Well, promised review is here.


Background
I have a 6800GT in my system and up until now I've been using Arctic Cooling Silencer 5 Revision 1 to cool the GPU. It wasn't a stellar performer neither in terms of temperature, nor in terms of noise, but it was still much, much quiter than stock solution. I didn't even really noticed how noisy it was until I moved my system to P180. After I got used to the new noise floor of my new system I began to notice how noisy and annoying ACS5 was. Even at 7V which is the lowest my rheostat can go it was about as annoying as my four hard drives, which should give you an idea how loud it was. The noise character was of course much different than idle whoosh of hard drives, but it was just as annoying. For one it clicked like mad, apparently a very common noise that appears after several months of use and the second there was a howl whooshing sound. Not good. The more I used the system, the more unbearable the noise was becoming, especially at night. So I decided to replace it. Among the choices were new revision of ACS5 - 3rd revision, Zalman VF-700, Aerocool VM-101/102 and Thermalright V-1/V1 Ultra. You can read further at the original thread I created here. For a variety of reasons I chose VM-101, expained once again in the original thread.


System
Asus P4P800 motherboard
P4 3.0C>3.4C Northwood CPU with Scythe SCNJ-1000 Ninja mounted
2 x 512Mb Geil Memory
PNY 6800GT>ULTRA
AudioTrack Prodigy 7.1 Sound Card
Promise TX-2 IDE controller
Seasonic S-12 600 PSU
Seagate Barracuda 80Gb IDE
Seagate 7200.8 400Gb SATA
Hitachi 7K250 250Gb IDE
WD 2500JB 250Gb IDE
(yep, 4 hard drives in total, 11 HDD platters)
3 Yate Loon SL12 fans 1350RPM stock speed
first is mounted on SCNJ-1000
second is softmounted in the lower chamber
third is softmounted on rear exhaust
Peripherials are inconsequential to the test


First Impressions
1. Initially, when researching, I was a bit concerned about the weight of the heatsink since it is installed on the front of the card (bottom when you mount it in the case) and pulls down the GPU block, which might have lessened the GPU contact. However when I took the heatsink out of the case it turned out to be really really light, so I don't believe it is actually going to be a problem for anyone.
2. The surphace of the GPU block is actually quite rough, about the same quality as ACS5 and nowhere near the polished quality of Zalman and Scythe heatsinks. I didn't try to lap it since I don't really have any sand paper and I don't know how to do it, so I let the AS5 do the job.
3. My third impression was "holy crap" there is no way this is going to cool down my 6800GT>ULTRA, it's gonna go down in flames and take my whold PC with it... Whether this is turned out true or not, keep reading...


Installation
VM-101 isn't directly compatible with 6800GT because mounting holes around the GPU are 2mm wide and the screws included with VM-101 are 3mm. Of course one could use 3mm mounting holes that are further away from GPU but that would put less pressure on the GPU core so I opted for buying some 2mm screws with washers from Ace Hardware and using those. This also gave me the freedom of using four arms to hold the GPU block. The manual says you should use only two arms diagonally to mount GPU block and there are actually only two mounting screws to do so, but technically nothing is stopping you from using all four arms. The cooler arrived on Friday, so as soon as I came home I unpacked VM-101 took all necessary screws and rubber o-rings and went to the Ace Hardware. I didn't know which length I was looking for so I bought four 12mm ones... which turned out to be too long. Unfortunately it was too late for a second visit to Ace Hardware as it already closed so the actual installation was delayed until Saturday morning. So off I went to Ace Hardware again on saturday. Apparently they had 12mm and 8mm screws, the first ones were too long and the latter too short. D-oh! So I had to shorten the 12mm ones. Took me about half an hour to snap and sand the ends.

The actual installation after that was not as hard as many reviews would make you believe, but it was quite awkward. The hardest part was putting the GPU block on the card. At first I tried alligning o-ring rubbers with the holes around GPU and then gradually lowering the GPU block with four arms attached but that didn't work out, you need a precision and steady hands of a machine to do that. If you can do it - that's great, but my shaky hands couldn't, something was always out of alignment. So in the end I just plopped the GPU block onto the GPU and the friction from AS5 was good enough to hold it in place. I then lowered all four screws into the place through the mounting holes and used nuts that came with ACS5 to secure GPU block. As you remember, I used 2mm screws so the nuts that came with VM-101 were no good. So I used the ones from ACS5 instead - well, at least it's good for something I then fastened the screws until I couldn't move the GPU block if I put a reasonable amount of force on it. The whole construction is actually quite secure, at least much better than what I was lead to believe from reading various reviews.

I used AS5 for both GPU and the heatpipes. Mounting heatpipe heatsink to the GPU block was easier, but I must say the four screws they included were complete shit. I tried three different philips screwdrivers and all three kept slipping. Very shoddy work, especially since after the heatsink it already in place you can't put any force on the back of the card. Somehow I managed to slip my fingers under the heatsink as far as I could and secure all four screws. They are holding, I think, but I'm kind of disappointed.


Temperatures
Ambient temperature was 23 degrees Celsius, at least according to both of my thermometers, I don't know how much you can trust those. When reading the following secion, keep in mind that 6800GT>ULTRA is probably one of the hottest cards on the market. I believe that the new 7800 cards actually run cooler. My 6800GT overclocked to ULTRA may not be as hot as the real ULTRA since those have higher GPU voltage than GT's, but it should be pretty close. Quite a beast it is...

1) There is no any direct forced airflow over the card, only the residual airflow from the YL mounted on the CPU heatsink and induced draft from the rear exhaust fan.
2) The CPU fan in the test was spinning at 900RPM and the rear exhaust fan was spinning at 950RPM according to the sensors, but I believe it was actually closer to 1000RPM
3) I left slot cover directly below the videocard open to let the rear exhaust fan pull air over the card.


Like all huge heatsinks with no forced airflow at them (think Reserator) it always took quite a while for the temperatures to get to the equilibrium point so the actual readings below may not be entirely accurate as I don't have a lot of patience to wait for the temperatures to stabilize.


2D Mode 200Mhz
I downclock my card to 200Mhz in 2D mode since it supposedly should give me lower temperatures (in theory).
In this mode the card idled at 60-61C.


2D Mode 400Mhz
A little relatively unknown fact is that when using certain applications accessing higher end GPU functions, GeForce cards will switch to 3D mode even if there is nothing 3D on your desktop. A typical example is running a movie. For this part I rewatched the entire Pulp Fiction, was fun....
Anyway, in this mode the card was ran at 61-62C

Subjectively, the GPU block is quite hot, but it does not burn my fingers, I can hold my fingers on it indefinitely. The heatpipes are slightly cooler, but not by much. The heatsink fins are reassuringly warm, but it's expected from such heatsink construction due to further heat loss and air flowing over the fins.

Quite interesting as you can see there is only 1 degree difference between 200Mhz and 400Mhz. I know some people claimed 5C reduction in temperatures when downlocking to 200Mhz, well, not in my experience. These results are actually consistent with what I had with ACS5 too, reducing clock to 200Mhz provided almost no benefit at all.

For a brief moment I turned my rear exhaust fan up to full 1350RPM speed and the temperatures came down to 59-60C, after I turned up the CPU fan as well to full blast at 1350RPM the temperatures went down to 55C. Not bad, however the noise was more than I was willing to tolerate so I ramped down the fans to their default speeds 900 and 1000 RPMs for CPU and Rear exhaust fans respectively.

So far so good, especially since ACS5 kept my card at 56-59C in idle. VM-101 temperatures were a little higher but considering it came at zero noise penalty it's actually quite good IMHO.

I know that some of you will say that these temperatures are a bit high but they were always like that no matter which heatsink I used. With stock fan my idle temperatures were around 60C, with ACS it was at 56-59. I've read that people had GPU temperatures at high 30's/low 40's when using Zalman VF-700 at 5 volts, and while there is no doubt that Zalman is much more effective than stock and VM-101 solution, it still seems a bit too low to be true from my experience. I don't know why I have such relatively high temperatures, it may be because of the difference in thermal probes calibration, or it may be because Zalman VF-700 is truly that much better. I simply don't know, but what I do know is that in 2D mode VM-101 is about as good as stock solution while absolutely silent.

OK, enough of that, lets go to the next part. Just for the future reference when gaming my typical temperatures with ACS5 were 67-69 degrees Celcius.


3D Mode 400Mhz
This is probably the weakest point of VM-101. I tested the video card with a relatively old game, Enemy Territory. It is an old game, but it still puts my PC down to 50-60 fps at times and frankly, I never noticed much difference in GPU temperatures when playing ET vs Far Cry. In this mode the temperatures quickly rose to 72 degrees and continued to rise up until 77-80C at which point they seemed to stabilize. The nvidia temperature monitor was a bit jumpy at this point with temperatures quicly changing from 77 to 79 and back. A bit unusual I think, but what do I know how thermal probe works, could be perfectly normal. Yes, this is 10 degrees higher than what I had with ACS5, but my PC was completely stable and once again, this came at zero noise penalty. So while I may not be totally satisfied with the results, who cares? Of course there is always a chance of long term damage as it happened with my Radeon 8500 which I ran fanless with stock heatsink no less for 18 months, but oh well, the things we are willing to sacrifice for silence. And besides I don't believe this will be the case with 6800gt.

Once again, for the fun I turned up both of my upper chamber case fans to the full RPMs and the temperture quickly came down to 69-70 degrees. So if worst come to worst I can always turn up the case fans while gaming for lower temperatures.


Noise levels after installation
Well, of course, now that I got rid of this noisy ACS5 my system is much more quiter. Of course that comes with a price as other noises are much more noticable. For example a coil whine that sometimes is coming either from the motherboard or from rheobus whenever I scroll through the webpages is much more noticable. The HDD noise is also much more annoying now, from what I hear one of the HDDs now has a coarse whoosh, or rather high pitched ringing, somewhat similar to the one one Seagate 7200.9 160Gb made during recent HDD roundup by SPCR. It's really getting on my nerves right now and I really hope it's not 400Gb Barracuda since I was planning to replace all other drives anyway, while leaving 400Gb barracuda in my system. 400Gb of HDD goodness it too appealing and I don't have any other computer to put this drive into. The noise is also very directional so it varies in loudness depending on which way I turn my head or how I shift in the chair, arghhhhhhh...


Conclusions
Overall, I feel that the heatsink performed quite good. I'm satisfied with its performance in 2D mode. The temperatures in 3D mode are a little high, but I can live with them and it's not unexpected from a totally passive heatsink coupled with one of the hottest videocards ever produced. There are a couple more things I can probably do (see below), see if it helps temperatures, in the worst case I can always turn up the fans a bit while gaming.

VM-101 is a good heatsink and provided there is a decent airflow in the case it will work fine. However, it does have its limitations. One is you gotta have a decent airflow in the system. If you are one of the guys with ultra quiet computers with Nexus fans running at 5V, then depending on the kind of videocard you have in your system, this heatsink MAY NOT WORK FOR YOU. It must have at least some airflow over it, 5 volted Nexus is not going to be enough. Overall (if assuming 6800ULTRA dissapates ~75W heat under load) then VM-101 can probably handle 40-50W with minimal airflow in the case, if your card is hotter than that, then it may not work. Other thing is that you gotta be prepared and expecting slightly higher temperatures than from a fanned solution.

One side effect I noticed, is that my CPU temperatures rose by 2-3 degrees. I can't say it's totally unreasonable, since the GPU heatsink is now directly below the SCHJ-1000, but I expected that the airflow from the CPU fan is going to be more than enough to carry heat away from CPU to the rear exhaust fan.

Another thing that remains to see is how VM-101 performs at higher ambient temperatures. It's winter right now and the temperatures stay at 23C inside, but we don't have central airconditioning, so the ambient temperature may rise up to 28C in the summer. I really hope VM-101 will be good enough to handle summer.


Possible upgrades to improve performance
1) As I said before there is some residual airflow coming from the each sode of the CPU fan. The thing is however, right now the CPU fan is mounted higher than the center of the heatsink, that's just the way I mounted it. I could always move is slightly down, see if that helps temperatures a bit. I'm kind of reluctact to do it because I would need to take out the video card again and the AGP slot on my motherboard is a little touchy. And it is also possible that I would have to remount CPU heatink again since it's quite hard to mount/dismount fan when heatsink is mounted already and it rocked in place quite noticably when I originally mounted the fan.
2) Another thing is to try to duct the air being pulled in the case through the open slot cover. Right now most likely it's being pulled over the edge of the card straight into the exhaust fan. I could probably make a little duct over the slot cover to ensure that the air is pulled over the GPU block/memory for slightly better temperatures.


Pictures
I don't have a direct host so I have to use bbzzdd. However it doesn't allow direct linking so you'll have to use copy&paste trick.
GPU block mounted
Completely assembled, GPU side
Completely assembled, Heatsink side
Profile, AGP bus side
Profile, top side
Installed in case 01
Installed in case 02

unable to see your jpegs "unauthorized referrals" says the server.

"Like I said, you have to use copy&paste trick. Open new browser and manually copy link into address bar. lol"

Gees i am so dumb... no wonder my wife doesnt respect me , won't even pick
my socks up.

Like I said, you have to use copy&paste trick. Open new browser and manually copy link into address bar. lol

Nice review there Jack!
I don't remember that it came with 4 mounting arms, I must sank them somewhere
Have you tried replace that "Data Structures and Algorithms in Java" with that fan on Ninja?(I suppose it would be 'noisier' but might lower temp on summer time)

Thanks.
The heatsink came with 4 arms, but only two screws, so you could only use two arms. Not really a problem though as you can always buy two extra screws/nuts at your local hardware store.
As for "Data Structures and Algorithms in Java", well, that's the best use I could find for it, being a college textbook it's pretty useless otherwise. Makes one hell of expensive top cover though, at $83 used it cost me almost as much as the case itself
On a more serious note, I don't want to use top fan for several reasons. One is that yes, there is some noise leakage, not as much from the fans as from the hard drives - the idle noise is loud enough to leak into the main chamber and out of the top vent. Two is that the top fan cannot be softmounted and three is because I like my two fan setup, with the modifications I've just made (read below) it's pretty good, taking the CPU fan off would be detremental for the video card temperatures and leaving it in would make three fans total in main chamber alone, and I wouldn't want that.



UPDATE
I did both of the proposed mods I've described in my original post. Namely the GPU duct and softmounting a fan in a way that it would blow across VM-101 too. I got the idea on CPU fan suspension from this thread and it worked very well. Thanks to the Ninja there is absolutely no increase in the CPU temperatures, however the videocard temperatures have come down quite a bit.


Pictures
To see the pictures you have to do copy&paste trick. Copy the link into the memory, open new webbrowser, and copy link into address bar.
Overall view
Close-up


Temperatures
All test conditions were the same. Exhaust fan rotated at 1000RPM and CPU fan at 900RPM. Ambient temperature was 23-24 degrees celsius.

Once again I must feel I must comment on the temperatures. I've read of people getting low 40's on the 6800GT GPU with zalman VF-700 at 5V. I've never got such temperatures, not even close. The lowest I've ever had was about 54-55C with ACS5 at full 12V. It's possible that my heatsinks never had good contact with GPU, it's also possible that there is a difference in thermal probe calibration. So I cannot directly compare effectiveness of VM-101 against VF-700 which I've never had, but I can comment on its effectiveness against stock heatsink and ACS5.

2D mode 200Mhz
56C

2D mode 400Mhz
56-57C

As you can see simply building a GPU duct and mounting CPU fan differently so that it blows over VM-101 heatsink as well provided a pretty significant difference of 4-5 degrees celsius. In 2D mode it's almost as good as ACS5, maybe even slightly better (because ACS5 could go as high as 59C) and all without any additional noise. Very nice, I'm totally satisfied with 2D temperatures.

As you can once again see there is very little difference, if any between 200 and 400Mhz. Strange, what can I say, but I will repeat myself again, these results are consistent with what I got before.

3D mode 400Mhz
This is where it gets a little strange. First, as I have discovered, VM-101 is even more inertive as I initially have thought. And second, it appears that my previous results of 80C max in the first post may have been incorrect. In that test I've ran the game at com_maxfps 76 which apparently forced the card to skip frames and not work at full power. In any case....

Like I said the heatsink is extremely intertive. I've ran the game for 45 minutes and it sort of stabilized at 65C, I almost thought that was it, that was the max temperature, but then it began to rise again, this time stabilizing at roughly 68-71C. At this point I've actually discovered that com_maxfps actually affects the temperature and set it to 400fps which brought the max temperature up to 74-75C. I continued like this for 30 minutes, so it would appear that this is the max temperature with the new setup. Still a pretty good temperature drop at the same noise level.

Just out of curiosity I felt the air coming out from my exhaust fan and it seemed a bit too warm for me so I raised the rear fan up to 1200RPM and the temperature came down to 67-70C. I'm not sure how to interpret results, but it would appear to me that the combined heat from the videocard and CPU was too much to handle for a 1000RPM YL so there was some heat build up. As soon as I raised the RPM on exhaust fan, subjectively the exhaust air temperature went down and the GPU core also went down, as I said to 67-70C which seems acceptable to me seeing as it's still on the same level or even lower than stock heatsink.

So for now it would seem that I would have to turn up rear fan RPMs while gaming, but that's totally fine by me. The noise increase is not that much and it doesn't bother me while gaming while keeping the video card pretty cool. I will probably do another stress test with Far Cry, but for now I'd say that the experiment is a total success - fairly good temperatures and no additional noise over what my system already produces. I may have to remove "well, mostly" from the thread title

How about CPU temp with that new attempt?
Is the duct really make any difference compared to new fan direction?

Like I said, CPU temps didn't change at all, 38-40C during gaming.


Don't know, it was my last day before going back to school so I didn't have time to test how much difference they make separately. But I assure you that there is a fairly strong draft around the slot cover in the back, probably because front intake in P180 is somewhat restrictive so the air is drawn through the empty slot cover. In any case, even if it doesn't help GPU temperatures it still must be good for the memory and overall board temps. 6800GT is so hot that even the metal reinforcement bar gets hot to touch, not burning, but very hot.

Theoretically, the effect of the GPU duct would have probably been bigger if I had VM-102 which has a big heatsink on the GPU side. However, I don't know if it would have been a better cooler overall compared to VM-101 in my PC because VM-102 has several downsides compared to VM-101. First, the GPU block is much thicker, second, the fins on the heatsink block are thinner and lastly, the fins on the GPU heatsink are perpendicular to the slot cover its not as effective as it could have been.