Soliciting brief insight on the "dead spot" of fans

[BonzaiDuck]

We've had some hot discussions recently about "AiO versus heatpipes" and related topics.

Someone else proved the advantage of eliminating the dead-spot on cooler or radiator fins from the fan hub(s) -- by inserting a "duct" or spacer between the fan and the fins. It occurs to me how easy it would be to do this, whichever heatpipe cooler one uses, and for some unspecified AiO designs. But it would be easier for the heatpipes.

For instance, everybody gets their panties in a bunch over the "weight of heatpipe towers," when the problem is really a matter of torque, and the biggest component to that torque is the 7oz fan or fans (14oz) attached to the cooler.

If I had only a single fan hanging on the heatpipe tower, it wouldn't significantly change things to insert a 25mm-thick fan-shroud between the cooler and the fan, and I could probably secure the assembly with the existing clips or mechanisms provided with the cooler.

But what sort of benefit would I expect? ALL the pipes and fins of a heatpipe tower are getting air at SOME parts of their surfaces. Obviously, eliminating the dead spot would make an improvement. I'm just wondering if it is "estimable."

 

[loladinas]

I've been pondering something along those lines myself as one of my NB-MF fans crapped out (starts vibrating so much at max. rpm that I can feel it through the computer desk itself) on me recently. To be honest they were a bit anemic anyway. I've ordered a couple of Delta AFB1212LE fans to play around with, that I should be getting in a couple of weeks. I'm still unsure on what heatsink to get though. Silverstone HE-1 seems like an obvious candidate. Neither weight nor RAM clearance is an issue in my case so just getting the biggest chunk of metal seems like the intuitive choice for ducting, no?

 

[BonzaiDuck]

I've been pondering something along those lines myself as one of my NB-MF fans crapped out (starts vibrating so much at max. rpm that I can feel it through the computer desk itself) on me recently. To be honest they were a bit anemic anyway. I've ordered a couple of Delta AFB1212LE fans to play around with, that I should be getting in a couple of weeks. I'm still unsure on what heatsink to get though. Silverstone HE-1 seems like an obvious candidate. Neither weight nor RAM clearance is an issue in my case so just getting the biggest chunk of metal seems like the intuitive choice for ducting, no?

Thanks for your response and remarks.

I've used 120x38mm Delta Tri-Blade fans, with thermal-control from the motherboard to keep the noise down. Looking at the spec for the AFB1212LE, I just think you could do a lot better than that.

I can revisit that thought with greater detail if you want.

On the matter of the HE-1. I would like to see a comparison review such as might be found at Hardware Secrets, which compares it with several known models.

If you can link such a review, that would be a big help, but I'm going to start searching . . . . right now . . .

The problem with the center fan, no different than my Akasa Viper 140R between the towers of my D14 cooler, is that it's offering a dead-spot to both heatpipe towers.

With a double-tower, the only way to get around that involves ducting the entire cooler and either sucking air through the duct and fins from a beefy case exhaust fan, or doing so in combination with a pusher fan on the forward tower installed with the 25mm fan-frame spacer I mentioned. The duct would have to completely close off the space between the towers to any other air source but the pusher fan.

Funny. I thought I was "getting out of the heatpipe cooler rut." But I'm always curious when someone raises my attention to a model I hadn't seen before.

 

[BonzaiDuck]

HardOCP has a review comparing the HE-1 to the Arctic Cooling Freezer 13 and i30. Someone here many moons ago had shown some results or comments about the i30, and it was presentable but nothing special among the heatpipes.

I CAN say, however, that the testing done with their i7-3770K @ 4.5 Ghz shows approximately the same temperatures with Prime95 that I would get from an NH-D14 -- without any special "mods." It could be more or less, I can't say. But the review wasn't entirely clear as to which test they used in Prime95.

If you want to "do things" with your rig and heatpipe cooler, the HE-1 probably has as much promise for the results as any other cooler I know.

http://www.hardocp.com/article/2012/09/13/silverstone_heligon_he01_cpu_cooler_review/3#.VWaxd8LbKUk

Whoops! Spoke too soon, and found this -- Frosty Tech always has some reliable comparisons:

http://www.frostytech.com/articleview.cfm?articleid=2689&page=5

Even the 212 EVO might edge out the HE-1. Otherwise, a 4C unfavorable difference from the NH-D14.

On the other hand. It has a nickel-plated base, or so it looks, and so I thought I just read somewhere. Lap that sucker and get rid of all the nickel. At that point, I'd lap the processor down to bare copper, also.

On our original topic here, it would be interesting to see how much the HE-1 would improve with the tedious mods we're discussing.

 

[loladinas]

Actually I do have both 212 Evo and NH-D14, in other computers, so buying something worse would be rather moot.

I can see why EVGA ACX is so effective; fan fits flush with the cooler, staggered heatpipes, moved off-center as to avoid the deadspot from the motor, closed off sides to avoid useless leakage, and there's probably some engineering magic that's not as obvious... However if I were to use another shroud to remove the fan further away from the fin stack, that should, in theory, lessen the deadspot problem. And if I were to encase the fin stack with art board that wouldn't that lessen the ACX advantage even further? At that point wouldn't it make sense to simply go bigger?

Do tell about the fans, I just picked those as they seemed to be able to scale down with voltage rather low and not be the stereotypical Delta "screamer". I do have a pair of Noctua F12s to test things out with, but I wasn't terribly impressed with them either.

 

[Micrornd]

Someone else proved the advantage of eliminating the dead-spot on cooler or radiator fins from the fan hub(s) -- by inserting a "duct" or spacer between the fan and the fins.

I'll give you something to contemplate. :whiste:

While it would appear you are eliminating a "dead-spot" by the quoted method (and you may be slightly doing that), what you are really doing is optimizing the distance of the fan from the HS or rad to achieve maximum velocity of the airflow.

Contrary to what is normally done (placing the fan directly on the HS/rad, both for compactness and cheapness), the maximum velocity of the airflow from the fan is not at it's face, but generally, at least the thickness of the blade pitch from the face.
Placing a fan directly against a HS/rad also explains the need for a high SP, since the airflow is less than optimized and actually compromised.

The same concept is also true for using the fan as a "puller". Maximum suction is not achieved at the face of the fan, but, again, a distance from it, similar to the above.

My testing has shown me that a high SP is not needed if the fin spacing of the HS/rad and fan spacing from the HS/rad are optimized.
When testing or using this concept, making sure that all air from the fan is going into the HS/rad is very necessary, as is making sure all airflow goes entirely through the HS/rad and cannot escape out the sides of the HS/rad.
Intel's wind-tunnel HS's on 603-604pin Xeons are a good example of this, and why I chose to investigate this further. :biggrin:

It is, of course, best to use air from outside the case for cooling the HS/rad, as long as any ducting is very short, as the fan should be pushing or pulling, but should not have be doing both, as that just increases the need for a stronger fan unnecessarily.
With proper fan spacing and ensuring all air goes through the HS/rad, even using the already heated air inside the case will show a measurable improvement over the accepted norm and will allow fans of less airflow (and noise) to be used, even in maximum cooling situations.

One last, but very important point to remember, is that if you do succeed in routing all the airflow from the HS/rad fan(s) actually through the HS/rad, then cooling of the MB components must be accomplished by others means, as the normal method of cooling the MB components in many cases (VRMs in particular) is spillover/leakage/deflection from the HS/fan.

 

[piasabird]

When I was building AMD Athlon computers we use to use a scoop that was about 1.5 - 2.0 inches from went from the base of the CPU cooling fins. It also reduced the aperture through which the air traveled and it worked pretty good.

Of course you could just use a Zalman cooler that looks like a flower. All the fins are in a circular pattern like flower petals.

http://www.newegg.com/Product/Product.aspx?Item=N82E16835118003

 

[BonzaiDuck]

I'll give you something to contemplate. :whiste:

While it would appear you are eliminating a "dead-spot" by the quoted method (and you may be slightly doing that), what you are really doing is optimizing the distance of the fan from the HS or rad to achieve maximum velocity of the airflow.

Contrary to what is normally done (placing the fan directly on the HS/rad, both for compactness and cheapness), the maximum velocity of the airflow from the fan is not at it's face, but generally, at least the thickness of the blade pitch from the face.
Placing a fan directly against a HS/rad also explains the need for a high SP, since the airflow is less than optimized and actually compromised.

The same concept is also true for using the fan as a "puller". Maximum suction is not achieved at the face of the fan, but, again, a distance from it, similar to the above.

My testing has shown me that a high SP is not needed if the fin spacing of the HS/rad and fan spacing from the HS/rad are optimized.
When testing or using this concept, making sure that all air from the fan is going into the HS/rad is very necessary, as is making sure all airflow goes entirely through the HS/rad and cannot escape out the sides of the HS/rad.
Intel's wind-tunnel HS's on 603-604pin Xeons are a good example of this, and why I chose to investigate this further. :biggrin:

It is, of course, best to use air from outside the case for cooling the HS/rad, as long as any ducting is very short, as the fan should be pushing or pulling, but should not have be doing both, as that just increases the need for a stronger fan unnecessarily.
With proper fan spacing and ensuring all air goes through the HS/rad, even using the already heated air inside the case will show a measurable improvement over the accepted norm and will allow fans of less airflow (and noise) to be used, even in maximum cooling situations.

One last, but very important point to remember, is that if you do succeed in routing all the airflow from the HS/rad fan(s) actually through the HS/rad, then cooling of the MB components must be accomplished by others means, as the normal method of cooling the MB components in many cases (VRMs in particular) is spillover/leakage/deflection from the HS/fan.

Nothing to dispute with any of that. On the "intake air" issue, the mainstream approach with heatpipe coolers follows the traditional design of cases with rear-exhaust and ATX motherboard orientation. But even enthusiasts tend to accept things as they are.

Consider other cases. I think there was a Lian Li, and possibly a Silverstone model, which reversed the orientation of the motherboard and mobo-pan to fit at the left side of the case. Now that I think about it, there was brief flirtation with a departure from the ATX standard. I think it was called "BTX."

OEMs like Dell were quick to jump at this possibility. I have seen older LGA-775 systems using Dell's own custom-designed aluminum heatsink ducted to the case front-intake. There is no doubt from inspecting those systems that ducting was a deliberate choice in their equation.

Someone else questioned the usefulness of surrounding a heatpipe tower with a duct-box: that somehow, it would deprive an opportunity to draw air through the sides of the cooler. But some of these coolers, like the ACX, have the fins closed off on the sides. Instead, putting a duct-box surface close to those sides with a gap of about 1/4" would force intake air to move past those exterior surfaces, and the restriction would improve the velocity and therefore cooling effectiveness. In other heatpipe designs, one would wish to make use of open fins for more airflow through the cooler in general.

As for the ACX, I also suspect some sort of "engineering magic." The fins are dissimilar from the usual thin "aluminum-foil" design. They're thicker. I suppose the only way to find out would be to cut one in half. Also, being thicker fins, they don't vibrate as much as those of your 212 EVO or Noctua. That's my tentative explanation as to why I need no acoustic mods for the AP-30 exhaust fan on the ACX-configured system -- or no more than a TR accordion duct provides. How significant it is, I can't say. but any noise transmission is obviously less.

On the matter of motherboard component cooling -- yes. The crude solution in my systems at the moment is a high-airflow 200mm NZXT or BitFenix Spectre-Pro fan, augmenting the front 200mm fan.

The better solution takes a cue from most of the ASUS Sabertooth boards. Anyone should be able to do it, but it's "time and trouble." Foam-art-board is the answer to replicating Sabertooth plastic. One would then fit one or two small fans to the art-board at chosen locations to force air between the motherboard and art-board duct-plate to exit from the sides.

I still have the stencil "PDF" I was making to use for cutting the PCI-E slots in a foam-board panel. I was using a photo of the Z68 board, and there will be all sorts of trouble just "getting it right" for a 1:1 precise stencil. Definitely, a lot of trouble, and a reason I've shelved that project for now.

 

[loladinas]

Consider other cases. I think there was a Lian Li, and possibly a Silverstone model, which reversed the orientation of the motherboard and mobo-pan to fit at the left side of the case. Now that I think about it, there was brief flirtation with a departure from the ATX standard. I think it was called "BTX."

That's the one I have, TJ08e.

So I've been browsing heatsinks and HR-02 Zero caught my eye, taking that case into consideration I think it might work out, putting one fan at the front of it, using a spacer, and putting another one on the exhaust went, and then ducting it to the heatsink. Thoughts?

 

[BonzaiDuck]

Tell me about the fans . . . make, model, size, spec. I'll leave it to you to either decide -- or find out -- if a duct box at the rear helps at all. It won't "hurt."

I could say any number of things. How about this? "I like it!"

I could wonder if the center fan is actually helpful; I could wonder if separating the front cooler fan a bit from the fins wouldn't help -- as Micrornd (and I) might think.

The center fan does, in a way "provide ducting," but its hub is obstructing fins coming and going -- no different than my Akasa Viper 140R on my D14.

If you're thinking about the HR-02 Zero, it's not the "best" for stock-installation performance, but it DOES raise some possibilities. Some pictures that popped up on the web show a black accordion duct, and from the way it fits, I'd assume it's a TR product. Losing the D15 clutter, you could do the same for the pusher fan as I've suggested.

I've never been that eager to replace a heatpipe cooler that was decently in the upper range of performance reviews, but it's always a temptation. It's your choice.

I sort of wonder how the original blue duct that TR made has come down in price to less than half what it used to be. And I'm wondering how it is that I haven't seen these new developments for the HR-02 Zero:

http://www.thermalright.de/en/accessories/

If you can't buy it, you can build it. But -- where can you buy it?

Hmmm. Where is -- and who is -- "Nan's Gaming Gear?"

http://www.nansgaminggear.net/ProductDetails.asp?ProductCode=FAN%2DDUCT%2D120

See -- I have to laugh, even at heatpipe notable ThermalRight, when they say "this or that duct is not compatible with the HR-02 Zero" or other cooler. Fa chrissake! It's flexible rubber/plastic! It doesn't . . . have . . . to fit . . . PER-fectly! You can trim it. You can "extend" it. You can glue art-board to it. "It's not compatible. . . " I'm going to split my side snickering.

$7 from Nan's or $3 from Performance PCs to get the old blue duct would allow you to find out if you can improve on the existing D15 configuration. All sorts of thoughts and possibilities. But . . . that's chump change . . . don't need to buy your $5 panel of art-board, sharpen your Xacto knife, make precise drawings, buy glue and "fiddle around."

That's the one I have, TJ08e.

So I've been browsing heatsinks and HR-02 Zero caught my eye, taking that case into consideration I think it might work out, putting one fan at the front of it, using a spacer, and putting another one on the exhaust went, and then ducting it to the heatsink. Thoughts?

 

[loladinas]

Well, currently I have NB Multiframe M12PS, Noctua NF12 at my disposal, and the said Deltas' on the way. So I guess I'd try the Deltas' first, using a butchered old CM fan frame as a spacer for the front of the heatsink, and a ducted Delta as an exhaust. I'm still using the AP181 that came with the case as an intake.

I'm not wary of buying a new heatsink since I've got three computers rigs running, and one of them is still with the stock Intel cooler.

Yeah, I was considering buying the HR-02 Zero, and it's being sold as a bundle with that duct.

 

[BonzaiDuck]

Well, currently I have NB Multiframe M12PS, Noctua NF12 at my disposal, and the said Deltas' on the way. So I guess I'd try the Deltas' first, using a butchered old CM fan frame as a spacer for the front of the heatsink, and a ducted Delta as an exhaust. I'm still using the AP181 that came with the case as an intake.

I'm not wary of buying a new heatsink since I've got three computers rigs running, and one of them is still with the stock Intel cooler.

Yeah, I was considering buying the HR-02 Zero, and it's being sold as a bundle with that duct.

I tend to favor the fans with 100+CFM potential, like the iPPC 3000 model. If I were to replace the even more outrageous AP-30's I use, I'd pick that iPPC model.

But Micrornd's point may be well-taken. What would it be like with more modest airflow if the hubs are at a slight distance from the fins? And why worry about high static pressure? All in all, for me, I wouldn't need to replace either the AP-30 or the iPPC 3000. There's nothing of noise to worry about the way they're running.

OK. I think I have to try it -- remove the iPPC fan from my ACX, fix it to a fan-shroud and secure the fan shroud to the ACX. Just tedium and a little sweat. REally! I'm getting so OLD! "Oh! just a drudge! Pulling the computer to the floor, opening it up, removing the fan . . " etc. etc. I'll get to it, though!

[and a few minutes later . . ] Another one bites the dust! An OCZ 0.13A "nothing-special-in-black" 120x25mm fan from an old "bundle." I clipped out the motor. Better save the PWM cable and plug. I'll think about it some more, and then move to deploy this . . . remnant . . .

[and an hour or so later . . ] I've been scouring the web for some really good comparison reviews that include the TR HR-02 [Macho] Zero, and only found one or two -- mostly Silent PC and Bit-Tech. Nothing at Hardware Secrets or Frosty-Tech pops up. But this is what I glean from what I've seen so far. The Macho Zero is maybe a couple C behind the D14 in performance. The same review has a more stern evaluation for the D15 when I'd though the D15 was maybe 3 or 4C better than the D14: It only shows a 2C difference. But these slight irregularities are all likely explainable by the choice of test-beds and fans. The rank-order is probably still consistent with much more comprehensive review summaries.

However. ALSO INTERESTING. They didn't test the Macho Zero with the special Macho Zero accordion duct. Of course not -- they'd have to find a basis of comparison testing common to all the coolers. So I'm guess that -- with the duct -- you get 3C better than the stock D14, and then only 3C worse than the unducted ACX. That still gives the ducted ACX maybe a 4 or 5C lead over the ducted Macho Zero. It's close enough that I won't even try and sell posters on the ACX. It also has a more uniform shape on both sides.

I would ALSO bet that you'd make up part of that shortcoming by lapping the nickel-plate off the base. The ACX cooler is one of those "direct-touch" designs, if I'm not mistaken, but I lapped the hell out of it anyway just to assure flat uniformity.

I should really stop obsessing over heatpipes. Really.

 

[BonzaiDuck]

Just to report my little project to insert a fan-frame/shroud between my Noctua iPPC 120mm fan and my heatpipe cooler.

If there is any measureable difference and improvement, it couldn't be more than a degree C or two. I just can't tell with any certainty.

It also might be more quiet, or so I imagine. No harm done, though!