This still looks inferior to (say) beQuiet! Polar Freezer. PF is very efficient if you open a few PCI slots near the video card. That forces the air to blow directly on the heat sinks. PF also has a very large HS very close to GPU. Maybe the heat transfers better?

It would be nice to see actual comparison of most popular passive VC coolers, lie Polar Freezer and Zalman ZM80D-HP.

hmmmm... I doubt it. That polar Freezer is really just a knock-off of the Zalman zm80... which is pretty limited without direct airflow. We dida review of an early version of the zm80 years ago -- and it struggled in a low airflow case to cool a nv ti4200. Newer versions of that cooler are little changed. The fact that the Condor has 3 heatpipes & probably at least double the cooling surface area suggest it should be far better than any zm80-type coolers.

rofl, i knew they recommended toothpicks to align mounting holes, but I didn't know they actually included a pair to make your life easier. On one hand you could say it's unprofessional, on the other you could say they provide the cooler with complete set of mounting tools you need to install it...


Regardless of the toothpicks I sort of expected these test results. I was really surprised when AeroCase tested 1900 at 88 degree load in 28 or so ambient temperature. That seemed unbelievable. Yes, the heatsink is brute, it is just two aluminum blocks pieced together and joined to the core with heatpipes. Perhaps Scythe Ninja is just as brute (at least in concept), but effective heat dissapation area of condor is much much smaller, and it has to deal with GPUs (1900XTX) that run as hot as hottest prescott CPUs.


I wouldn't go as far to say as you can't use condor with 1900 because 1900 is not suited for a quiet computer in the first place, like review did. People have different tastes, and wanting to have quiet gaming computer is not unreasonable. It depends how much you value framerates against how annoying computer noise is to you. However, I wouldn't recommend using condor with 1900 for a different reason: the cooler clearly cannot cool the card effectively without help of external fan. With external fan, the noise level is higher than that ov VF900. And that's with 22 degree ambient temperatures which is not always the case in real world (we have enough people in SPCR forums that come and say they built a system in winter, but it's summer and it overheats).


I suppose one could still use condor passively even on 1900xtx with fan off in regular 2d mode and switch on the fan only when gaming, but that's not an ideal solution. So all in all I feel that while many were led to believe condor will be an ideal all-purpose high-end solution, it's really not. It does have its merits, but who knows how Aerocool VM-102 would have performed on the same SPCR test rig?



PS on a feminism sidenote, I wonder kind of PR would a company get if they publicly stated they are proud to have all-man management team... It goes against the company "professionalism" making it look more like an attitude. So, IMO it's ill advised to make such statements, be your management team all male or all female.

Nice review
Thanks Devon.

In case it's of use to anyone, Here you can find a thread with a fair amount of already asked questions.

The article confirmed my belief, that the product is over-hyped.
And it can kill your expensive gfx card by shear weight, too.

If only I could find a way to manufacture my own design, that has over 4,000 cm2, 6 heatpipes and shouldn't weigh over 350g...

One down side I see on the Condor is that its fins are alinged in the opposite direction of the airflow that most of the users have in their cases.

When mounted, the fins can only take advantage of natural convection and if there's a front intake blowing towards it, the airflow from that intake would be ineffective.

This is just terrible.

With only one case fan in the upper chamber of a P180, a nexus at 7V and a BeQuiet! Polar Freezer mounted on a 6600GT (50W load) I'm running at 70C load (ambient somewhere around 20-25C). With one open PCI-slot, even lower.

Ok, the Condor can't cool a 1900XTX passive. The big question is what is the fastest card it can cool passive. It should be way better than the VF900 passive, right?

Personally, I think it looks like a fairly good proposition for a fairly high powered card like the nVidia 7900GT. There's no way you can argue an X1900XTX is power efficient, and it doesn't make a good match for a near-silent system due to its ridiculous power draw - which is the only place you'd want a passively cooled graphics card anyway...

As for cards bending, I've never heard of such a thing killing a card before. If it does, I'm sure we'll hear about it in short order.

Finally, just as a nitpick - I spotted this: "16 sample anti-aliasing and 6x anisotropic pixel shading" - do you mean 16x sample anisotropic filtering and 6x multisample anti-aliasing?

*Edit* Oh, and Ackelind - it sounds like you have a very well optimised airflow setup there - just what this cooler would need, and what it seems it didn't really have in the SPCR system. It's also worth bearing in mind the 6800GT they tested with has twice as many pipelines as your card and the heat from the AGP-PCIe translator chip to deal with too.

It seems to me it shouldn't be too difficult to support the heatsink itself directly, either by blocking it up from the case floor, or hanging a wire from the top of the case (should you have one with the fins above rather than below the card). Supporting the corner of the card won't take any of the cantilever stress off of it, but directly supporting the heatsink will.

Without trying to sound rude, I am not just dissapointed with this review (the only one on SPCR so far), I am also shocked that as mentioned on page 3 "Another trick is to remove the covers of the PCI slots to encourage additional in/out flow near the heatsink." Why was this not tested.???

The slot cover (spine) directly below the graphics card seemed to be empty, but this was not specifically mentioned, and not having that slot covered will do very little to aid the Condor.

The Condor review should be tested with one or more PCI slot covers removed, this may make the temps drop by 10+ C.

Likewise, the "reversed-wing" version was tested, which is a much worse scenario for cooling than the "normal-wing" version, it looks like there is enough room in that test setup to fit it. This of course would be a whole new review which I wont expect as I fully understand how long they take, however some products have been reviewed twice before with very different setups - the P180 to be precise.

Please dont think I am "SPCR" or "Devon" bashing, but this review is not up to the usual SPCR standards I have come to expect. If I was an "Aero-Case fanboy" I would suggest that SPCR has deliberately shown the Condor in the worst possible light, and deliberately kept as much airflow away from the condor as possible.!!!


Andy

It's obvious that you cannot "passively" cool any >100W video card, and that forced airflow via case fans is absolutely necessary.

Yes, we might have done more to increase airflow to the Condor without actually adding a fan in front. Opening up all the PCI slot covers would probably have helped a bit, at the expense of HDD cooling in the front. Would it have been enough to run this system with a single 120mm Nexus fan? Highly unlikely, and as others have already noted, room ambient temperatures will often be much higher than the 22-23C of our lab.

Remember, too, that Aerocase recommends both intake and exhaust fans in systems that use their VGA coolers. All we did in our test setup was to position the front fan a little closer to the vidcard and cooling fins.

More importantly....

Given the ramp-up of the PSU fan -- even though it was separately ducted for cool intake air! -- trying to run the Condor "fanlessly" with the X1900XTX under such loading conditions was pointless. Perhaps some of you did not read the text carefully enough: The system noise (minus VGA cooler fan) under full cpu/vga load was ~30 dBA@1m. As noted in the review, "no other fan-cooled power supply would fare better."

That might be a slight overstatement -- perhaps the fan in an S12-600 in a P180 would ramp up less and run a bit quieter at the same load. Still, when the DC power demand is near 250W, there are no PSUs we know of that would NOT ramp up its fan at least some. Even at 80% efficiency, you're looking at 50W inside the PSU -- no normal PSU can dissipate this heat without forced airflow.

That leads us to the question of a passive PSU system. However.... I cannot think of a single passive PSU + case for a gamer that would accommodate a hot CPU (like the 2.8GHz prescott) and an X1900XTX with a single Nexus 120.... that I'd be happy and comfortable with in the long term. (IE -- that I didn't think would fail early due to overheating.)

Which then gets back to the question of whether one can or should expect to run a super hot vidcard like the X1900XTX passively with the Condor in a quiet system. Yes -- if you accept 30 dBA@1m as quiet. No, not at SPCR preferred 22-23 dBA@1m levels. (Besides, the only reason for a vidcard like this is to play games; gamers get mesmerized and immersed in their games, have the sound up and hardly are aware of anything else when they are gaming. I don't see what's wrong with a switch for the front fan to turn on when gaming.)

But having said all that, if we'd had a 35W A64 3800X2 or similar for the CPU and a "normal-wing" Condor, then the single Nexus 120 & the quiet CPU fan might have been enough to cool it all. Then we'd only have to solve the issue of the 30 dBA/1m PSU. An external passive PSU, right? We should have tested it that way, right?

------------

Finally, be reasonable, people. We have umpteen products to review and only 24 hrs in a day. Surely there is more than enough data and information in the review to reach the same conclusions I did above.

I have a question about power draw in graphics cards.
I've seen lots of W numbers for various cards, many of them coming from xbitlabs. I have no idea though how much the GPU itself draws? There are VRM modules, RAM and other chips on the board, besides the GPU. My guess is that the GPU uses 70 % at most. What do you think?

I don't think this is an important question as the total power draw tells you enough about the GPU, but maybe something to remember before you start comparing it to CPU numbers (which also sometimes includes VRM loss though).

Mike, all of your points are totally valid, but please also note that I never mentioned the X1900XTX.

The vast majority of gamers cant afford such an expensive graphics card, and on top of that most of the readers on SPCR will understand that they will have to make a compromise somewhere.

A lot of people will compromise on the graphics card, and likewise the cooler of the graphics card, the Condor is not as cheap as some coolers, however I divulge........

A huge amount of SPCR readers will be put off of buying the Condor, and not because of its price, but because most people will feel that their graphics cards (7600GT, 6800GT, 7800GT, and 7900GT's a few common cards) dont deserve to be tortured at 90C whilst playing games with 1x120mm nexus fan running at a mere 7V.

However if those temps were 5-10C lower, and proven so by an SPCR review, then the Condor would be a perfect compromise for all of those people who have compromised with every other component, be it by price, heat, or performance.

My system (spec in Sig), has been a long list of compromises, and if I was in the market for a new Graphics card it would probably be a 7900GT, which I am certain could be cooled passively by the Condor, its power draw is similar to the 6800GT. I wouldnt exactly be thrilled by the GPU hitting 80+C, however I would be very happy if the case fan ran faster when needed as I am not in the habit of playing games without sound, and would be exstatic to have such a fast GPU when gaming without the enormous cost of an X1900XTX, and have 1 less fan when not gaming.

My system is not perfect (I am working on it still....), but I am very happy with my current setup, its very very quiet when not gaming, and gets louder as it gets hotter (gaming).

RE: The point about the PSU becomming the noisiest part of the PC, their was no mention of the PSU ramping up with the 6800GT, even with a PressHOT in the system to add to the power draw. The system hit 232W, does that mean that the PSU did NOT rapm up at all.??? Does that mean that the system with the Nexus at 7V with the GPU @ 90C, the system was at 23Db.???

Could the noise levels for the WHOLE system be added to the case fan speed comparison table on page 5 for the 6800GT, there is no point adding those figures to the results of the X1900XTX.


Andy

Sorry Andy, but I think this is presumptuous: You can speak for yourself, but not for "a huge amount of SPCR readers". How do you know what they think? Besides, the stock cooler will let those cards reach such temperatures anyway! And the system does not have to be running with the Nexus at 7V to be at 22-23 dBA -- that is with the Nexus 120 at full tilt.


AFAIK, the PSU did not ramp up with the 6800GT, or if it did, not by much. I am not as intimately familiar with this part of the test; I am sure Devon will chime in when he gets a chance to do so. 232W is only ~180W DC power, and the PSU is ducted, so it would do much better than in our review test... where it ramped up from 21 dBA at 150W to 26 dBA @ 200W... with the intake air temp at 33C and 36C at those power levels. With the intake vent/duct, the PSU runs much cooler than in the test rig; the intake temp stayed at 22C the entire time during testing.

i'm thinking about getting the condor's little brother, the raven...
does anyone have any idea how the raven would fare on a 6800GS?

even the condor didn't cool a 6800GT particularly well and since the raven only has 2 heatpipes and less surface area, i'm wondering if it could handle one. my setup would be the standard wing (wing over the cpu cooler/exhaust fan). the aerocase site says that the raven can handle a 6800 but looking at the condors results, i'm a little wary of that claim.

anyone willing to extrapolate an estimate of the raven's performance?

Can't seem to find any info on this fan. Anyone managed to do so?

Whoa this is one of the most argumentative ive seen SPCR before personally I think that this is a good review you show what this Heatsink can do with the Hottest cards in bad airflow and from the results of this I can extrapolate the differences to lower powered cards and better airflow, I would be more than willing to use one of these on a card such as the 7900GT from what the results say.

Fat-bloater-dave, this is no argument, this is a discussion, and I am sure that Mike will say the same thing.

The discussion from my point of view is not so much about the performance of the Condor, but of how the Condor affected the whole system. I feel the review does not contain quite enough information, like the overall system noise. I understand that this test rig has been used on many many graphics card coolers so the results of one cooler can be compared directly to the next, however with the 6800GT and the Condor, the noise results for the whole system are not listed, unlike other graphics card cooler reviews, such as the GPU cooler roundup including the Zalman VF900 (which persuaded me to buy one for my brothers 7900GT).

Maybe as the Condor is SPCR's first "passive" graphics card review its not fair to compare it on the current rig, I know that more "passive" GPU coolers will hit the market, and as such SPCR may require a new "passive" GPU rig so that the rig represents how the end user would cool the passive coolers.

AC have a couple of passive GPU coolers in the pipeline, they are now delayed until October, Aero-Case have 2, how long will it be before Zalman jump into the fray (again).


Andy

Interesting test results!

Read the footnote of the last table. It is clearly stated there.

Sorry Andy to pick you as an example. But I see a lot of posts in this thread from people who discard at least half of the things that can be learned from this review. And then make their own negative conclusion.

When the Scythe Ninja was reviewed, it was put on the standard two level plywood platform testbed. It couldn't cool the processor passively that way. I didn't see a lot of comments that the Ninja was not up to it's task then. More like everybody agreed with the reviewer then, that with a better planned airflow it could work. See the similarities with this review?

Tibors, I do see a similarity with the Ninja review, however the foot note in table 2 is for a different graphics card, and the power draw is a fair bit higher, that is as relevant to testing the ninja horizontally or vertically, it is different.!!!

Most people who have been reading SPCR for a while will have their own thoughts, opinions, and ideas, the possibilities of the Condor are superb, however as I have said several times before, the test setup with the Neo HE + 1x120mm Ninja the Condor and the 6800GT has no noise references at all.

I appreciate what SPCR are trying to do with their GPU cooler reviews, they are trying to isolate the noise of the cooler with X setup, which makes one review directly comparable to the next, but IMO the most funamental part of the review was missed, the overall system noise. Everyone who has read the review can see the test setup in minute detail, and its not an unreasonable setup, but stating the overall system noise seems to be the only important factor missing. If the overall system noise was recorded then adding it to the site would be very easy.

I have read the whole review twice to make sure I have not missed anything, and I am certainly not trying to be negative for the sake of being negative, I consider myself part of the SPCR community, and as such I want to help my fellow silencers, and where possible help the SPCR team with constructive critisism, and I consider this review flawed, and as such am expressing my opinion. Whether the review gets updated or not depends on whether its quick and easy to do as much as whether the SPCR team agree with my points and want to polish the review.

If any of the SPCR team say that they dont want to edit the review, please just tell me, my point has been made, and there is nothing more that I can do or say, and the answers that have been added have cleared up many of my questions anyway


Kind Regards Andy

A GF7900GT card runs a lot cooler than the older GF6800GT models: It's why you can buy passively cooled GF7900GT cards off the shelf.

For comparison a GF6800GT fitted with a ZM80D-HP heatsink and 80mm Acoustifan Dustproof at 5v idles at 64c whilst a GF7800GTX 256mb card fitted with the same heatsink and cooling idles at 47c.

The newer 90nm core GF7900 series cards are cooler again.

http://www.vr-zone.com/?i=3335&s=8


(taken from the VR-Zone.com review above)

Thanks for the update WR304, very useful info.


Andy

andyb --

For your benefit, I've dug up measurements for the system noise with the 6800gt and added this into to the first results table in the article. The PSU fan did ramp up a bit, just enough to bring overall noise levels up ~2 dBA@1m. With the back Nexus going at 12V, the overall noise was at ~26 dBA; with it at 7V, it was ~24 dBA. What this suggests is that adding a quiet front fan at 7V would result in no increase in noise, but certainly aid the Condor cooling substantially.

It would be nice to try the Condor passively with all the PCI slots open, and perhaps the front vent closed up, forcing the 120mm case fan to use the PCI slot openings as intake vents, which would then bring more air across the Condor HS.

However, no retesting with our 6800GT is possible: Just after this review was completed, the AOpen card finally gave up the ghost on us. It's a bit of a surprise that it lasted as long as it did -- well over a year with umpteen cooler changes and much abuse in the lab.

We're trying to replace the card with another of the same or very close power dissipation, but do not have one in our possession yet. A 6800XT was purchased, but it turns out to have substantially lower power.

The closest candidates according to xbitlabs power measurements are:

eVGA 7900 GT CO
7800GT
Radion X1800XL

any 6800GS/GT will also do, of course, but these are harder to find these days.

If anyone wants to make us a donation of any one of these cards, we'll happily accept -- and give you a brand new 6800XT in exchange, if you like.

Hello everyone,

I’ve wanted to stay out of this one to keep the whole review process as uncluttered and independent as possible … but, at the same time I do feel a part of the SPCR community. So I have a just little bit to say as a member.

First, for the record, I do believe the review was fair … given the testing environment. However, therein lays the problem. Was having Devon and Mike test a X1900XTX fair in the first place? Probably not. As they pointed out, you are NOT going to run a 120+ watt video card in a silently configured machine; period. We’re talking a total of over 200+ watts of heat inside an enclosure just bigger than a breadbox. We could build a nice oven and BAKE bread with that much heat …

The fact is, before this year no one would ever have imagined video cards this hot. PC enclosures were just never designed for these kinds of power and thermal loads. Trying to keep our old technology (i.e. last years) cool with a minimum of fans was a challenge. Now, given these cards, it’s nearly impossible in a conventional case. Which is exactly our point at Aerocase. As a community we need to start thinking OUTSIDE the box. One way or another, our future cooling is going to be external. NVidia has already announced their external quad card box, complete with it’s own PSU, and I’m sure AMD/ATI won’t be far behind.

Aerocase isn’t just about high-end video cooling … it’s about silent cooling. We can’t possibly keep up with 250 watt crossfire and 400 watt quad systems passively … there is just no way. We started this company out with idea we could eliminate fans (or nearly) all together by building enclosures mounted with large external heatsinks, and to do in less expensively than Zalman. The video cooling was an afterthought … a way of providing our new heatpipe technology at a entry level price. Then the X1900 and 7900 hit ... so we said, why not go for it? And, it works, *IF* you have enough airflow. But, it doesn’t work, nor was it ever designed to, without front and rear case fans moving 40+ CFM through the case.

Anyway, in conclusion, I should point out that our smaller Raven mounted on a 40-50 watt card really *IS* a good match in an SPCR low noise environment and probably should have been the one tested. It still needs airflow of course, but it can be done with just the rear case fan.

So, thank you to AndyB and those members coming to bat for us on this review. But, let’s stop bashing Devon and Mike. They did a great job, and my complements go out to them.

Thanks for reading,

Ann Goodheart
Product Manager
Aerocase, Inc.

You guys give up too easily!

I managed to cool a 325W system built around an overclocked 150W 830D with four slow fans at less than 25 dBA. To me, a 135W video card paired with a Conroe or AMD CPU doesn't seem that hard.

I'm actually in the middle of upgrading my system: I've got an eVGA 7900 GT KO already, and an E6600 on a P5W DH Deluxe, with a Condor on the way. My plan is to use only three fans: one in the PS/HD chamber, one on the CPU, mounted on an XP-120 and ducted out the rear, and one cooling the Condor and the TV card in a segregated chamber (bottom half of the MB area).

Will write this up when done.

Thanks Mike and Devon, those Db's are fantastic, and with the input from WR304 almost anyone can pair a condor with a 7900GT an make an extremely quiet system whilst gaming, and a near silent system when idle.

I do honestly appreciate the extra info and detail of the "late" 6800GT, I would also like to give my condolences for your loss


Andy

You make some very perceptive points Ann, and I'm sure everyone here is very interested to get some input from the "other side of the fence" as it were. The simple fact is GPU's now are going through what CPU's were in the Netburst era: progressively higher and higher power consumption, and consequently heat dissipation. Really, in the long run the solution is for GPU makers to get the "performance per watt" bug like Intel and AMD have, but that does not seem to be happening yet.



LOL, soon people are going to have two computers, one to do the GFX and one to do everything else!! ROFLMAO, insane!