I love experiments. Even if only fin spacing!

Cooling Processors quietly

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

I love experiments. Even if only fin spacing!

Post by Felger Carbon » Mon Feb 12, 2007 11:19 am

There is much commentary to be found in the SPCR reviews (such as the recent SI-128) and in the forums that wider fin spacing is best for low airflow cooling. Certainly the Ninja seems to prove that... but the Ninja also has 110mm-deep fins!

Suppose we were to compare a standard Freezer 64 Pro (passively or actively cooled at low RPM) with its standard ~1.5mm air gap with one that had a 3mm air gap, but half as many cooling fins? No other change to the HS/HSF? What would be your prediction as to the relative performance?

We're gonna find out, so place yer bets now! :D

3mm is .118". The heatpipes on the 64 Pro are spaced in a way that makes it relatively easy to bend the cooling fins apart. I just placed an order with McMaster for 4ea 1394A16, which is 6" by 1+" by .110", and 2ea 19155A38, which is almost exactly as long as the 64 Pro fins are deep, and is .118" thick!

I'll use my dremel-alike with cutting wheels to taper the "entrance edge" of the above, so I can hammer the steel in between the fins, bending them apart. Start with the .110" and finish with the .118".

Then, test an unmodified and a modified 64 Pro and record the results.

Anybody brave enough to go on the record in advance? Sure, wide spacing is better... but with half as many cooling surfaces?

You know, if this works out, my Scythe Andy Samurai is next, if I can locate the correct thickness steel. That thang has 'way too many fins, too closely spaced. :(

Howard
Posts: 275
Joined: Wed Feb 08, 2006 5:33 pm

Post by Howard » Mon Feb 12, 2007 11:36 am

What fan and fan speed?

EndoSteel
Posts: 308
Joined: Sun Jan 25, 2004 7:50 am
Location: Moscow, Russia
Contact:

Post by EndoSteel » Mon Feb 12, 2007 1:15 pm

I've already conducted the same experiment on an Igloo 7700 but my goal was to achieve fanless operation. The best variant to cool a 3000+ Newcastle was: 16 fins spaced 4mm apart. Originally the heatsink had 45 \ 1.5mm.

J. Sparrow
Posts: 414
Joined: Wed Jan 17, 2007 7:55 am
Location: EU

Post by J. Sparrow » Mon Feb 12, 2007 1:18 pm

My expectation is that performance could drop about 30% in terms of heat dissipated at full speed (using the stock fan, of course)

The reasoning behind that is: it will have less exchange surface but air should flow more easily on the full fin depth (while now it exhausts more air on the sides)

If you're looking for more modding ideas for the Freezer, I'd really like to see one with an Infinity-like fin setup :D

Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Mon Feb 12, 2007 1:37 pm

J. Sparrow wrote:My expectation is that performance could drop about 30% in terms of heat dissipated at full speed (using the stock fan, of course).
I agree that any HSF will perform better, with stock fan, at the stock fin #s and fin spacing. Otherwise the design engineer(s) should be fired for incompetency.

My experiment is only to determine what the performance will be under low airflow conditions, i.e. passive cooling. I may not use a fan at all, since I'm completely uninterested in high airflow, and I want to minimize the number of variables.

J. Sparrow
Posts: 414
Joined: Wed Jan 17, 2007 7:55 am
Location: EU

Post by J. Sparrow » Mon Feb 12, 2007 2:51 pm

Too bad :( I've probably skimmed through the post too fast.

However, your experiment might be interesting even with the fan at full speed because of the noise reduction. Maybe you can find a sweet spot which is little noisier than low airflow while being cooler.

I don't think you could go passive with that heatsink, unless your thermal load is really low.

McBanjo
Posts: 671
Joined: Thu Dec 08, 2005 1:40 pm
Location: Sweden
Contact:

Post by McBanjo » Mon Feb 12, 2007 3:46 pm

Have you read my modding of the freezer 4?
viewtopic.php?t=31327&start

Might answer some of your thoughts

Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Mon Feb 12, 2007 8:44 pm

McBanjo, your photos have expired, so I can't see what you did. Endo performed an interesting experiment, but one which presented the broad side of the HS to the PSU and not the exhaust fan.

The Freezer 64 Pro permits mounting in all 90 degree directions, and comes standard with the 104mm width presented broad-side to the exhaust fan. And, I'm gonna duct the HS to the exhaust hole, but not increase the width of the HS beyond the std 104mm. So my experiment is fundamentally different from Endo's... and also from yours?

I only ordered 3 line items from McMaster this morning, so there's hope I'll receive the stuff by this Fri or even Thur.

Also, I plan to offer the delta temp (die - room ambient) before and after, at several different airflows - as determined by the voltage applied to the big side fan. Neither you nor Endo did this.

However, I do appreciate you calling my attention to the previous effort. As you may know, I very recently took a Sonic Tower partially apart and modified it very similar to Endo's effort, and posted pics of that partial effort. So I'm very respectful of anybody who completely dissembled a HS's fin structure! :D

McBanjo
Posts: 671
Joined: Thu Dec 08, 2005 1:40 pm
Location: Sweden
Contact:

Post by McBanjo » Tue Feb 13, 2007 3:11 am

Oops, I removed the pics from my site and forgot about it but yes it was more or less like Endos project

You might try to have 1 fin at 1 heatpipe and 1 fin at 2 heatpipes? Making it longer rather than wide and thereby keep the area while getting twice the amount of space between the fins?

Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Tue Feb 13, 2007 4:11 am

McBanjo wrote:You might try to have 1 fin at 1 heatpipe and 1 fin at 2 heatpipes? Making it longer rather than wide and thereby keep the area while getting twice the amount of space between the fins?
If I were planning on completely dissembling the HS's fins, yes, that would be an interesting experiment. However, looking down on the HS from the top, I don't plan to change the outline of the fins, meaning I don't have to disassemble the fins. Just use brute force to change their spacing, alternately.

I'm planning a different experiment than you or Endo performed. No claim of "better", just different. :wink:

McBanjo
Posts: 671
Joined: Thu Dec 08, 2005 1:40 pm
Location: Sweden
Contact:

Post by McBanjo » Tue Feb 13, 2007 12:11 pm

Felger Carbon wrote:If I were planning on completely dissembling the HS's fins, yes, that would be an interesting experiment. However, looking down on the HS from the top, I don't plan to change the outline of the fins, meaning I don't have to disassemble the fins. Just use brute force to change their spacing, alternately.

I'm planning a different experiment than you or Endo performed. No claim of "better", just different. :wink:
Hmm, using brute force might not be such a good idea since it might alter the shapes a bit too much. Removing 50% of the fins will most likely be a bether result.
Brute force will also result in 50% of the fin-space is VERY tight and might block the air and create turbulence

Just some thoughts. Not based on much fact :wink:
Just as you I enjoy experiments so I hope you do this anyway :)

Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Thu Feb 15, 2007 3:02 am

This experiment just got cancelled on account of sunny weather. It turns out that an HSF that's never been reviewed and yet is available at Newegg, HeatSinkFactory, SVC, and other places has 58 fins on a 134x137mm grid (fins 40mm deep). That's either 2.35 or 2.4mm fin spacing (57 spaces between the 58 fins) without munging the fin spacing, and is a more interesting candidate than the Freezer 64 Pro if I do want to munge - there's nearly a 4mm air gap available that way, within ~10% of Ninja's air gap!

So I don't know if the munging experiment is cancelled or just postponed. I do know I should get my Dominator from Newegg by the middle of next week because it's already on order. :)

McBanjo
Posts: 671
Joined: Thu Dec 08, 2005 1:40 pm
Location: Sweden
Contact:

Post by McBanjo » Thu Feb 15, 2007 6:14 am

Got an url?

oso
Posts: 38
Joined: Thu Feb 15, 2007 8:05 am

Post by oso » Thu Feb 15, 2007 8:07 am

I think he's talking about this one:
http://www.newegg.com/Product/Product.a ... 6835129048

Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Thu Feb 15, 2007 8:08 am

McBanjo wrote:Got an url?
To get a url, I'd have to log onto Newegg and search "aerocool dominator". Since you want the url, why don't you do it instead? :D

McBanjo
Posts: 671
Joined: Thu Dec 08, 2005 1:40 pm
Location: Sweden
Contact:

Post by McBanjo » Thu Feb 15, 2007 10:27 am

Felger Carbon wrote:
McBanjo wrote:Got an url?
To get a url, I'd have to log onto Newegg and search "aerocool dominator". Since you want the url, why don't you do it instead? :D
You didn't post the name otherwise I could :wink: or atleast I thought you didn't, I assume that Dominator was refering to Corsairs memory :oops:

Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Thu Feb 15, 2007 11:52 am

McBanjo wrote:You didn't post the name otherwise I could :wink: or atleast I thought you didn't, I assume that Dominator was refering to Corsairs memory :oops:
You're right! Sorry, I wuz up late last night chasing this fin spacing thing. I think I found a winner spacingwise but I have no idea how the HSF as a whole works. I'll find out, though. No reviews, not even in foreign languages? :?:

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Thu Feb 15, 2007 12:04 pm

Looks similar to the thermalright SI-128,maybe slightly better performance due to bigger fan?

J. Sparrow
Posts: 414
Joined: Wed Jan 17, 2007 7:55 am
Location: EU

Post by J. Sparrow » Thu Feb 15, 2007 12:32 pm

Is there any good (quiet) 140 mm fan, anyway?

McBanjo
Posts: 671
Joined: Thu Dec 08, 2005 1:40 pm
Location: Sweden
Contact:

Post by McBanjo » Thu Feb 15, 2007 12:59 pm

Felger Carbon wrote:No reviews, not even in foreign languages? :?:
Who's your daddy? :P
http://www.pro-clockers.com/reviews.php?id=129&page=1

Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Thu Feb 15, 2007 2:20 pm

Neat! Looks like a decent performer, just beaten by the TT Big Typhoon as far as cooling is concerned, with a very hot CPU, overvolted and overclocked, being tested. I just want to cool a Sempron64 2800+ or 3400+, which are the two varieties I have. Thanks for the url!

About the fan: I bought one of the 140mm x 20mmm fans a while back, never had reason to turn it on. I just did. Specs are 73CFM and 31.72dBA at 1500RPM. That much CFM and RPM on a 13-blade fan doesn't say "silent", and at 12V it isn't.

My copy starts at 5.0V when you slowly increase its voltage with the Sunbeam Rheostat, which is worst-case. Like most sleeve-bearing fans, it will run below its starting voltage once started. Using my free desk-lamp stroboscope (the lamp wasn't free, but its strobe capability is: it comes with the flourescent bulb) I got the following RPM/voltage points:

8.69V = 1088RPM (2 blades per light pulse)
5.82V = 816RPM (1.5 blades per light pulse)
3.69V = 544RPM (1 blade per light pulse)
at 60Hz, I get 7200 light pulses per minute. 13 blades, remember!

at 5.5V, I can just barely hear the fan when held a foot away. In my computer, it will be just over 2 feet away and on the other side of a 1/8" rubber sound barrier.

There is nothing to say that smaller fans - the TT 13cm Silent Wheel, the GW NCB 120mm - cannot be used. Many XP-90s, including mine, were run using 80mm fans.

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Thu Feb 15, 2007 4:01 pm

J. Sparrow wrote:Is there any good (quiet) 140 mm fan, anyway?
Yate Loon D14SL/M-12

Evil Gnomes
Posts: 24
Joined: Tue Nov 01, 2005 12:40 am
Location: Sydney, Australia

Post by Evil Gnomes » Sat Feb 17, 2007 3:31 pm


Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Sat Feb 17, 2007 5:25 pm

Thanks for that reference. The reason I decided to try the Dominator:

It has the second widest fin spacing of all the current HSFs I know of, and the widest that will fit in my available 133mm. Wide spacing works best with low air flows. I'll be using a low power Sempron64 CPU. And I'll be presenting the cooling fins with a 2-stage fan, not one-stage as with the Madshrimps review. That means I can get about the same fan pressure at roughly 60% RPM, resulting in a 10dBA noise advantage over a 1-stage fan. So, that's 3 things that favor the Dominator in my PC case.

In fact, if I can also fit a 20mm-thick fan to the bottom of the HSF as well as the top, I'll have a 3-stage cooling fan system on the HSF. I may replace the top fan with a 14cm Yate Loon; that will fit with 8mm to spare. :)

And if you haven't already concluded that I'm totally insane :twisted: , I'm actively constructing a 4-stage cooling fan system around my new big-fan case and the Freezer 64 Pro. In addition to the 220mm fan providing positive pressure for the PC case, I'll have 3ea Scythe 100mm 1000RPM nominal fans, one in front of the Freezer and 2 behind, all in a duct.

This is how the experiment has morphed: I will then be able to test the Freezer's cooling at different voltages for 1, 2, 3, and 4-stage fan systems. I've already done this for 1 and 2 stages. I do like experiments! :D

Evil Gnomes
Posts: 24
Joined: Tue Nov 01, 2005 12:40 am
Location: Sydney, Australia

Post by Evil Gnomes » Sun Feb 18, 2007 11:59 pm

2 stage fan?

Stop confuzzling me! :P

Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Mon Feb 19, 2007 12:43 am

Evil Gnomes wrote:2 stage fan?
2 stages, meaning 2 fans in series. Same CFM as one fan, but twice the pressure. Good for HSFs with closely spaced cooling fins.

Also, my engineering analysis sez two fans running at ~2/3 speed can cool better, and more quietly, than one fan running full speed. I need to do more experiments to gather data to prove (or disprove!) the point, thus the 3/4 stage experiment coming up.

spookmineer
Patron of SPCR
Posts: 749
Joined: Sat Nov 11, 2006 6:02 pm

Post by spookmineer » Mon Feb 19, 2007 2:23 pm

The two fans need to be some space apart from eachother (maybe one pull, one push with the heatsink inbetween - in case of a tower heatsink).

I read somewhere, you can put 2 fans parallel to eachother (placed on top of eachother) to get bigger pressure, but for best effect, they both need to turn another direction.
It has something to do with the vector of the airflow coming from the fan: air leaving the fan has an axial speed, but also a "rotating" aspect. The second fan will reinforce the rotating aspect even more, and not increase the axial speed/pressure as much. If the second fan turns the opposite direction of the first fan, the energy is not put into the rotation but into the axial speed/pressure (which is wanted).

Most fans turn the same direction though...

Felger Carbon
Posts: 2049
Joined: Thu Dec 15, 2005 11:06 am
Location: Klamath Falls, OR

Post by Felger Carbon » Mon Feb 19, 2007 5:42 pm

Good point. I'll have 1 fan on one side of the Freezer fin stack and 2 on the other, separated by ~.1". So the third fan may not provide good results - but that's what an experiment's for: to confirm or refute theories. :D

Post Reply