Fan shroud, it works!

Started this thread because I didn't want the last one to get more threadjacked than it already was:

CualvezAsteria said:

Those are incredible results imo. With my Sunbeam with the stock fan at high, I get 60C at 1.5V. Have you considered adding a shroud to the push fan? I don't really know if it will change your temps (mine maybe 1C?) but it is kind of a fun quick and easy mod.

Quick little guide, take an old 120mm fan, gut it so you just have the casing. Then attach it to the intake side of your fan so that it acts as an extra little casing... I don't really know the science of it but to me, intuitively, it seems like it would help.

Click to expand...

Bobnova said:

Don't confuse the 212 with the 212+, they're different.

I did the same thing

The 212 came out in 07 and was an ok cooler, but the two tower thing is really just them cheaping out on the aluminum, it hurts performance. It had a solid copper base.
The 212+ came out quite recently (comes with i5 brackets) and is a more classic tower cooler with HDT and full width aluminum fins and a newer type fan. It has better reviews then the previous one.


The science of the duct is thusly:
A spinning fan pushes (pulls, really, but that's a different story) air away from the hub, partly forward and partly out towards the rim. If you hold a fan in your hand and feel where the air is going, the majority leaves the casing at a 45* angle out from the outer edges of the fan, none from the very middle, and very little from the point between the hub and the blade tips, also almost no air is actually headed straight forward. As a result, the vast majority of the air goes into a 1" band around the outside of the HSF where the blade tips are.
A sealed chamber between the fan and the HSF gives the air time to collect itself and form a more balanced flow, it takes the high velocity air from the blade tips and slows it down, resulting in pressure instead. That pressure then forces the air through the HSF where ever it can, which is the entire surface rather then just that 1" band. Ideally you would have more like 6" between the fan and the HSF, but that just isn't practical. Even a 1" spacer helps balance out the flow a lot.
When you see fan blades with wildly different profiles at the hub and the blade tips what you are looking at is the engineers attempting to get more of the air to flow though the middle of the fan instead of just at the tips.
Some older Big Box computers (dell?) had a rear case fan that was ducted to the CPU heatsink, the heatsink itself had no fan, the rear fan pulled all the case air through the heatsink before spiting it out the back. It worked decently, as the flow through the heatsink was very balanced, as opposed to a traditional (at the time) square lump that had the hottest part of the CPU (the core) centered directly under the fan hub, where it got no airflow whatsoever.

(As a side note, a turbocharger uses this in the extreme, it spins the air out to the blade tips at extremely high speeds, where it goes through a narrow restriction (further increasing speed) before going into a much more open chamber, the air stops almost dead and in doing so the energy put into it in the form of velocity is kept, and is changed into pressure.)

Click to expand...

CualvezAsteria said:

Awesome, thanks for the info. So where can I buy me a turbocharger lol.

I believe then I've done this completely backwards. Right now I have my heatsink with the fan directly on it as is standard. Then on the back of that fan I have the shroud, the shroud is on the intake side...

Like this:
....<===== air flowing this way
000[xx][ ]
000[xx][ ]
000[xx][ ]
HS/Fan/Shroud

I'm going to put the shroud on the HS itself then the fan away.

I did notice how the air blows out in an angled directions away from the center not towards it. The way my HS has the fan on it, there is about an 8mm gap from the top of the fan and the top of the HS, and I felt a lot of air blowing out above the HS. I fixed this by electrical taping over that gap connecting the top of the HS to the top of the fan.

Will report back with results. For the record, right now I'm idle at 43C and load at 59-60C.

Click to expand...

CualvezAsteria said:

Idle was about the same, actually higher but it wasn't true idle; load is actually 57-58! So a 2C drop just from the shroud, wow! I checked the air flow and it definitely is still pushing out the sides of the heatsink. I think that putting another 120mm shroud would actually lower temps more!

Sorry OP for the thread jack, I'm going to start a separate one right now

Click to expand...

Basically the important thing is that I put a fan shroud on my CPU heatsink and it lowered temps 2C!

It's a 120mm fan casing attached to the front of the HS fan, so the casing is connected to the heatsink and acts as a mini duct of sorts to concentrate the fan's air flow through the heatsink instead of out its sides. A fan actually blows air out away from its center, not straight (perpendicular to the fan) as I previously imagined... so the shroud helps it gain a more straight through flow.

I currently have just one 120mm casing acting as a shroud but I'm going to double that up right now to two, will report back with results.

The reason that putting a duct between the fan and heatsink has dropped your temps is because there is a dead spot at the center of the fan, where the hub is. Putting the duct there helps the air spread out, and reduce the dead spot. The longer the duct, the smaller the dead spot becomes, although I haven't actually seen any real world testing done. Another thing that will make a large difference in temps, would be using 38mm thick fans, which have a higher static pressure than a standard 25mm thick fan. Using a pair of 38mm thick fans in push-pull will yield you even better results.

Any particular 38mm's that you recommend with good noise/perf? I don't mind noise but don't want a vacuum cleaner either (e.g. this lol).

shrouds are indeed nice.. i think the space issue is what limits most people though, besides the actual killing of a fan

if i was to ever upgrade my heatsink, id definitely take a couple san ace 1011's with some 120mm shrouds in a push/pull megatron setup

e6600 said:

shrouds are indeed nice.. i think the space issue is what limits most people though, besides the actual killing of a fan

if i was to ever upgrade my heatsink, id definitely take a couple san ace 1011's with some 120mm shrouds in a push/pull megatron setup

Click to expand...

Hmm would you be putting a shroud on the pull part? My first thought is that would be a bad idea as it'd reduce the pressure the fan would produce on the actual pulling of the warm air.

CualvezAsteria said:

Any particular 38mm's that you recommend with good noise/perf? I don't mind noise but don't want a vacuum cleaner either (e.g. this

lol).

Click to expand...

The San Ace L's would work very well, so long as you can find them. They seem to be quite hard to get ahold of these days. Also, a much easier fan to find, would be Panaflo mediums. I personally have a Panaflo ultra high, and the thing is loud. It doesn't bother me any, I have been more concerned with performance than noise. Right now, I have a pair of high speed Yate Loon fans in a push-pull on my Ultra 120A, and even being 25mm thick fans, are performing very well. The higher static pressure of a 38 mm would help me heaps, but because of the tall heat spreaders on my Dominator RAM, I can't quite fit a 38mm thick fan in.

I think I'm going to get this: http://www.performance-pcs.com/cata...t_info&manufacturers_id=167&products_id=24236

EBM Pabst W2GT110-AF45-19 120 x 38mm
Voltage: 12VDC (6-14VDC)
Current: 2.95 Amps
105.9 CFM
45 dB
All Metal Construction
Ball bearing
Power: 5.1W
Dimensions: (120x120x38mm)
Connection: 3-wire 3-pin fan connection 12"

Anyone know if I can plug this into my mobo directly on the CPU fan connector? I've never had a fan this powerful so I don't know if there's risk of blowing anything out. I know for the San Aces that risk is listed on sites.

Alright 2nd 120mm shroud installed and sufficiently tested. Results time!

No shroud: idle at 43C and load at 59-60C
1 shroud: idle I didn't test and load at 57-58C
2 shrouds: idle at 41C and load at 56-58C

With 2 shrouds the performance gain was minimal but noticeable to me just watching the temp patterns. Anyway this was fun and I'll be sure to post my results with a new 38mm Pabst w/double shroud action.

Most mobo's can't handle more that 0.5 amps on the mobo headers, but it is said that Gigabyte boards can handle up to 1 amp. I wouldn't recommend any more than that though. For a fan with that much power draw, you will need a pretty good fan controller. Also, being rated at 45 db, you will find that fan to be quite loud. I think my ultra high speed Panaflo is rated at about 46 db.

Cool, so I'll definitely power it through a 4-pin converter thing. Not going to get a fan controller as I plan to run it at stock...

The current fan that comes with my HS is at 51.3db (according to frostytech review) and I'm totally fine with it... Although I wonder if the fan rating for this Pabst is lower than what it actually is. If it is actually around 60db that would be for me. HRMMM.

deadlysyn said:

Most mobo's can't handle more that 0.5 amps on the mobo headers, but it is said that Gigabyte boards can handle up to 1 amp. I wouldn't recommend any more than that though. For a fan with that much power draw, you will need a pretty good fan controller. Also, being rated at 45 db, you will find that fan to be quite loud. I think my ultra high speed Panaflo is rated at about 46 db.

Click to expand...

Panaflo NMB MAT FBA12G12U
I'm pretty sure that's your fan. Holy crap that thing is loud! Now I'm scared of db ratings... hmm

CualvezAsteria said:

Panaflo NMB MAT FBA12G12U

I'm pretty sure that's your fan. Holy crap that thing is loud! Now I'm scared of db ratings... hmm

Click to expand...

That's the one. There's a reason my wife has nicknamed it "The Beast". I love it, but she hates it. I don't actually have it in my rig now, but eventually I will find a place for it again.

Alright, new choice! 42db version

http://www.performance-pcs.com/cata...t_info&manufacturers_id=167&products_id=24238

EBM model W2G110-AN29-87
Voltage: 12VDC (5.5-14VDC)
Current: 2.8 Amps
97.1 CFM
42 dB
All Metal Construction
Ball bearing
Power: 3.9W
Dimensions: (120x120x38mm)
Connection: 2-wire 3-pin fan connection 12"

Okay after listening to fan videos on youtube all night, I'm really scared of getting any of these 38mm fans so far.

Now here's my question, if I get a lower speed 38mm, in particular the one I'm looking at pushes 68.9CFM, would it still have an advantage over a 25mm that would push maybe even more CFM?

So here's my match-up:

120mm x 38mm NMB/Panaflo FBA12G12L-1BX Low Speed Hydro Wave Bearing
Fan Size: 120mm x 120mm x 38mm
Speed: 1700 RPM
Airflow: 68.9 CFM
Noise: 30 dBA
Bearing: Hydro Wave bearing
Voltage: 12 V
Current: .24 Amp
Static Pressure: 0.13 in. H2O
Connector: 3pin


Versus


120mm Yate Loon D12SM-12
Fan Size: 120mm x 120mm x 25mm
Speed: 1650 +/- 10% RPM
Airflow: 70.5 CFM
Noise: 33.0 dBA
Bearing: Sleeve
Voltage: 12 V
Current: .3 Amp Max
Fan Life: 30,000 hours
Connector: 3pin and 4pin pass-through connector

Is the 38mm fan in this case the Panaflo going to be better than the Yate for cooling the heatsink with my shroud set-up? They both push about equal air according to ratings.

deadlysyn said:

Most mobo's can't handle more that 0.5 amps on the mobo headers, but it is said that Gigabyte boards can handle up to 1 amp.

Click to expand...

On both my server and main system, I run desktop grade Asus boards (server is AM2, main system is 775). On that, I run 2x 92mm ultra high speed don't-stick-your-finger-in-the-fan fans. I made a dual 3 pin to single 3 pin converter and run them off the main CPU fan header. They pull just over 2 amps. Haven't had an issue yet. They have really improved the amperage those headers can supply. I remember in an old p3 system, I burned the traces on the CPU fan header from running a 120mm fan with less than 0.3 amps.

CualvezAsteria said:

Now here's my question, if I get a lower speed 38mm, in particular the one I'm looking at pushes 68.9CFM, would it still have an advantage over a 25mm that would push maybe even more CFM?

Is the 38mm fan in this case the Panaflo going to be better than the Yate for cooling the heatsink with my shroud set-up? They both push about equal air according to ratings.

Click to expand...

Yes, the 38mm will be better because it has more pressure. Remember that free-flowing CFM's don't really mean much here where pressure makes a HUGE difference. I went from Yate Loon medium fans on my TRUE to Panaflo's, and saw a sizable temperature drop while being able to run them at a lower speed.

thideras said:

On both my server and main system, I run desktop grade Asus boards (server is AM2, main system is 775). On that, I run 2x 92mm ultra high speed don't-stick-your-finger-in-the-fan fans. I made a dual 3 pin to single 3 pin converter and run them off the main CPU fan header. They pull just over 2 amps. Haven't had an issue yet. They have really improved the amperage those headers can supply. I remember in an old p3 system, I burned the traces on the CPU fan header from running a 120mm fan with less than 0.3 amps.

Yes, the 38mm will be better because it has more pressure. Remember that free-flowing CFM's don't really mean much here where pressure makes a HUGE difference. I went from Yate Loon medium fans on my TRUE to Panaflo's, and saw a sizable temperature drop while being able to run them at a lower speed.

Click to expand...

On my old ASUS A8N5X (now wifes rig) I burned out a fan header using a 1.5 amp blower.?!?! The MOBO works fine other wise ....just the CPU fan headder no longer puts out power..It will still read the RPM from the Tach wire though.
SO... from that point on when ever I have a High amp (over .40A) I split the tach and power wires, power from PSU tach to MOBO. Works great! And While I have no physical proof...ever since I have started to do that I a have been able to sustain a more stable overclock...

As for the the whole Shroud thing.. I am a huge propionate of using shrouds...Too bad that using a shroud often takes up more space than available in a case....

My real query on this topic comes from the difference from Blower fans vs Axial fans in over all CFM/static pressure...

DUSNOETOS said:

On my old ASUS A8N5X (now wifes rig) I burned out a fan header using a 1.5 amp blower.?!?! The MOBO works fine other wise ....just the CPU fan headder no longer puts out power..It will still read the RPM from the Tach wire though.
SO... from that point on when ever I have a High amp (over .40A) I split the tach and power wires, power from PSU tach to MOBO. Works great! And While I have no physical proof...ever since I have started to do that I a have been able to sustain a more stable overclock...

As for the the whole Shroud thing.. I am a huge propionate of using shrouds...Too bad that using a shroud often takes up more space than available in a case....

My real query on this topic comes from the difference from Blower fans vs Axial fans in over all CFM/static pressure...

Click to expand...

Don't think I've ever seen someone mention having a Blower fan in their set-up. Do they even have them that are designed/sold for computer use?

What length shroud do you use? Bobnova mentioned 6" being optimal, however unrealistic. Though in my case I could actually do it, just don't have anymore 120mm fans to scrap.

CualvezAsteria said:

Don't think I've ever seen someone mention having a Blower fan in their set-up. Do they even have them that are designed/sold for computer use?

Click to expand...

Why does it have to be made for a computer? Just mod it!

thideras said:

Why does it have to be made for a computer? Just mod it!

Click to expand...

Speaking of mods, here's a particularly cool one: Electric Turbocharger on Computer Case
I'd love to see that somehow set-up to push through a TRUE.

I wonder if my duct explanation is good enough for a mini-article.


Blowers do away with the narrow band of air, for obvious reasons.

A fan/blower moves air based on three things (in no special order):
1) Number of blades.
2) Profile and length of blades.
3) RPM of blades.

We'll get back to the profile/length (BTW: We're talking about the length from the leading edge to the trailing edge, not the length from the hub to the tip), it's secondary to the other two.

Number of blades and RPM of blades are both crucially important, a fan with a given blade profile moves air based on how many blades go around in a given time. A low rpm fan with a lot of blades and a high rpm fan with a few blades both move the same amount of air.
There is an important difference in pressure between the two though, and it comes down to RPM.
A low rpm high blade count fan puts a little bit of energy into a lot of air, this works great for moving a lot of air, but when it comes time to generate pressure (when pushing it through a heatsink, for instance) the air doesn't have very much energy in it, and because of that can't generate much pressure. When the air hits the heatsink it stops, and rather then having the air behind it stack up and push it through (pressure!), it just backs up to the fan and leaks through it.
A high rpm low blade count fan puts a lot of energy into a smaller amount of air, the unrestricted result is the same (lots of flow), but when you put a restriction after it the fact that the air has more energy in it causes it to stack up and pressurize itself rather then simply backing up through the fan.

A squirrel cage blower can be looked at as a fan with a lot of blades, because that is exactly what it is, it's a classic example of a low rpm high blade count fan. They move a ton of air, but have lousy pressure (as a whole, there are ways around this as usual).

Now blade profile and length!
Profile is important obviously, but it's also highly time consuming to talk about and is best suited for aerodynamics classes in college. Also it is different for blowers and fans, so we're going to do the short version:

Fan: A steep pitch moves more air, requires more electrical power, and creates more pressure. A shallow pitch does the opposite, and is quieter generally speaking.

Blower: The curve is the important part, a straight blade (no curve) is most efficient, and is balanced between flow and pressure. A swept back blade blade (one that curves away from the direction the blower is rotating) generates more pressure, but less flow. A swept forward blade generates more flow, but less pressure. If you look at GPU fans they're all swept back violently, as they need a lot of pressure. If you look at a swamp cooler blower, those blades are swept forward as it needs flow, not pressure.

Now lastly, the length of the blade, this is important both for blowers and fans, and is really fairly simple.
The goal a fan is to put kinetic energy into the air, which forces it to move (or vice versa), the longer the air is touching the blade, the more energy can be put into the air. It's that simple.
A longer blade works better then a faster blade, as beyond a certain speed the air tears rather then flowing, and that renders the blade profile irrelevant, makes a lot of noise, and seriously decreases efficiency. You can get around it by making the blade a shallower angle, but now you've lost pressure and flow. A longer blade can also tolerate a higher RPM before the air tears, hence the really high RPM Delta fans being 38mm or more in depth, a deeper fan allows a longer blade at a decent profile, allowing more pressure and higher RPM.


EDIT:
Blowers don't have a dead spot like fans do, they use the effect that hurts airflow in fans to generate all their airflow in the first place.