Aerocase Raven passive GPU cooler v.3...

Closely moderated reviews by forum members willing to share their experience and accept stiff peer review. Open only to registered forum members.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar

Post Reply
mr. poopyhead
Patron of SPCR
Posts: 376
Joined: Thu Jun 29, 2006 8:37 pm
Location: Mississauga, ON
Contact:

Aerocase Raven passive GPU cooler v.3...

Post by mr. poopyhead » Tue Sep 12, 2006 9:16 pm

v.2 EDIT: now includes photos and analysis.... if you don't care for my rambling, please skip to the results section, :P

v.3 RETEST: adjusted space between raven and CPU heatsink as suggested by stealthgirl. added photo of new setup. sorry for the crappy cell phone photo. my parents took my camera on a trip.

intro:
last month, SPCR ran an article about the condor cooler from aerocase... the article can be found here: http://www.silentpcreview.com/article650-page1.html

but of course not all of us have the latest and greatest GFX cards. i happen to be among those in the aforementioned group. i own a BFG 6800GS and i didn't see a need to have such a huge cooler (did you see the torque it was putting on the card? [shudders...]). but i did want a passive cooler to replace the stock cooler i had, which was only on there because my AC NV5 silencer decided to call it quits after a month. so i decided to give the raven a try. despite a lack of reviews or any kind of feedback from anyone, i went ahead and bought the raven anyways... so here i am. sort of a raven guinea pig. hopefully what i post here will be useful to someone. please don't expect a professional SPCR review. i lack any kind of lab equipment. all i have is my eyes, ears, hands and any other appendage that might come in useful. on with the review...

ordering:
ordering the raven was a breeze. paid through paypal. no problems there... their ordering system could use a little work. i would suggest confirmation emails and maybe some kind of status updates from their end. i did have some complaints about their shipping but they have been resolved and need no further attention in this review. all i can say is that their customer service is fantastic and quickly turned my complaints into praise.

the raven arrived today, neatly packaged in a box filled with bubble wrap, and foam peanuts. all the little accessories came in a plastic baggie (yup... those famous toothpicks too, :P) along with my invoice and an instruction booklet. aerocase also provides one extra screw and one extra rubber/brass spacer just in case you lose one. good idea... more on that later.

installation:
when i got home i took out the heatsink and got to work. one thing i noticed right off the bat was that the heatpipes weren't bent 90 degrees like the diagram. i knew this would be trouble right away. more on that later... the instructions seemed pretty straightforward but i did manage to run into a few snags.

first of all, the toothpick method wasn't as easy as it looks in the instructions, or the SPCR condor review. i'd suggest everyone go grab 2 extra toothpicks and put them through each hole because i found that the 2 non-toothpicked holes kept getting misaligned from the rubber pads. (sorry for the lack of photos. just look at the condor review. it's pretty much the same deal...)

also, when they suggest that you can reuse the stock backplate, DON'T. just use the little brass rings they give you. my stock backplate wasn't agreeable with the screws aerocase provided. i had a HORRIBLE time trying to get them in. i threaded one of the screws and another one actually got stuck. when i tried to force it through using brute force i actually BROKE the head of the screw... stupidity on my part, but aerocase may want to look into stronger screws? also, maybe they should provide more than one bonus screw for idiots like me...it's not hard to imagine losing or breaking more than one screw. i messed up 2 of them and i started to panic... after i got the headless screw out (a mentally and physically taxing process that i will never speak of again...) i abandoned the stock backplate and went with the brass spacers. easy as pie...

now comes the screwing in part. the instructions tell you not to over-tighten. but it also doesn't tell you when to stop. you kind of have to eyeball it (unless i missed something in the instructions). i just kept screwing until i could feel some reistance.... DO NOT DO THIS! when i looked at my board, it was bending.... i guess the advice i can give is to just watch it as you screw it in... look for solid contact between the GPU and heatsink. don't rely on the rubber pads as an indicator. your board may start to bend and creak before the rubber even starts to compress...

as noted earlier, the heatpipes weren't bent at a 90 degree angle. it was probably something like 70 degrees or something. i knew right away that this would mean that my XP-90 heatsink and fan would get in the way of the raven. sure enough when i tried to put the card in the AGP slot it did encroach on my CPU heatsink space. it took a little time before i mustered up the courage to bend the heatpipes myself. i was so scared of breaking them. i hope aerocase makes sure that future builds are consistent with the specs the provide on the site.

i put some ramsinks on my ram even though the official SPCR condor review says that this is not necessary. although i have no temperature measure i must say that the memory gets REALLY hot. the only quantifiable measure i can give you is to say that i can only touch the ramsinks for less than a 1/2 second before my finger gets burned. so while the RAM debate goes on somewhere on these forums, i feel safer with the ramsinks on...

ok. so finally the cooler is on and the installation nightmare is over. i've got multiple cuts all over my hands now and bleeding all over... what a nightmare. at this point i'd like to make clear that most of my installation mishaps were due to my own stupidity. it should have been a breeze to install although there were a few things i thought they could fix as noted above.

cooler is on. let's take some performance measurements...

test set up:
my computer
--AMD athlon64 3000+
--thermalright XP-90 & nexus 92mm fan
--enermax noisetaker 420W PSU
--antec P180 case
rthdribl 1.2-stress GPU (also pushes CPU load to 100%)
rivatuner 2.0-monitor GPU temps
speedfan 4.28-monitor CPU, NB temp
stock cooler mounted on BFG GeForce 6800GS OC 256MB DDR3 (unlocked to 16pp,6vp)
raven passive GPU cooler mounted on BFG GeForce 6800GS OC 256MB DDR3 (unlocked to 16pp,6vp)

ambient temperature in my room is approximately 25 celsius

photos of whole setup:
Image
one case fan has to clear out all the heat from the CPU and GPU now. the raven also seems to block the nexus fan...

Image Image
LEFT: original test setup. not too much room for the heatsink. if you have a tower cooler, then you can order the wings reversed so that they cover the PCI slots instead.
RIGHT: retest setup. note the extra clearance after adjusting the raven as suggested by stealthgirl.

method:
run rthdribl.exe until temp remains constant for 5 mins. take temps for GPU, CPU, NB, and ambient (some sensor on the GFX card).
i'm including CPU and NB temps so that you can get an idea of how the passive heatsink affects other components. all values are in celsius...

results:
stock cooler idle/load
GPU-61/88
CPU-44/55
NB-39/44
Am-46/67

raven idle/load
GPU-63/96
CPU-45/60
NB-40/49*
Am-52/83

* after the raven was installed, i had enough room again to get rid of the noisy NB fan and put on a zalman NBJ-47

raven idle/load - RETEST**
GPU-68/~100
CPU-47/~58
NB-43/~47
Am-56/~84

**the retest was stopped when the GPU hit 100 degrees since it was clear that the new configuration was detrimental to performance.

analysis:
at idle, the raven is virtually on par with the stock cooler but obviously MUCH quieter. i even had the added bonus of freeing up some real room for my zalman NBJ-47 chipset heatsink. so really, i got rid of two fans with the raven.

under load the GPU temperatures really soar. surprisingly, given the close proxmity of the raven to the XP-90, the CPU temperatures weren't as high as i expected them to be. overall, the load temps are pretty high leaving no room for anyone with overclocking aspirations (unless you have really excellent air flow in your case).

RETEST
it's almost counter-intuitive but the extra space given to the raven caused a DECREASE in performance both in the idle and load states. my best guess is that the raven is now too far from the airflow path created by the rear case fan and too close to the side panel of my case. hopefully aero-case can shed some light on this. i'm guessing the optimal position is somewhere between the original spacing and the new spacing.

two variables that may be inconsistent with the original test are the ambient temps in my room and the break in time of the thermal paste. my room SHOULD be the same temp as the original but the weather outside is different this week. thermostat still says 25. the other error factor i can think of is the break in of the thermal paste. my original test was done within a few hours of installing the raven. basically enough time for my computer to return to normal internal temps. this test was conducted over week later giving time for the thermal paste to settle. in theory this should increase performance which means the retest results might've been a little worse had i conducted it immediately after the original.

conclusion:
i think 96 degrees is a little high for my tastes. the chip doesn't throttle down til 120, but i never want to get that close. however, i think that it's a reasonable temperature to endure for complete silence.

i've used 3 different coolers on this card now. NV5 silencer, stock and now the raven. performance wise, the NV5 silencer was defintitely the best. as i recall my load temps were somwhere in the 70s with the silencer. however the raven provides the benefit of a completely silent solution and if you're reading the SPCR forums then i'm assuming that will be a huge selling point. if you have anything lower than a 6800GT, then the raven should be pretty safe and you probably won't be pushing the thermal limits of your components like i am. personally i am quite happy with this cooler. the silence part of it anyways. not too wild about running my card at almost 100 degrees but sacrifices must be made right?

with a little tweaking of my system airflow, i'm certain i can probably lower my temps by a few degrees across the board.... perhaps i will try using a different thermal paste. the raven came with some white stuff on it and an unmarked tube of extra thermal paste. maybe some arctic silver 5 may drop my temps a shade. but i probably won't try it cause the screws on the raven are nearly threaded and i don't want to risk taking it off for any reason. maybe stealth girl can send me some extra screws for my experiment? :P

i'm writing this really late. i bet i missed a whole bunch of stuff. if you have questions let me know and i'll try to fill in the gaps at work tomorrow...
Last edited by mr. poopyhead on Wed Sep 20, 2006 9:48 pm, edited 5 times in total.

StealthGirl
-- Vendor --
Posts: 77
Joined: Fri Jul 23, 2004 8:59 am
Location: Marin, CA.
Contact:

Post by StealthGirl » Fri Sep 15, 2006 10:59 am

Sorry, for the delay in responding to this, I just saw your review this morning. First I’d like to say thank you for all of your work in posting this ... it takes a lot of effort to write this much detail and post these great pictures.

Now, here are some of my thoughts and comments as both SPCR member and Aerocase Product Manager:

With a Wing that is at 70°, it sounds like it may have been squished in shipping (maybe even by our shipping folks … so we’re checking if it could have been the way the box was packed). All of our heatpipes are bent in a jig so we know it could not have been manufactured that way.

As for the heatsink installation, yep, we know we have some occasional problems … which we are actively working on. With so many variations between manufacturers of even the same GPU, making a heatsink that is “supposed to fitâ€

mr. poopyhead
Patron of SPCR
Posts: 376
Joined: Thu Jun 29, 2006 8:37 pm
Location: Mississauga, ON
Contact:

Post by mr. poopyhead » Fri Sep 15, 2006 11:19 am

StealthGirl wrote: Now, for your high temps, our engineers believe the Raven heatsink wing is too close to the CPU fan. Not only is it blocking the CPU airflow, but also the Raven’s. If you can just bend the heatpipes up about 10° more to give the heatsink about an inch of breathing room you will see a temp drop on both the GPU and the CPU. The thermal paste is Arctic Silver Ceramique, which is almost as good as the Silver 5, so there is no need to remount the heatsink if you think you have good contact.
is it safe to continue bending the heatpipes? i've already bent them through about 20 degrees... once i get the green light, i will bend the wing back a little more and measure my temps again. thanks for the tip.

NeilBlanchard
Moderator
Posts: 7681
Joined: Mon Dec 09, 2002 7:11 pm
Location: Maynard, MA, Eaarth
Contact:

Post by NeilBlanchard » Fri Sep 15, 2006 11:39 am

Hello,

What about reversing the CPU fan -- so that it blows on the Raven? It still is close, but the Thermalrights are supposed to work just as well that way; and it surely might cool the Raven better? The only downside might be an increase in the temps on the VRM components on the mobo...

StealthGirl
-- Vendor --
Posts: 77
Joined: Fri Jul 23, 2004 8:59 am
Location: Marin, CA.
Contact:

Post by StealthGirl » Fri Sep 15, 2006 1:08 pm

Yes, go ahead and bend the Wing up an inch. The heatpipes will be OK to bend that far. If there is a problem we'll replace the unit, so go for it.

And, no I don't think reversing the fan is the way to go. Leave it blowing down. The Raven was designed for natural convection so it works best if there is very little air distrubance over the Wing.

Good luck!

mr. poopyhead
Patron of SPCR
Posts: 376
Joined: Thu Jun 29, 2006 8:37 pm
Location: Mississauga, ON
Contact:

Post by mr. poopyhead » Tue Sep 19, 2006 10:58 pm

i moved the raven as suggested by stealthgirl and ran the test over again. for those who have already read the review and don't want to hunt for the new text, here it is below:
raven idle/load - RETEST**
GPU-68/~100
CPU-47/~58
NB-43/~47
Am-56/~84

**the retest was stopped when the GPU hit 100 degrees since it was clear that the new configuration was detrimental to performance.

analysis:

RETEST
it's almost counter-intuitive but the extra space given to the raven caused a DECREASE in performance both in the idle and load states. my best guess is that the raven is now too far from the airflow path created by the rear case fan and too close to the side panel of my case. hopefully aero-case can shed some light on this. i'm guessing the optimal position is somewhere between the original spacing and the new spacing.

two variables that may be inconsistent with the original test are the ambient temps in my room and the break in time of the thermal paste. my room SHOULD be the same temp as the original but the weather outside is different this week. thermostat still says 25. the other error factor i can think of is the break in of the thermal paste. my original test was done within a few hours of installing the raven. basically enough time for my computer to return to normal internal temps. this test was conducted over week later giving time for the thermal paste to settle. in theory this should increase performance which means the retest results might've been a little worse had i conducted it immediately after the original.

JazzJackRabbit
Posts: 1386
Joined: Fri Jun 18, 2004 6:53 pm

Post by JazzJackRabbit » Wed Sep 20, 2006 9:23 am

Part of the problem is, I think, that even if the part is optimized for low airflow, high airflow never hurts it. You may not gain much using high airflow, but you definitely not going to hurt performance. Unfortunately opposite is not true. You decrease airflow and even if the part is optimized for low airflow you are going to see increased temperatures for sure. Perhaps raven is not really optimized for low airflow as much as you think it is, so in retrospect it was probably best not to bend the heatpipes.

vg30et
Posts: 105
Joined: Tue Aug 22, 2006 5:14 am

Post by vg30et » Mon Dec 11, 2006 8:07 am

I've had the Raven (reverse wing) for several months now and thought I'd post my experiences with it. I'm surprised there haven't been more posts about this cooler after the SPCR review and special offer stickied in the forums.

Anyhow, the Raven was originally intended for my 9800SE but is now doing a good job cooling my X850XT PE. I know the Condor is better suited for the X850XT PE but I am pleasantly surprised that the Raven is holding its own.

System specs: AMD AM2 3200+ w/ Scythe Infinity, ATI X850XT PE w/ reverse wing Raven, P180B case w/ Yate Loons controlled by Zalman fan controller (5V to ~10V). HDD enclosed in suspended Scythe Quiet drive. System basically inaudible when fans at 5V except for low whooshing noise from Corsair power supply.

At 25C ambient room temp, temps are:

CPU 33C idle, 42C load
GFX 45C idle, 70C load

This is with a Yate Loon placed in the upper HDD bay in the P180b which blows directly at the reverse wing Raven. I run it at 12V whenever at load so it is audible but definitely not over my speakers.

I was previously using a Arctic Cooling Silencer which performed better - I believe temps were at least 10C lower. However, my "Silencer" was making clicking noises even at the lowest fan setting and was very audible at max speeds.

Post Reply