FEATURED FANS 101: Going Scientific.

[mastrauckas]

Congratulations bluezero5!



Positive airflow is more cfm into case than out.

Yep, I understand that part.

I consider heatsink and GPU cfm, but only to be sure case is flowing more air than heatsink and GPU. If case does not than some of heatsink/GPU heated exhaust is going to be mixing with case air and heatsink / GPU intake air will be hotter than room ambient.

It should be flowing more air since my GPU has a exhaust so it shouldn't be mixing with case air.

Heatsink fans.. push/pull will be same as what one of the fans is rated at, not the combined total of both... I suppose if they are super high pressure there might be a slight increase in cfm, but minimal.

Yes, but I was more wondering do you count the heatsink CF toward intake or exhaust?

It's not really *intaking air* since it's not going to bring in new frest air.

However my heatsink fan is in push and it's pushing it toward the back exhaust fan. However, it's not exhausting heat but pushing toward the fan that does.

Or am I overthinking this?

HDD rack fans? Most cases don't have separate fans on HDD, only the intake fans in front so no.

You don't need exhaust fans if you have good flow design an low resistance exhaust vents. (remove exhaust grill like ehume's cases show)

So your saying it's better off removing the hdd cage if I don't need it and put the Reeven ColdWing 120mm Performance fan on the side?

Anyway, thanks for your time.

 

[doyll]

Yep, I understand that part.


It should be flowing more air since my GPU has a exhaust so it shouldn't be mixing with case air.

You need more CFM intake case than your GPU & CPU fans suck into GPU & CPU.

Yes, but I was more wondering do you count the heatsink CF toward intake or exhaust?

My intake fans are "cooler" fans. They have similar static pressure ratings to what my cooler fans are. So one of them moves as much air into case as cooling intake fans moves into cooler. I have 3x TY-140 74cfm case intake fans with 2x TY-140 (push/push) on CPU cooler and 1x Scythe Setsugen 120mm 45.5cfm. That's 222cfm in to supply cool air to 119.5cfm to coolers... Intake is restricted by filter and grill.. coolers are restricted by fins adn pipes. I think the grills and filters are at least as much if not more restriction than heatpipe radiators are.. and we also want more are moving thorough case than coolers are using. Picture a stream flowing along with a couple of boxes with screens on each end setting in it. The stream moves same speed through boxes as it does outside of boxes.. Our cases are the stream and out coolers are our screened boxes.

That make sense?

It's not really *intaking air* since it's not going to bring in new frest air.

Your GPU and CPU cooler are "intaking air" into themselves inside of case. If you don't have as much or more CFM coming into case they will eat their own exhaust/hot air. Your case intake fans have to *feed* enough air to your GPU & CPU coolers so the get all the air they want to *eat* without taking any air from anywhere else.

Picture the stream as case again. If boxes (coolers) take more than stream is flowing there so stream running any more.. and the boxes will start sucking the water coming out of them back into themselves.

CPU uses 80cfm + GPU uses 60cfm so you need at least 140cfm to keep up.. And that's in a perfect all air coming in goes to them. You also need enough additional cfm to force all the heated exhaust air coming out of GPU & CPU out of case.

Case exhaust fans can only assist in removing air. The air has to be coming into case before it can be removed.

Think stream and exhaust fans are pumps. If the stream isn't flowing enough water the pumps have ni water to suck.

However my heatsink fan is in push and it's pushing it toward the back exhaust fan. However, it's not exhausting heat but pushing toward the fan that does.

We have 2.. well, 3 *intake and exhaust* operations going on. 2 of them are inside of the case and are dependent on the case intake/exhaust to supply them with cool intake air and remove their hot exhaust air.

Or am I overthinking this?

Not unless I'm overthanking it too.

So your saying it's better off removing the hdd cage if I don't need it and put the Reeven ColdWing 120mm Performance fan on the side?

Anyway, thanks for your time.

Remove the hdd cage helps front intake fans because it removes resistance they are fighting.. Suck in front and push toward back. If hdd cage is not in their way they don't have to push against it.

Reeven ColdWing fans push lots of air.. and we need to be careful about moving air across the front of an intake fan too fast. Because if it is moving too fast it creates turbulence and the intake fan cannot suck in the air it wants. Think of it kinda like drinking out of drinking out of hose. If water is running out slow you can get a drink no problem, but if its running too fast you can't get a drink at all.

I know I've ran the same theme over and over in different ways. Hope that's okay.

No problem and thank you for thanking me. All I've done is confuse you

Edit: Instead of hijacking this thread any more than we already have how about you message me or better yet start a new thread. Maybe we can get a kind moderator to move all this so bluezero5's thread isn't a bodged up with my babble.

 

[bluezero5]

3 EVGA GTX 580 stock clock, will it be enough?, i am running it in push and pull

Just enough. For 3-way SLI u do need ard 480 rad to be quiet. Running push/pull will overcome most static pressure issues. Temps should look nice.

 

[Dragoon]

A good read. I have a question about CPU fans though.
I was looking into fans and came across several variations of fans and heatsinks.
Stock CPU fan sits on the processor and blows down onto the CPU and motherboard.
A lot of aftermarket products use heatsink and tubing and have the fan blowing a radiator that is sitting perpendicular to the motherboard and CPU.

Is there benefits or disadvantages of one over the other?

 

[doyll]

The down flow is much harder to get cool air to intake because the heated exhaust curls back off of mobo, RAM, GPU, caps, I/O housings, etc back up and often gets sucked right back into cooler. The heated air also mixes with other case air making case hotter.

Heat pipes rapidly move heat from CPU out to the fins. Some stock coolers have heat pipes too. Towers usually set facing the back where there is an exhaust vent and fan. They intake air from the from the front and exhuast it toward the exhaust fan. Not near as much potential for the heated exhaust air to mix with case air.

 

[Dragoon]

That makes logical sense. Thank you.

 

[trueblack]

very good read. actually helps my fan picking, thanks.

 

[Leegit]

Awesome read. Thanks.

 

[ehume]

The down flow is much harder to get cool air to intake because the heated exhaust curls back off of mobo, RAM, GPU, caps, I/O housings, etc back up and often gets sucked right back into cooler. The heated air also mixes with other case air making case hotter.

Heat pipes rapidly move heat from CPU out to the fins. Some stock coolers have heat pipes too. Towers usually set facing the back where there is an exhaust vent and fan. They intake air from the from the front and exhuast it toward the exhaust fan. Not near as much potential for the heated exhaust air to mix with case air.

I can vouch for this. In the past when I tested fans on heatsinks I put my ambient temp sensor 30cm in front, to sample representative air that could go into the intake. When I tested the Noctua NH-L12 (here) I tried positioning the temp sensor above the intake. I got ambient temps that were all over the map, even though the CPU temps were steady. I had to move my temp sensor completely away from the rig -- beneath it, actually -- to get a clean ambient temp measurement. Then my net temps over ambients became repeatable.

So, while I would consider a blow-down heatsink for a non-OC'd rig, I would only consider tower heatsinks or water for an OC'd rig.

 

[xsuperbgx]

Does anyone know the difference in bearing types? I know that durability/ longevity take quite a hit with certain types of bearings. I bought a bunch of the cheap yates. They have almost all quit working, add a little lube and they work better, but low quality nonetheless.

I have been thinking about doing the ap-29, ap-30, ap-31 gentle typhoons, but it seems that no one has them in stock. I was looking for higher static pressure fans that could handle a radiator and a filter. I think I will look for PWM 120x38's. I want them quiet for 24/7, but they need to crank up well for those rare times that I need to do a bit of benching. Plus a long life, any recommendations?

 

[ehume]

Does anyone know the difference in bearing types? I know that durability/ longevity take quite a hit with certain types of bearings. I bought a bunch of the cheap yates. They have almost all quit working, add a little lube and they work better, but low quality nonetheless.

I have been thinking about doing the ap-29, ap-30, ap-31 gentle typhoons, but it seems that no one has them in stock. I was looking for higher static pressure fans that could handle a radiator and a filter. I think I will look for PWM 120x38's. I want them quiet for 24/7, but they need to crank up well for those rare times that I need to do a bit of benching. Plus a long life, any recommendations?

The Yates we have access to have sleeve bearings. When you have one you don't need any more, take a small punch and tap one on the tip where it shows through the washer. You'll get an education.

GT's are high static pressure fans with ball bearings. You can convert the high highspeed versions to PWM.

Noctua NF-F12 PWM's are designed to produce relatively high static pressure and stay fairly quiet. NF-P12 PWM's are quieter, more an all-around fan; great for filters and heatsinks, but the F12's were designed for rads and heatsinks.

Cooler Master has come out with 4th generation fans -- basically sleeve bearing fans (anatomy lesson here) with POM bearings, a high durability plastic with low friction and high lifetimes. Sickle Flow X. Waiting on other high SP fans from them.

Look around at reviews, especially from Martin210 over at OCN.

Personally, I like San Ace fans for high-SP work.

 

[Conumdrum]

Yates have a limited lifespan. Especially if mounted in a horizontal way. I never had one fail, but only used them in vertical, and have one still on my wifes rig. 6 years? I myself used them for about 2-3 years before I got the AP-15 1850 (GT's) I used the medium Yates.

 

[xsuperbgx]

Sorry, to clarify, you think the sleeve bearing is ok, because you can always re-lubricate?
or would it be better to get double ball bearing or extra fluid dynamic bearing?
Some fans do not give hour rating, but can I assume that fans with certain bearing types have similar life span?
I do not know what all of these bearing types are. I assume the ball bearing is best, but I do not know what ex-fdb has that is different from the other hydrodynamic bearings.

noctua bearing says "SSO2 Bearing"
gentle typhoon says "Double Ball Bearing" 100,00hr mtbf
scythe kama flow 2 says "EX-FDB (Extra Fluid Dynamic Bearing)" 120,000mtbf
Swiftech Helix says "Z-Axis " 60,000
Yate loons i think are just a sleeve bearing
sickleflow r4 says"Long Life Sleeve" 50,000 hrs
akasa HDB (Hydro Dynamic) 50,000 hours

 

[ATMINSIDE]

Ball bearing fans are more expensive, but that's how I would go. They're typically quieter, as well as last longer.

 

[Conumdrum]

If horizontal, the sleeve rubs on the 'tube' at the bottom. It gets rough and no longer spins freely. That's the issue with sleeve bearings.

If it says sleeve bearings, it's a tube with a center shaft.

If it says magic rifle with peanut butter on top it's not a sleeve bearing.

There are many bits of tech magnified by marketers who get paid for awesome words.

The Helix for example..... It's not a sleeve.

 

[Mr.Scott]

Nobody mentioned the MagLev type.

 

[xsuperbgx]

After browsing for a couple hours, it seems that the noctua's are the better rated fans, but they are the most $ and they come in the horrible colors. So unless the ap-29, ap-30 etc can be found somewhere, I will have to get something different. For pwm fans on frozencpu, going by specs alone these or these look good to me depending on the space I have available. Reviews don't seem to favor that fan though, and there seem to be conflicting descriptions on teh bearing type, but perhaps that is a newer revision of the fan.

 

[Bobnova]

Many of the super long life non-ball-bearing things are sleeve bearings with creatively ground grooves in them to circulate the lube.
They're long lasting and quiet, where sleeve bearings are quiet but die and ball bearings are long lasting but noisy.

 

[NiHaoMike]

Nobody mentioned the MagLev type.

I have encountered one in an old NAS. The NAS has been operating for a few years 24/7 before it was discarded due to bad capacitors on the motherboard (which I have since fixed) and the fan still seems to work well.

There are also FDB (Fluid Dynamic Bearing) fans. They generally last a very long time, but beware that bearing life can be dramatically shortened if slowed down too much. Consult the datasheet for the minimum voltage. PWM type variable speed FDB fans will not let you slow them down too much.

I also didn't see any mention about the fan's inverter/motor drive. There are (for the most part) square wave inverters and sine wave inverters. Square wave inverters are used in almost every cheap fan and have a tendency to "buzz" due to the sharp transients. Sine wave inverters can be divided into two subtypes: resolver based drives and sensorless FOC (Field Oriented Control) drives, sometimes called DSP (Digital Signal Processor) drives.

Resolver based drives tend to lose performance at higher speeds due to the resolver filters and motor inductance causing the drive current to lag behind. Some fans compensate for that with phase advance which is a circuit inside the inverter chip that engages at high speed. Variable speed Sanyo and YS Tech fans (and older variable speed Nidec fans, though those PWMed a 5 step waveform rather than a sine wave) use resolver based drives, with some higher end models featuring phase advance.

Sensorless drives do away with the resolver altogether and rely on sensing the voltages and currents coming back from the motor (back EMF) in order to properly time the drive signals. They digitize the back EMF and use what is essentially a tiny computer to interpret the rotor position at any given instant, then synthesize the waveforms that actually control the power MOSFETs in the inverter. The digital control gives a lot of flexibility, such as transitioning to trapezoidal waves and/or applying field weakening at high speed in order to get a little more power out of the motor. Variable speed Delta and (newer) Nidec fans use sensorless drives.

Here are some typical waveforms for a sensorless drive, top trace being phase to ground and bottom trace being neutral to ground:

You'll notice that even with the phase to ground looking like a trapezoidal wave, the phase to neutral (what the fan motor actually sees) is still very close to a sine wave.

Also do not confuse the PWM speed control signal with the PWM the inverter uses to synthesize sine waves. PWM was chosen because it's very easy to interpret in an analog circuit and very easy to generate with a digital circuit. The PWM the inverter uses is based on an oscillator internal to the controller chip and is generally just above the audible range as that gives great efficiency without making an audible whine.

 

[Conumdrum]

I have encountered one in an old NAS. The NAS has been operating for a few years 24/7 before it was discarded due to bad capacitors on the motherboard (which I have since fixed) and the fan still seems to work well.

There are also FDB (Fluid Dynamic Bearing) fans. They generally last a very long time, but beware that bearing life can be dramatically shortened if slowed down too much. Consult the datasheet for the minimum voltage.

Deleted the stuff that don't matter.

NiHaloMike, I bet those fans were a vertical orientation? Not horizontal like PC cases usually can have?