Gentle Typhoon 4250RPM/116.5CFM

[WoodButcher]

Cone in the center w/ bondo on the struts to smooth the airflow?
ROFL,,,,
You're twisted Duck, check-up from the neck-up for you!

You know, ducting for noise never crossed my mind. Rubber screws or double stick foam tape was as far as I ever bothered with. As I alluded to earlier I tried the AP30's but needed more quiet. Right now that build has yates med or lows I don't recall offhand but even those I have on a DPDT switch for 12 or 5 volts.
Do you do anything with the square or rectangle ducts to smooth the inside 90*? a chamfer or cant strip to make it octagonal? would it make any difference?

 

[BonzaiDuck]

Cone in the center w/ bondo on the struts to smooth the airflow?
ROFL,,,,
You're twisted Duck, check-up from the neck-up for you!

You know, ducting for noise never crossed my mind. Rubber screws or double stick foam tape was as far as I ever bothered with. As I alluded to earlier I tried the AP30's but needed more quiet. Right now that build has yates med or lows I don't recall offhand but even those I have on a DPDT switch for 12 or 5 volts.
Do you do anything with the square or rectangle ducts to smooth the inside 90*? a chamfer or cant strip to make it octagonal? would it make any difference?

Gee -- per your last para, that goes in the direction of aerodynamics or some such discipline. All I thought about was air pressure in chambers. There was a Wiki article on "ducted fans" which had a purely aeronautical focus: round donut shaped ducts around propellers. In that type of example, changes in the angle of the duct have noticeable affect on lift. But for this, I think all you'd be concerned about is the displacement of air within any confined volume, and how that might facilitate entry of more air.

The interior shape of the duct might affect turbulence, though, and therefore modestly affect the level of white noise. But with my idea, a fan with high static pressure and potentially high airflow that is smaller than the heatsink fins that the fan exhausts might see that pressure and airflow translated into the volume of air moved within the larger box, which in turn just fits the back of the cooler fins. If it doesn't cover the cooler fins entirely, it will at least draw air through the uncovered portion and it will do so more effectively the closer the fit of the duct to the edges of the fins.

So one might nevertheless taper the interior of the box in the corners to make the airflow more uniform. How much work you'd want to put into this may have a limit in terms of what is practical versus any benefits and time spent.

Put it another way: I occasionally expend a great deal of personal human wattage doing these things, when the greater temperature improvements might come easier with the initial trouble of installing a water-cooling loop.

But taken the other way around, since fans are integral to both water-cooling and air-cooled rigs, and since there are quite a number of fans required for - say- two 3x120 radiators, any benefits to these "art-class construction-paper" projects may cross the boundaries between one type of cooling and the other.

 

[WoodButcher]

I see your point, effort vs. coin, one way or the other or these hobbies we dive into demand both, if we allow it too much of each.
A thought for those chamfers though, bondo. I should think if it were applied to the corners with a taper from almost nothing on the heatsink side (to allow the air contact w/ all the fins area) to a thickness at the fan end to match the diameter of the fan it would not only direct the air and eliminate dead zones but also add mass to the duct reducing the sound level even more. If you wanted to go whole hog you could lay up fiberglass to make he duct.

 

[BonzaiDuck]

I see your point, effort vs. coin, one way or the other or these hobbies we dive into demand both, if we allow it too much of each.
A thought for those chamfers though, bondo. I should think if it were applied to the corners with a taper from almost nothing on the heatsink side (to allow the air contact w/ all the fins area) to a thickness at the fan end to match the diameter of the fan it would not only direct the air and eliminate dead zones but also add mass to the duct reducing the sound level even more. If you wanted to go whole hog you could lay up fiberglass to make he duct.

Well, I'm still going to post my "journal" on this project, but I'll say up front that I'm pleased with the outcome.

I think the point where a plain G. Typhoon AP-30 would begin to affect noise tolerance of the most discriminating ears is about 2,100 RPM. The 20% duty-cycle for the fan is about 1,200 RPM. Over that range, the fan is virtually silent. After that, the noise is a combination of white-noise from air-turbulence and a "tone." The tone increases in frequency to the 4,250 RPM top-end, and also increases in volume. So adding one noise-abatement feature after another would be successful if you could raise that 2,100 threshold. I think I moved it up to above 3,200.

Put it another way: the 70% duty-cycle for this fan translates to 3,250 RPM. My notes show that it sounds like what you'd expect from any OEM computer, except for the white noise. There's nothing to do about the white noise, except to reduce airflow, which defeats the purpose of that particular fan. The tone is a different matter.

With all the noise-abatements in place -- the duct, Spire rubber on the duct exterior, thin vinyl foam on the inside, a 10" square of Spire with foam-board backing stuck to the right side-panel so there's no mess to remove, a foam "collar" around the fan to mate up with the duct, and even foam circles or "Bozo's noses" on selected fan hubs, the tone still seems excessive in the 3,600 to 4,250 RPM range of the fan. That last 500 RPM has no effect on temperatures. But the noise-tolerable speed-range for the fan has been increased from 1,200-2,100 to 1,200-3,400 or so.

Well, I'll take some more photos and finish writing it up.

The whole point of getting a fan like this is to provide the assurance that your air-cooler is functioning to its max potential when load-testing overclocks. Once you know how more airflow ceases to improve temperatures, you might set the fan's "curve" to a lower maximum speed, and that speed might only arise during heavy gaming or other intensive processes.

For instance, I ran IBT set to "Maximum" so that the fan would run at 3,250 above 63C. After 20 minutes 76F room-ambient, I get about 67.5C as an average of the maximum core temperatures. Now it looks as though IBT-"Maximum" is a pussy under air-cooling -- with this case, the fan-strategy, the ducting, and now -- serious noise reduction.

 

[BonzaiDuck]

UPDATE:

OK. O-O-O-KAYYY!!

Give me up to a week to write this up and append it to my originating post.

There may be a crosswalk of implicational value between air-cooling and water-cooling on the fan issue. I cannot say for sure. How would you duct fans for water-cooling -- or even "Why?"

On the other hand, thinking back when I bought a couple packages of the stuff in 2005 or 2006 (it keeps in storage REEE-ally well!!) -- the Spire acoustic rubber foam need never have been applied to the entire inside of a computer case. The best strategy is to target specific sources of noise, and apply the foam rubber first to the sources. Ducted airflow makes that very convenient.

There are two types of noise characteristic of high-output fans: air-turbulent white-noise, and a tone that has a rising pitch and volume. There is very little to be done about the first type or white-noise. You accept that fact that under severe load and a fan profile that ramps up a Gentle Typhoon AP-30, it's going to sound at best like a nearby air-conditioning vent on a hot day. So that's a given.

But it is the tonal sound that drives people crazy. It is the tonal sound that makes people obsessive about keeping their computer quiet, and leads to choices of low-output noiseless fans.

There is also a limit to how much CFM contributes to heatpipe cooling effectiveness. So what I've discovered is this. In the range of perhaps 3,600 to the top-end 4,200 RPM, there is no added cooling benefit for that 600 RPM delta. None at all. In fact, it just may be that the limit with the AP-30 is more like 3,200 RPM.

Without any kind of noise abatement, the quiet range of the AP-30 degrades quickly after about 2,100 RPM. 20% duty-cycle for the fan is around 1,200, and the fan is dead-silent at its lowest operable speed through 2,000. Again, keep in mind there's nothing to be done about air-turbulence, and there isn't much that's disturbing about it. You might be aware that your computer is "running," but air-turbulent white-noise with no tone is a lot better than an older refrigerator kicking on (with a tone), or your microwave heating dinner (with a tone.)

What I've done so far is to make the AP-30 "tone-dead" silent up to 3,650 RPM, and I haven't yet tested the range beyond that, because it WILL include a tolerable level of tonal noise, even over a small range above 3,650 RPM.

So . . . . BIOS profile set to 4.7 Ghz OC setting. Ran IBT with the "Maximum" stress option and the existing fan-profile which I'd created at 4.6 Ghz. Core maximums averaged: about 71.5C. And according to the fan profile, and monitoring software, the fan ramps up to 3,650 RPM.

WITH . . . NO . . . . FREAKING . . . . TONE!! NONE! NICHTS!! NADA!!

As I said, in a week I'll post the write-up in my original post, pictures with circles and arrows on each one.

[So much trouble and tedium for quiet, high-CFM air-cooling. Next year, as I've said many times, there will be a new system with water-cooling. Even if the desktop goes the way of the steamboat . . . . ]

You can politely humor me if you are splitting your sides with amusement over this low-tech project. That's all I ask . . . Even so -- I can take a joke pretty well . . .

 

[krnmastersgt]

I can only think of 2 reasons that people would duct their fans on a radiator in a watercooling set up. The first is similar to your whole experiment and the noise as the air pattern from many of the fans used in most water cooling set ups (either super powerful or designed to be a balance of quiet but still decent airflow) are not the same straight from the fan blades as it might be even 1 inch from it. As the fan blades displace air at different rates along the blades path a different noise response can be seen in some radiators. Some super high fins-per-inch rads actually serve to amplify a lot of the noise just because much of the airflow itself does not penetrate well along the majority of the fan blade. So with that in mind when you duct the fan you create a pocket that can better even out the different air behavior to a more uniform one that allows it to create less noise overall and often times even improve airflow.

That actually ties into the second reason and the reason I think most people ever bother to duct their fans to a radiator, as fans displace air differently they also have different levels of pressure depending on where along a blade you might measure. Not to mention the very center point of a fan typically has a dead spot as the motor housing has no direct airflow behind it, so you'll often see dust patterns on rads having a completely clean spot in the middle. So when you duct the fan the pressure like the aforementioned airflow is evened out in the pocket and the fan performs better acoustically and performance wise as now all the fins are getting a relatively equal amount of airflow across as opposed to some getting none and others getting more than the rest. This is also the reason I've decided to try and fit push-pull into every radiator in my forthcoming rebuild as a push fan can provide better temps (marginal but pretty consistent in this regard) but seems to waste a lot of its efforts in air that gets deflected on a high FPI radiator. A pull config seems much easier to clean and maintain because it performs much more evenly with a slight tolerance between the fan blades and the fins themselves but combine the two and you'll get the theoretically best airflow as 1 is gently forcing air against the fins while the other is creating a negative pressure environment on the opposing side encouraging more airflow between. Also because more fans = lower rpm but same airflow rate = lower noise!

 

[Essence_of_War]

push fan can provide better temps (marginal but pretty consistent in this regard)

When people like LinusTechTips have tried this, their data does not show this. Rather, when high quality static pressure fans are used, both 'push' and 'pull' configs come out identical up to experimental error.

 

[BonzaiDuck]

When people like LinusTechTips have tried this, their data does not show this. Rather, when high quality static pressure fans are used, both 'push' and 'pull' configs come out identical up to experimental error.

That may be, and I don't pretend to address the water-cooling configuration. But between the metal radiator fins and the number of fans, I can see how noise can become a nuisance when a major purpose of water-cooling had always been that of noise-reduction in addition to the quantum improvement in load temperatures.

But Sarge has a point there about ducting "evening out" the airflow. That was one reason I initially attempted to add a 25mm/38mm "exhaust pocket" or duct between the case-metal and the exhaust fan's output side.

This became a problem, because there is "extra torque" from a beefy fan pulling down on rubber-rivets: some of these gum-rubber rivets such as those bundled with AKASA fans are very "flexible." But it was even worse trying to design, insert and secure a duct from a fan sitting close to the cooler -- without trying to hang the fan on the cooler itself -- when the exhaust fan vent is off-center from the cooler.

I'll add some more points about this with my "pictures & arrows" post to the original. I don't have a decibel meter, so I can't scientifically measure noise reduction; I can only analyze the types and sources of noise, and make qualitative conclusions that this or that aspect of the project reduced this or that type of noise.

I had originally obsessed myself with ducts when they were much more effective with earlier heatpipes and heatsinks. With these coolers, though -- Noctua, ThermalRight and so forth, you might get 5C degrees improvement -- I confirmed this the other night when I temporarily removed the duct and ran the "test" provided in my ASUS software. About 4C degrees. Instead, with this beast of an AP30 Typhoon fan, the benefits lean significantly toward the noise-reduction.

At that point, the builder-user-enthusiast has to ask him/herself whether it's worth it to buy a high-output fan. Over some airflow range, there are as much as 5C degrees to be gathered over some of the limp stock fans with these coolers -- keep in mind I'm using the NH-D14, which has been slightly eclipsed by newer models of various makes. Beyond some level with pressurization, you won't find additional gains for higher CFM and higher RPM from the fan's top-end. So the choice of such a fan simply widens the range, leaving a little surplus that only increases noise levels.

Even without using that surplus, you likely won't be eager for a fan like this unless you can attenuate the noise.

 

[WoodButcher]

Yeah Duck, I was tempted to reply to that but I would just take your thread off on a tangent. I almost take offense reading some of linustechtips about being an old timer when he referred to heatercores. Those of us that used them know the value of ducting, pros and cons of push pull and have seen first hand where and when the point of diminishing returns comes into play.
You've been ducting and playing with the concepts as long as I can remember here so I'll close my yap here.
In any case, let's see what you came up with!

Where's the pics????

 

[BonzaiDuck]

Yeah Duck, I was tempted to reply to that but I would just take your thread off on a tangent. I almost take offense reading some of linustechtips about being an old timer when he referred to heatercores. Those of us that used them know the value of ducting, pros and cons of push pull and have seen first hand where and when the point of diminishing returns comes into play.
You've been ducting and playing with the concepts as long as I can remember here so I'll close my yap here.
In any case, let's see what you came up with!

Where's the pics????

In my camera. I need to upload, do the circles and arrows thing, and stick them in the text -- which I'm writing. Seems weird, but lately, things just seem to take me longer. It's probably a "real" "old timer" difficulty . . .

 

[BonzaiDuck]

LOL Duck, credit card shock cured me.
Well, maybe not.
My focus just got redirected to my van.
My current project is to put a V8 in a minivan.:whiste:

OK, where's the pics?

OK!! OK!!

Check the initial post to this thread, and follow to the beginning of the titled piece "DUCTING FOR DECIBELS." I had to ditch a couple pics, because I was limited to 10.

 

[JEDIYoda]

In my camera. I need to upload, do the circles and arrows thing, and stick them in the text -- which I'm writing. Seems weird, but lately, things just seem to take me longer. It's probably a "real" "old timer" difficulty . . .

or your having a senior moment......

 

[WoodButcher]

Sorry not to respond right away, I needed to have time to sit, read and digest.

In your opinion would it be possible to quiet 3 or more of these on a rad for watercooling? Keep in mind the spacing between fans on a radiator is nil in most cases, I wonder what effect the combined fans would have on sound levels?

 

[BonzaiDuck]

Sorry not to respond right away, I needed to have time to sit, read and digest.

In your opinion would it be possible to quiet 3 or more of these on a rad for watercooling? Keep in mind the spacing between fans on a radiator is nil in most cases, I wonder what effect the combined fans would have on sound levels?

Here and there, something has been said about the effect in decibels by adding multiple fans of the same specification to a case. I mean this in general -- without acoustic enhancements, and I can't remember the verdict. If a verdict were forthcoming, I'd bet a little money on the noise-level being less than linear with the number of fans, but I don't rightly know for sure.

If it were me, I might attend to first isolating the radiator itself from case-metal -- with some sort of sturdy, reliable rubber mounts. I might then consider building a duct-box in the shape of the radiator, but even so just a little thicker than the radiator itself or otherwise the thickness of the fans. If there's room for it, a little thicker might be better. You really don't want the fans to come into direct contact with the radiator.

The fans would be mounted in the box itself, and could be enhanced by layers of acoustic foam-rubber inside and out. If the thickness of the box permitted, and if the exhaust of these fans is blowing into the radiator, you could add the little Spire foam circles to the fan hubs -- even for just a quarter-inch layer. This would mean that the fan would be separated from the rad by 1/4" or more, and you'd have some small "dead spots" at the radiator, but probably nothing of much significance. If you don't like the dead-spots, then isolating the fans from the rad might be enough.

You could "collar" the fans, and extend the duct-box a bit on the intake side, but that's filling the case with a lot of duct-box if you have a 3x120 radiator.

Also, you can opine here and I'd be interested; but it seems that the noise isn't "less" on the intake side. You might remember the Doppler effect, exemplified by a moving train. In this case, the air is moving toward a fixed object, and the noise waves set up in the opposite direction. That's why I just suspect that the rising pitch of my fan tone isn't simply a function of its RPM and the mechanical parts of the motor. It's the RPM, the motor and the direction of airflow.

There's at least a 50-50 chance I'll build a water-cooled "E" system machine between the processor's release-date and six-months hence. The 50-50 chance depends on what Noctua's Jakob Dellinger give us with the release of the D15 next month or sometime thereafter. I want to see the reviews -- first decision-point; and second, if the reviews prove the cooler, I'll buy one just to test it on this existing machine. I can leave the fan in place! I only need to build a different duct-box!

Probably the chance of water-cooling is greater than 50% even if the Noctua cooler astounds the community. I still want to build a bong, too, and I need a water-cooled rig to do that.

First - the reviews of the D15; then, maybe I'll get one. After that, the E-machine -- likely with water, and especially because I eventually want to build a bong.

AFTERTHOUGHT [and hope you don't miss it here]:

Depending on the size of the case and other limitations, I'd think you could make one or two fans do the work of three if you can distribute the airflow reasonably to the radiator fins. In that case, a single fan could sit some distance behind the radiator -- lots of ducting and acoustic opportunities -- but the greater the distance the more crowding inside the case. Even so -- attenuating the noise for one fan might be easier than doing it for three. If you could get 110 CFM out of the fan with less annoyance and 140 with somewhat more, I'm only thinking that might be worth something there.

MORE . . . Truth is, I only guess you could build a duct box around the radiator together with the fans. If the computer sits above the floor on casters, you might put a radiator outside the case and the fans inside. Offhand I'm guessing this could be an "intake" setup. More trouble with noise with the fans outside the case . . .

And like I said -- I only could guess or wonder about acoustic-enhancement with a radiator . . .

 

[WoodButcher]

Ya know,, I'll bet one of these AP-30 ducted to pull through a triple rad would do the job.

 

[BonzaiDuck]

Ya know,, I'll bet one of these AP-30 ducted to pull through a triple rad would do the job.

Well, looking toward my "big plans" for later and into next year, It's something I would consider. You'd want to see that airflow were somehow spread over the radiator, but the air resistance of the tightly-spaced fins of the radiator might see to that sort of effect.

If I were wrong, I don't think you'll sue me for this:

http://www.coolerguys.com/840556014003.html

. . . or the $5 to $20 you might spend on foam board. Since you might want to move sooner on this idea than I actually can on my own, I'd sure like to know how it turns out.

I forgot to put a tip in the "big original-post amendment:" the art-board bends to any angle you want along a line if you score one side through the paper backing along the line where you want the corner, and maybe two lesser scores a 16th inch on either side of that line. This time for my project, I just made one cut through the paper.

You can punch holes through that stuff with a standard pliers-type office hole-punch and use appropriately-sized nylon screws with it. You may even be able to make it work with rubber fan rivets. As far as I can tell, it's about 5mm thick.

 

[BonzaiDuck]

The one picture I didn't take or forgot to post was the 1.75" dia., 2"-thick barrel-shaped piece of vinyl packing foam with same diameter Spire that I stuck on my case-exhaust port so that it just matches the center of the fan-motor hub.

I decided this helped a little bit when I first positioned it over the vent to see what happened to the noise with and without.

Meanwhile, I'm going out in a minute to pick up some Chinese carry-out and visit the drug store. At the drug store, I hope to find those wooden tongue depressors doctors use -- oversize popsicle-sticks. I'll need two of 'em with my dremel to cut a sort of lap joint.

What are the noise-suppressing qualities of a cork from a wine bottle? I think you could make it wider after the construction with a strip of spire and some Super Glue. . . . I shall cut two 2mm-wide slots at one end of the cork in a perfect right-angle cross.

Got plenty of Super Glue and Pit Crew's. But will it pay off?

[That was your idea by the way -- WoodButcher!!]

 

[WoodButcher]

I gave the idea some considerations w/my last build but didn't have the room in my case.
I won't be building again for quite a long while other than upgrades or repair as my attention is elsewhere (engine swap). My PC, currently a Q8400, 8800GT, six gig ram and 4 notebook drives, one an SSD for the main will meet my needs for a while.
My rig is in a case I got when win 98 came out.

from this,


to this notice the AIO watercooler, I think I did that in '05



and this the latest itineration/ incarnation

That case has been resurrected many times between those pics and I've built many others as well but I think unless someone changes the ATX format it will likely be one of my last.

I know your using the HAFs now but what happened to that server case you had? The silver one, the name slips my mind,,,

 

[WoodButcher]

The one picture I didn't take or forgot to post was the 1.75" dia., 2"-thick barrel-shaped piece of vinyl packing foam with same diameter Spire that I stuck on my case-exhaust port so that it just matches the center of the fan-motor hub.

I decided this helped a little bit when I first positioned it over the vent to see what happened to the noise with and without.

Meanwhile, I'm going out in a minute to pick up some Chinese carry-out and visit the drug store. At the drug store, I hope to find those wooden tongue depressors doctors use -- oversize popsicle-sticks. I'll need two of 'em with my dremel to cut a sort of lap joint.

What are the noise-suppressing qualities of a cork from a wine bottle? I think you could make it wider after the construction with a strip of spire and some Super Glue. . . . I shall cut two 2mm-wide slots at one end of the cork in a perfect right-angle cross.

Got plenty of Super Glue and Pit Crew's. But will it pay off?

[That was your idea by the way -- WoodButcher!!]

Dooonnnn''t blame me! lol remember the case that had jet cones on it?
That and the silenx hubs, it really makes sense when you think about it,
A. noise is made from the blades hitting the "still" air,
B. a fan will make less noise in a breeze blowing in the same direction
C. a cone can eliminate a dead air space that is "still" that gets picked up

smooth the flow? maybe,, I don't know

 

[BonzaiDuck]

Dooonnnn''t blame me! lol remember the case that had jet cones on it?
That and the silenx hubs, it really makes sense when you think about it,
A. noise is made from the blades hitting the "still" air,
B. a fan will make less noise in a breeze blowing in the same direction
C. a cone can eliminate a dead air space that is "still" that gets picked up

smooth the flow? maybe,, I don't know

Well, on the "down" side -- drug-store doesn't have 'em. On the "up" side, the medical-supply store on a corner I often pass would have them. And before I go down there, I might try to weld together two popsicle sticks and see how that works. If it doesn't, I'll go out to the garage wine-chiller and get another bottle. And plan my next errand-trip.

. . . the cream-cheese wontons were pretty good, too . .

On this . . . cork-and-Spire X-e-fiction. If it's sturdy enough by itself and properly installed, it may help orienting the duct-box so it never crosses the path of the fan blades, although that would be difficult even now. And I think if it doesn't work, I can just cut it out of the box without leaving damage . . .

 

[WoodButcher]

I know your using the HAFs now but what happened to that server case you had? The silver one, the name slips my mind,,,

You spoke of it recently, what was the name?

 

[BonzaiDuck]

You spoke of it recently, what was the name?

It was a Compaq Proliant case. It was at least one and a half times the height of my HAF 922 midtower, and it was several inches higher than any spec full-tower I'd ever had or seen. And it was about two inches wider than my HAFs, which were marketed touting their 200mm LED fan in the front panel. There was a tall drive-cage attached with machine screws from the case-interior, which held some four or five of the OLD SCSI drives, or double the height of your mainstream WD Blue SATA-6 desktop drive. There was circuit-board behind the drives but inside the case for "hot-swap" connection. I found a part from a Windows 95 IBM midtower and pop-riveted in a hinged power-supply cage that just unlatched and popped out of the top, and a DELL midtower drive cage, which I made mountable to the bottom of the big ProLiant cage. It had rubber-acoustic mounts on the drive rails.

Somewhere, there are pics, and I can link them, but I thought maybe I was spotlighting an old, old project going back to forum posts here in 2007 -- even if I wish I could get it back now.

I always build a system without tearing apart an old system first, and if it replaces my main computer, there has to be an overlap of some several months. I also pass them around the family -- hand-me-downs. So I was so work-weary in late 2011 that I finally told my brother across town that I wasn't going to rebuild his "new" machine in a HAF case, but I'd just give him my "Chrome Window" machine with the box, the 3-inch braked double-casters, the hot-swap SATA setup, the front-door with the latch and key, and the special red LEDs which each blinked according to activity on its own HDD of four in RAID5. I had plans for HDD-array-driven "lightning" in the window, which was perfectly prepared using automotive mirror window film -- tricky with Lexan versus glass.

Somehow, I'd have to scheme with my brother to get that case back. On the other hand . . .

. . . It was just . . . so . . . darned . . . big . . . .

 

[WoodButcher]

Proliant, that was it. IIRC you started ducting that monster to improve the air around the cpu.
I played with a rocketfish for a while and in the end gave it to my kid with a TR cooler ducted in a similar fashion as yours because the darn thing was too big to evacuate with just the one fan.

 

[BonzaiDuck]

Proliant, that was it. IIRC you started ducting that monster to improve the air around the cpu.
I played with a rocketfish for a while and in the end gave it to my kid with a TR cooler ducted in a similar fashion as yours because the darn thing was too big to evacuate with just the one fan.

That's a general prognosis, I think. With the ProLiant case, I'd modded the (massive) rear of the case to add a second San-Ace exhaust identical to the one which fit into the ready-made vent. There was a single 140mm fan at front-bottom, two 140mm fans in an aluminum screened cage on the bottom. There was a pusher-fan in front of my (big) duct for the TRUE cooler. So . . . 2 . . 4. . . a total of 6 fans. I think I may have removed the single front-panel 140mm because the duct pusher fan was pulling air past the drives in the drive cage. I had ready-made screens and Lexan panels to block the holes of "deleted" fans.

Adding the two aluminum handles on the case-top, the "aircraft" toggle switch for the power-button, the little red toggle switch for the circular neon/cold-cathode light added to the duct pusher fan as completed features -- I wish I'd finished the "bling" refinements I'd mentioned before.

But of course . . . it was SUCH a BIG case . . .

THE CORK'n'SPIRE X-e-fiction: I think however this develops, the cork needs to be shorter or it will be too close to the cooler fins. The cork becomes a convenient construction device, but I think something similar and stubbier can be accomplished with the art-board and Spire. It is just important that it is solidly-constructed and precisely installed so there's no risk of fouling the propeller.

I'm . . . thinking about it. Maybe -- use a small dovetail-saw. The right kind of glue should make it solid.

 

[BonzaiDuck]

Well -- this is the update.

Instead of having the cross-braced or "X-design" corner-to-corner framing of popsicle sticks to hold a wad of cork and SPIRE foam in place just forward the moving fan-hub on the intake side, I only used a single popsicle-stick brace from one corner of the box to its diagonal. It leaves about a centimeter of space between the spinning fan and the "cork'n'rubber-cone."

I can only say at this point that I THINK it further reduces the volume of the Typhoon tone -- insofar as the tone transmits itself to the cooler fins. I THINK that I can increase the fan speed from 3,200 to between 3,400 and 3,600 -- ending up with the same level of noise as before, but the project has done so much to attenuate the noise of the fan that it is harder and harder to tell for each successive augmentation of this strategy.

It certainly doesn't affect airflow: the temperatures under the earlier fan profile are the same.

I think that the Typhoon is still transmitting vibration directly to the case despite my gum-rubber fan rivets. I can feel some of that vibration on the rear case panel. But other than rubber fan rivets, I'm not sure what else to do, or that I really want to do anything more.

So as far as I can tell, with the noise-mitigation strategy, one can exploit the fan's speed range between 1,200 and ~3,500 RPM. You would still hear something of the fan's tone at that upper boundary, but it is not very loud, nor nearly as loud as the fan had been at lower RPMs without this "project." For those speeds, even the white-noise -- which dominates the tone now -- seems subdued.

The only other thing that might help is a different duct design that puts the fan suspended midway between the metal exhaust-vent and the cooler. And I'd shown my "sliding rail" design for the duct in such an approach. There's only so much room between the fan and the cooler (noticeably more promise for a single-tower unit, like an NH-U14S). More distance between cooler-fins and exhaust vent would make it more feasible for that type of fan-placement.

Since everything is reasonably "tip-top" at this stage, I may wait until I feel inclined to build a better duct. I can see how refinements can be made. But most people might think this exercise tedious -- even a bit nutty.

Cool and quiet, though!! I got that! And with the beast of the Typhoon model line, I think! Folks will decide for themselves whether a high-output fan is worth installing with the noise-mitigation. For me -- it works. Even if I spent a lot of time fiddling with "construction paper," glue and foam-rubber.