu further up the card causing the ram sinks to get very close to the base. I just didnt use them, since my oem cooler just hovered over the mem anyways.
I find it quite good. recently w/ the warm weather in Toronto i've had ambient temps in the room of 27c (maybe up to 30c, darn A/C doesnt work too good upstairs). and it's idling @ 51c loading at 65ish.
the loudest part in my system is the s478 cooler (oem) but that will be fixed w/ an XP-120 on the way
As for the modification to allow this vf700 to mount to the cards fan plug all you need to do is pry the female plug on the fan off. it's just pressure fit on the pins, then the regular fan header should fit.
you can use RivaTuner to quiet down the fan during 2d use, and ramp it up to 85% during 3d use. works good.
I usually say, what a great review, but I am quite surprised to see a total lack of different voltages from the Arctic-Cooling unit.
Please can you re-test the AC VGA Silencer in the same system while using a peice of software to regulate the fan speed.
I think that AC got a really rough deal inthat review, and deserve to be re-tested, even if no-one is totally sure of the voltages, a % of the speed, and knowing what version of what software was used to obtain that speed would be great.
As an example, I HAD an Arctic Cooling VGA Silencer 4 Rev2 for my X800XT PE, I set various speeds using ATI-Tool, and even had it off completly for a while. It is an exceptional cooler in many respects, such as, the fact that it will run at 22% hapily and cool, and when a game is fired up it can the spin up with the temperature.
Therefore a lower noise level is attained when the graphic card is not doing a lot, but gets noisier when games are being played.
Please level the playing field.
Andy
Please can you re-test the AC VGA Silencer in the same system while using a peice of software to regulate the fan speed.
I think that AC got a really rough deal inthat review, and deserve to be re-tested, even if no-one is totally sure of the voltages, a % of the speed, and knowing what version of what software was used to obtain that speed would be great.
As an example, I HAD an Arctic Cooling VGA Silencer 4 Rev2 for my X800XT PE, I set various speeds using ATI-Tool, and even had it off completly for a while. It is an exceptional cooler in many respects, such as, the fact that it will run at 22% hapily and cool, and when a game is fired up it can the spin up with the temperature.
Therefore a lower noise level is attained when the graphic card is not doing a lot, but gets noisier when games are being played.
Please level the playing field.
Andy
I would like to know this as well. I wonder if most SPCR readers' cases have as effective airflow as the case shown in the review, and wonder even more if a higher performance vid card would have produced very different comparative results (esp. the ambient and CPU temps of the different coolers). I have read in the SPCR and other forums some users say that replacing their performance video card coolers with Zalman and other third party coolers has resulted in major jumps in case temperature, due to the inability to exhaust all that heat generated by the newer generation of performance graphics cards. By comparison, I know that even on my ATI 9800pro (45-55W under load) the Arctic Cooler Silencer dropped my case and CPU temps several degrees in comparison with stock cooling.Bobendren wrote:Awesome review. I'm suprised though that the Arctic Cooler had a consistently high cpu temp considering it chucks all the gpu heat outside the box. Why does this happen?
I think this is a terrific review as usual, but I agree with andyb that some attempts should have been made to reduce the fan speed on the AC unit via ATI Tool or if possible Speedfan to see if a better balance of cooling/noise was possible with the NV5. I'm not a die hard fan of AC units as I think they have some quality control issues. But we have to consider that: (1) vid cards are getting more and more power hungry and heat producing (the 6800GT at ~55W under load is average to low compared to vid cards today and on the horizon); and (2) the demands of quiet PC users often result in compromising case air flow to some degree. Therefore I wonder if if the AC Silencer style of GPU cooling solution doesn't still have a critical advantage *if* the fan speed can be controlled to bring its noise level down to that of the other VGA coolers.
It's always the way, you write a great article, with rigorous test results clearly presented and all you get is; yes good but can you .....
I think this shows several things:
That we're not afraid to test and modify things,
that we're always "pushing the envelope",
that we put great weight and faith in official SPCR reviews and
that we're a demanding bunch
To people asking about various things reading through the forums to see what experiences people have had can help, most products have been tried by someone here and while we might vary in our standards we're in the same ball park. Unlike the "gamer" sites that find most stock coolers fine and, for example, can't hear the V1 Ultra at full 12V / 2500rpm over the noise of the rest of their system yet Devon describes it as "too loud to consider in a quiet system"
However I too was surprised that no attempt test the cooling ability of the Silencer at different voltages as many cards allow fan speed control and many of us are capable and prepared to do some simple wiring the control the fan speed in some way. If the plug is anything like the one on my 9600 Pro you can just jam it on to a normal 3 pin fan connector. Yes the pins have to bend a bit but they bend/spring back and it’s an instant, easy way of connecting it to a fan controller. On the last page there are sound recordings of it at different voltage levels so at least that's covered, it's then a question of how the cooling performs at reduced speeds.
With regard to the Accelero SPCR can only test what they have and they're almost totally dependant on what samples manufactures will send them. Perhaps we should club together and buy an Accelero and send it for testing or just all email the manufacture and demand they send one!
Regards, Seb
I think this shows several things:
That we're not afraid to test and modify things,
that we're always "pushing the envelope",
that we put great weight and faith in official SPCR reviews and
that we're a demanding bunch
To people asking about various things reading through the forums to see what experiences people have had can help, most products have been tried by someone here and while we might vary in our standards we're in the same ball park. Unlike the "gamer" sites that find most stock coolers fine and, for example, can't hear the V1 Ultra at full 12V / 2500rpm over the noise of the rest of their system yet Devon describes it as "too loud to consider in a quiet system"
However I too was surprised that no attempt test the cooling ability of the Silencer at different voltages as many cards allow fan speed control and many of us are capable and prepared to do some simple wiring the control the fan speed in some way. If the plug is anything like the one on my 9600 Pro you can just jam it on to a normal 3 pin fan connector. Yes the pins have to bend a bit but they bend/spring back and it’s an instant, easy way of connecting it to a fan controller. On the last page there are sound recordings of it at different voltage levels so at least that's covered, it's then a question of how the cooling performs at reduced speeds.
With regard to the Accelero SPCR can only test what they have and they're almost totally dependant on what samples manufactures will send them. Perhaps we should club together and buy an Accelero and send it for testing or just all email the manufacture and demand they send one!
Regards, Seb
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Jumping in from Taipei because I'm pretty sure Devon is immersed in a several day film shoot (his other -- real? -- passion):
It was a very good review, I agree with those who said that, and Devon deserves kudos. We spent a lot of time thinking, trying, and retrying different testing schemes, and I think the one we ended up with is good enough to stand as a reference methodology for some time, which means future vidcard reviews will happen much quicker. Some of those coolers have been in our lab for months, I'm chagrined to say.
Yes, I agree adjusting the voltage on the AC NV5 would have made it more complete, and perhaps we can do that when the Accelero is tested. That cooler wasn't done because we never requested a sample -- didn't want any more vidcard coolers when we weren't sure how to test the dang things. Several new ones will be tested in the near future -- a new Thermalright, Accelero and possibly, one from Scythe. There may also be new brands -- Gigabyte, and some other big Taiwan brands are doing tons of coolers, some VGA coolers among them. We'll see whatComputex nets in the coming weeks.
It was a very good review, I agree with those who said that, and Devon deserves kudos. We spent a lot of time thinking, trying, and retrying different testing schemes, and I think the one we ended up with is good enough to stand as a reference methodology for some time, which means future vidcard reviews will happen much quicker. Some of those coolers have been in our lab for months, I'm chagrined to say.
Yes, I agree adjusting the voltage on the AC NV5 would have made it more complete, and perhaps we can do that when the Accelero is tested. That cooler wasn't done because we never requested a sample -- didn't want any more vidcard coolers when we weren't sure how to test the dang things. Several new ones will be tested in the near future -- a new Thermalright, Accelero and possibly, one from Scythe. There may also be new brands -- Gigabyte, and some other big Taiwan brands are doing tons of coolers, some VGA coolers among them. We'll see whatComputex nets in the coming weeks.
Since the only VGA coolers listed on Scythe's site are the ones from Arctic Cooling, I'm hoping/assuming you mean a new product, something we haven't heard of before. Could it be this thing?MikeC wrote:and possibly, one from Scythe.
Looks pretty hard to fit in any but the most roomy cases... and is that a heatlane? I haven't seen that since the Zen.
This is definitely something I've been waiting for though, a fanless VGA cooler that seems like it was actually made for fanless operation. Zalman's offerings in this area do their job, but their heatsinks never seemed to be shaped for low-airflow performance.
The picture is from this Computex report.
-
- SPCR Reviewer
- Posts: 1850
- Joined: Sun Sep 21, 2003 11:23 am
- Location: Vancouver, BC, Canada
Ok, there's a lot to tackle here, but let me try:
Desktop cases that don't trap heat under the VGA card.
Will the Antec NSK2400 suffice? You might block some air vents though...
VF900 as a low-profile CPU cooler
I think the main issue would be the mounting system, not the size of the base. Modern CPUs tend to ship with heatspreaders, which aren't as fussy as bare silicon about how much of the surface is covered. For example, the base of the original Arctic Cooling Freezer is actually noticably smaller than AMD's heatspreaders, but it works just fine.
With all of the different mounting holes on the VF900, I suspect that at least one of them could work, but the problem may be the height of the die, not its area. Even if the mounting holes could be lined up correctly, I think it's likely that the posts wouldn't be tall enough. I would expect to be faced with some significant customization...
Lack of data for the AC NVSilencer 5
I'm surprised people are still interested in this cooler with the Accelero set to succeed it. As MikeC stated, it was reallly only tossed into the mix as a reference because we had it on hand. In retrospect, I probably should have tested it both ways — once affixed to the VGA header, and once controlled externally. I was actually unaware that RivaTuner could control fan speed; I only discovered this program at the very end of the testing. Neither SpeedFan nor ATITool allows the fan on our test card to be controlled.
At this point, it will be quite some time before I will be back in the lab with enough time to retest it, so things will have to stand as they are for now. Perhaps we can revisit it again for a third time (!) when we look at the Accelero.
CPU temperature with the AC NVSilencer 5
I didn't really think a 3°C difference from high to low was worth commenting on. I would blame measurement variance and slightly different ambient temperatures on different days of testing.
Desktop cases that don't trap heat under the VGA card.
Will the Antec NSK2400 suffice? You might block some air vents though...
VF900 as a low-profile CPU cooler
I think the main issue would be the mounting system, not the size of the base. Modern CPUs tend to ship with heatspreaders, which aren't as fussy as bare silicon about how much of the surface is covered. For example, the base of the original Arctic Cooling Freezer is actually noticably smaller than AMD's heatspreaders, but it works just fine.
With all of the different mounting holes on the VF900, I suspect that at least one of them could work, but the problem may be the height of the die, not its area. Even if the mounting holes could be lined up correctly, I think it's likely that the posts wouldn't be tall enough. I would expect to be faced with some significant customization...
Lack of data for the AC NVSilencer 5
I'm surprised people are still interested in this cooler with the Accelero set to succeed it. As MikeC stated, it was reallly only tossed into the mix as a reference because we had it on hand. In retrospect, I probably should have tested it both ways — once affixed to the VGA header, and once controlled externally. I was actually unaware that RivaTuner could control fan speed; I only discovered this program at the very end of the testing. Neither SpeedFan nor ATITool allows the fan on our test card to be controlled.
At this point, it will be quite some time before I will be back in the lab with enough time to retest it, so things will have to stand as they are for now. Perhaps we can revisit it again for a third time (!) when we look at the Accelero.
CPU temperature with the AC NVSilencer 5
I didn't really think a 3°C difference from high to low was worth commenting on. I would blame measurement variance and slightly different ambient temperatures on different days of testing.
-
- *Lifetime Patron*
- Posts: 495
- Joined: Thu Apr 15, 2004 3:53 am
- Location: In Front of PC
- Contact:
the Accelero is not a replacement for the Silencer series, it's different product line. that's what AC told me at leastAccelero set to succeed it
The V1-Ultra shows it muscles when using a high-end (hotter) VGA card, the 6800GT is quite cool compared to 7800 and X1900 series, the scale will tip in favor of the V1-Ultra due to its massive heatpipe construction which comes into play at higher heat outputs. the VF900 will keep up nicely and even the Silencer will offer a great balance between noise and performance. the VF700 will be overwhelmed easily though, and higher temps than the stock cooler can be observed.
afaik it's not available for retailAny chance of SPCR testing the VF100?
not before the end of the year according to Scythe, it should be a passive HSF , much like the Thermalright one at their boothand possibly, one from Scythe.
Regarding the interest in the Silencer. I dont know for sure, but I believe the two main reasons for the interest is:Devonavar wrote:Ok, there's a lot to tackle here, but let me try:
Lack of data for the AC NVSilencer 5
I'm surprised people are still interested in this cooler with the Accelero set to succeed it. As MikeC stated, it was reallly only tossed into the mix as a reference because we had it on hand. In retrospect, I probably should have tested it both ways — once affixed to the VGA header, and once controlled externally. I was actually unaware that RivaTuner could control fan speed; I only discovered this program at the very end of the testing. Neither SpeedFan nor ATITool allows the fan on our test card to be controlled.
At this point, it will be quite some time before I will be back in the lab with enough time to retest it, so things will have to stand as they are for now. Perhaps we can revisit it again for a third time (!) when we look at the Accelero.
1) The basic design is brilliant.
2) Many of us had it / have it.
Personally, I believe AC could have taken it a class or two up, if they improved quality. Better fans, better polishing of the contact area between GPU/HSF. And maybe add in a heatpipe or two. Enough about that, back to the review
I think the main reason for people commenting on voltage for the Silencer, is the point about using the Silencer as a reference.
For me, a geusstimated voltage of 10 volt, dont look like much of a reference
Another thing to consider, is there's several reports of different VGA card's PWM regulation of the fan, which introduce noise. Especially with Silencers. Maybe, the Silencer is more quiet than the VF900...
All in all, I think using a FanMate2 will make it a better choise as a reference.
But dont get me wrong, I really like this review:
1) It's simply the best VGA review I've read (and I've read a LOT)
2) You've done a great job by creating and developing the test methology.
3) Well chosen coolers.
4) Re-use of usual SPCR standard way of doing things.
One more point: If this review wasnt as great as it is, I wouldnt spent time on commenting it.