120mm Fan Testing on an MCR120 Radiator Round 6

[Manicdan]

^ +1 I'd be interested in this too!
Manic I don't suppose you know the answer to any of the questions in the post prior to yours?

Thanks all/night.

i knida know the answers

basically you can easily run the 4th wire on the motherobard IF you have 4pin PWM connectors. mine only has one, but there are some with like 7.

his idea is to try and run them all in sync, so if you have to read a rpm, just pick one and plug it in and ignore the rest. if you want to read the rpms of them all, you still can as long as each has its own plug

if you have extra PWM, you can easily use them up for more dynamic controls. but that might not really matter unless you have them for different needs (like intake fans might want to trigger max speed at 40c, and exhaust at 50c)

[Trice]

Lots of information here! Almost too much for me Is it possible that someone can lead me in the correct direction for my fan choice? I had 2 San Ace 109R1212H1011 which I put on a fan controller and ran them at 5v I think? I now have a TA 120.4 rad, and need 2 more of these, or should I go for a different solution?

[OldChap]

How much do you need the ability to run fans at high speed ?? If you always run at ~5v then you are working in the 25cfm range at around 45decibels and a noise quality rating of 6

If you stick with 25cfm then the gt ap15 is running 6 decibels quieter with a sound quality of 7-8 but compared to your current fans maximum being 61 cfm at 60 decibels and qual. of 7 the the ap15 tops out at 35cfm very quiet at about 45decibels (similar to your current fans at 5v) and quality 7-8

Going the other route and having fans capable of high cfm the quietest at any comparable cfm is the Delta BUT the slowest it runs is around 35 cfm at which point it is 3decibels louder than the ap15 running full speed but mostly 5decibels or more quiet than the rest including the fans you have. It gets a quality mark of 6-7 when running slower but it can move a lot more air than your current fan at top speed. at the same noise output as your current fan running max the delta is 9-10 cfm more.

As things stand with the tests done so far for a fan up to 35cfm buy a gt ap15, for a fan 35cfm and up buy a delta. for something that overlaps these speeds buy two more of what you have.

BUT there is more testing to be done and maybe we (well Martin actually) will find the elusive fan that will do both reasonably fast and quite slow quietly

[jalyst]

AP-15's are apparently still the best noise-to-CFM ratio....
There's now some GT's with much higher RPMs in the form of AP-29,30,31, but these aren't PWM capable, so it remains to be seen how well they under-volt.
If you don't care so much about noise and want some real blowers: san ace, delta, or there's even some 6000RPM GT's about (but must import so $).

BUT there is more testing to be done and maybe we (well Martin actually) will find the elusive fan that will do both reasonably fast and quite slow quietly

This is what I'm hoping the new, higher RPM (PWM-aware), GT fans will be capable of.
Large variability of RPM, but excellent noise-to-CFM ratio maintained across that range.
It just sux that Scythe aren't making these versions available...

[Trice]

How much do you need the ability to run fans at high speed ?? If you always run at ~5v then you are working in the 25cfm range at around 45decibels and a noise quality rating of 6

If you stick with 25cfm then the gt ap15 is running 6 decibels quieter with a sound quality of 7-8 but compared to your current fans maximum being 61 cfm at 60 decibels and qual. of 7 the the ap15 tops out at 35cfm very quiet at about 45decibels (similar to your current fans at 5v) and quality 7-8

Going the other route and having fans capable of high cfm the quietest at any comparable cfm is the Delta BUT the slowest it runs is around 35 cfm at which point it is 3decibels louder than the ap15 running full speed but mostly 5decibels or more quiet than the rest including the fans you have. It gets a quality mark of 6-7 when running slower but it can move a lot more air than your current fan at top speed. at the same noise output as your current fan running max the delta is 9-10 cfm more.

As things stand with the tests done so far for a fan up to 35cfm buy a gt ap15, for a fan 35cfm and up buy a delta. for something that overlaps these speeds buy two more of what you have.

BUT there is more testing to be done and maybe we (well Martin actually) will find the elusive fan that will do both reasonably fast and quite slow quietly

Awesome info! I actually never ran my San Aces at full tilt and always ran them at the lowest possible voltage before they wouldnt spin anymore. Sounds like the AP15s are the fans I need to get. Would I need a fan controller for these as well?

[OldChap]

I would say probably not if you have enough fan headers on your MB that are going to ramp the voltage up under load. Actually, considering the fact that at full speed the Gentle Typhoon ap15 is a very similar volume to your current fans at minimum you could choose to leave them at full speed all the time.

If you do run a controller then running slower, even though you will have more fans, will seem quieter and the main advantage is You will control them as and when you want.

[the finisher]

@ OldChap, nice sumation of results so far

@ Trice, the 109R1212H1011 is discontinued, I haven't been able to purchase new in the US. I had to get a 9G1212H101's, I use the M speed version as I don't need 2600 rpm.
You could probably find 2 more around here, post a thread.


I realy want a higher speed GT, I think I have a higher noise threshold than many. Or maybe I'm just going deaf.

I have seen that Nidec makes a GT@2150, this is probably my sweet spot. Should be very good from 1500 and up. Since Scythe is selling AP15' like hot cakes,they probably have no insentives to offer a AP16

Since Nidec won't give us a 140 GT, we should hammer Scythe US Marketing Dept. for faster GT 120's, that they can easily get now

[Martinm210]

FYI, I got several more tests done last night, although I had a little light rain. I'll see if the check in test was acceptable. Travelling out of town for the day, but if the check in is good, I'll start processing some new results tomorrow. Subjectively, it seems like the older Deltas were louder than the newer rev 3 models and the AFB1212HE rev 3 didn't want to start below 9v..

The only real interesting thing with PWM that jumped out at me was the slipstream ran from 500-2000 rpm which is quite a range and better that voltage control for range...noise I think was about the same per my ear.

[the finisher]

I think you might've missed some of the posts earlier in this thread.
AP-29, 30, 31 are available soon, alas they're not PWM capable

OOh, thats good news I miss alot, live on the road mostly, work constantly:/

From what I've seen these industrial fan manufacturer's mostly do PWM on very high speed fans.

I should dig through Sanyo Denki's 120 PWM fans, might be a gem there

[Martinm210]

Yeah, unfortunately thbey are not showing the 2150 GT model as one of the new options..nor PWM, but apparently that 5000 RPM version will PWM control down to 1000rpm.

Regarding using 3pin fans in a MB controlled system. I think it's possible looking at the 3 pin pwm controller I built from pcbheavens design. If I understand it right, it basically takes a PWM signal from the 555ic and uses a power mosfet to amplify it. Seems logical you could do the same thing to a motherboard pwm signal. Just need a mini circuit that draws 12v from a molex connector, and convert the MB signal to 3 pin PWM. That might be just the ticket in converting MB temperature control over to auto adjusting your favorite 3 pin fan.

I'm just not that comfortable with the electronic details to design the circuit myself...

[Trice]

I would say probably not if you have enough fan headers on your MB that are going to ramp the voltage up under load. Actually, considering the fact that at full speed the Gentle Typhoon ap15 is a very similar volume to your current fans at minimum you could choose to leave them at full speed all the time.

If you do run a controller then running slower, even though you will have more fans, will seem quieter and the main advantage is You will control them as and when you want.

So do you think 4 Scythes are best or just get 2 more San Aces and regulate them?

[Martinm210]

So do you think 4 Scythes are best or just get 2 more San Aces and regulate them?

I'd suggest pulling up the video on each one, and compare the audio at like air flow levels. I think everyone has a slightly different preference on noise and noise level is only half of the story. GTs are a different sound, very low in noise level, but I also know of many people who really like the san aces at lower volts.

I personally don't like any of the 38mm fans at really low volts, but that just me. I'm also not entirly thrilled with the sound quality of the GTs. The noise they produce is absent of you typical wind like noise (which can be good or bad). I think it's the key to the lower noise levels, but it also makes the motor whine a bit more obvious since there really isn't much swooshing air type noise.

I'd pick the GTs myself over the San Aces..

[Martinm210]

THIS TESTING ROUND 6 IS CLOSED!!

I'm not quite sure what happened when I switched between the 120 rad to the 140 rad, and back again, but the results from this batch did not pass my 1dbA limit.

It was a good run, but this batch and all future test will not be comparable to previous tests...sorry guys. I was still getting starting dbA readings at 31.0...but something else has changed when I swapped rad templates and it's throwing the results off..

Anyhow, I figured this is a good place to cut off this testing round. I ran about 8 fans through this morning and I don't want to waste them, but I also don't want to include them if it's off this far.

Here is the failed check in....this next batch can not be compared.

[Church]

Martin: out of curiosity, can it be that something changed for fan used for baseline after that many remounts? Or anemometer started showing different results? (as in to cross-check against old results with another fan), if not fan failing, in such cases why not adjust results adding or distracting delta of baseline/check-in fan results?

[jalyst]

Yeah, unfortunately thbey are not showing the 2150 GT model as one of the new options..nor PWM, but apparently that 5000 RPM version will PWM control down to 1000rpm.

Not sure there is a 5k RPM...
This monster & the 5400 (pwm version) can get down to 1k RPM, but alas one has to import all these higher-end models

Regarding using 3pin fans in a MB controlled system. I think it's possible looking at the 3 pin pwm controller I built from pcbheavens design. If I understand it right, it basically takes a PWM signal from the 555ic and uses a power mosfet to amplify it. Seems logical you could do the same thing to a motherboard pwm signal. Just need a mini circuit that draws 12v from a molex connector, and convert the MB signal to 3 pin PWM. That might be just the ticket in converting MB temperature control over to auto adjusting your favorite 3 pin fan.

3-pin PWM, how is that possible, doesn't the PWM spec. call for a fourth pin?

So you're saying it's possible to integrate an add-on circuit which essentially allows your MB to control PWM (4-pin) fans, if your MB doesn't have any fan-headers that are PWM (4-pin) capable?

Could this circuit allow for higher amp fans, which would otherwise fry your MB if plugged directly into it?

[Martinm210]

Check in

I don't tthink it was the fan wearing, results were better, not worse. Also, a wearing fan will show high dbA throught, this was about the same to 2dbA better, although it does point the finger at the anemometer. I've had to change batteries a few times now, and I did again in this last batch. The face template spun around at a slighly different angle from before, so maybe the anemometer temp sensor is reading slightly different. Or ??

All I know is I got up way too early and was way too tired after the batch to throw the results away and start over and likely get the same bad check in.

I know ideally we would like everything comparable, but its just not practical after so many tests. I also don't believe in correcting results, that just invites the possibility of correction error and the video..I can't correct.

The nice thing about smaller batches is the group comparison it a lot easier to read, and I don't have to stress about the check in.

PWM
Yes, I think it can be done wih just one 4 pin MB header. If I understand the circuit, the power mosfet just sort of works like a relay (A switching power supply). Each MB PWM pulse simply activates the mosfet gate drawing power from the other power source(molex), so the power limit is really a function of the mosfet. I think with the right circuit and perhaps multiple mosfet and heatsinks on those mosfets, you could probably thermally drive as many fans as you want from a single MB header. It would only use the pwm signal to activate the power mosfet..the power would draw from the molex/power supply. You can do this same thing now using 4 pin pwm fans, you just have to wire the red and black over to the molex (same thing as the new swiftech PWM pumps).

Soo..yes a PWM fan is 4pin (power,ground,rpm,pwm sensor), but you can make a 3pin pwm fan controller..there is a downside though...no more RPM readout.

But, that's just my limited understanding, and I could be very wrong...just something I've been thinking about since I built the 3 pin controller. I just really want to explore a thermally controlled pwm setup and it would be great if I could just reuse my 3pin fans. It just doesn't make sense to run fans at full speed all the time or manually controlled. Let the motherboard do the work adjusting. Sure you can buy fan controllers that do this too, but I'd like to make my own as a fun project..

[Church]

Martinm210: i just guessed that fan maybe culprit because it's not as simple +#% cfm, but a little bit different looking curve as well.

[Martinm210]

Martinm210: i just guessed that fan maybe culprit because it's not as simple +#% cfm, but a little bit different looking curve as well.

You could be right, maybe I grabbed tthe wrong SM fan? I do have a few lying around, and I've already made that mistake using an SL.

If that the reason I better mark that fan yellow or something..

[skinnee]

I've been marking fans and other bits for a while now, helps reduce those 'Son of a...' moments.

[jalyst]

Hey folks,
Just a few older posts missed by some of you, plus some newer ones I'm responding to.
Bing, if you got a minute could you please respond to these two (below):

Why not connect each PWM fan to each of the PWM fan-headers instead of just one PWM header?
I have 2x PWM headers and will probably only have 2x PWM fans, the rest will be AP-15 & 14, which I believe are not PWM.
That way each fan gets power & passes/receives RPM/PWM signals independently of one another.
Of course any fan/s that are too powerful for the mobo's power regulation would need their +/- wires rigged to a molex connector.
And then be connected directly to the PSU, but otherwise....

Wouldn't it be better to have discrete RPM sensing for each individual fan?
e.g. if you have 4 fans on the rad, but all 4 fans are represented by only one RPM sensor, then you wouldn't always know if one fan dies!
I guess that's just the limitation of not having enough headers, sigh, perhaps longer-term I'll need a multi-channel fan controller.

You can see in a video I made for bing....the PWM version runs at about 2400rpm at the lowest duty cycle with Gigabytes CPU header....but with a proper PWM controller, at 0% duty cycle, the PWM version will run at about 1000PRM.....so, if you want the best, petition Scythe for the PWM version

Does this mean that if I try to run 1-or-more 5400RPM PWM GT's, off 1-or-more of my MB's PWM fan-headers...
That I also won't get decent PWM control w/a duty cycle at 0%, & hence won't be able to lower it to 1000RPM?

Regarding using 3pin fans in a MB controlled system. I think it's possible looking at the 3 pin pwm controller I built from pcbheavens design. If I understand it right, it basically takes a PWM signal from the 555ic and uses a power mosfet to amplify it. Seems logical you could do the same thing to a motherboard pwm signal. Just need a mini circuit that draws 12v from a molex connector, and convert the MB signal to 3 pin PWM. That might be just the ticket in converting MB temperature control over to auto adjusting your favorite 3 pin fan.

Aren't motherboards with 3-pin headers essentially linear voltage controllers?
So any fans with 3-pins, you could just sit em on those headers.
Then you could up/down speed in a similar fashion to PWM, just not as efficiently?
And you could still get RPM feedback etc...

PWM
Yes, I think it can be done wih just one 4 pin MB header. If I understand the circuit, the power mosfet just sort of works like a relay (A switching power supply). Each MB PWM pulse simply activates the mosfet gate drawing power from the other power source(molex), so the power limit is really a function of the mosfet. I think with the right circuit and perhaps multiple mosfet and heatsinks on those mosfets, you could probably thermally drive as many fans as you want from a single MB header. It would only use the pwm signal to activate the power mosfet..the power would draw from the molex/power supply. You can do this same thing now using 4 pin pwm fans, you just have to wire the red and black over to the molex (same thing as the new swiftech PWM pumps).

So you're saying that with this circuit...
One can have higher wattage PWM fans coming-off any 4-pin PWM header it's rigged to, than would normally be possible?
(I got 2x 3-pin and 2x PWM 4-pin)
Why not just connect the higher wattage fans directly to the PSU, does it get too messy?

Soo..yes a PWM fan is 4pin (power,ground,rpm,pwm sensor), but you can make a 3pin pwm fan controller..there is a downside though...no more RPM readout.
I just really want to explore a thermally controlled pwm setup and it would be great if I could just reuse my 3pin fans. It just doesn't make sense to run fans at full speed all the time or manually controlled. Let the motherboard do the work adjusting. Sure you can buy fan controllers that do this too, but I'd like to make my own as a fun project..

I'd love if there was a way to make all headers on my mobo PWM-capable, but in doing so retain compatibility with any 3-pin fans.
And hence have the choice of whether I want to keep RPM monitoring for those fans, or drop it so they can have PWM!?

Thanks all.

[Martinm210]

I'm not sure about the latest motherboards(my gear is over two years old now), but my DFI X48 only has static 12v feed out of the 3pin headers and one 4pin PWM header. The only difference between the two is that fourth PWM pin and I can control one PWM signal on temperature limits of the CPU. The power limit on that and the other headers is limited.

The whole idea is to take that one signal and split it out for many radiator fans, but instead of drawing power there, do so from the PSU. You can do that easily with 4pin fans now, just basic wiring.

But, to control 3pin fans, you would have to build a circuit to convert the signal to a 3 pin switching power supply type controller, if it's even possible..I'm just guessing. But again the idea is to automatically control the fans. I just know that my 3pin pwm controller I built loses rpm signal (gaps in power mess up the signal).

For simple manual feed, connecting directly to the power supply is best. And there are good fan controllers out there too, but I would recommend the voltage control types over PWM from my limited testing so far.

[Church]

Martinm210: For 3-pin you mean something like mCubed's FanAmp which regulates upto 20W (from PSU) similar to voltage it's fed from motherboard fan header? (and i might be wrong, but IIRC it passed rpm signal back to it's regulating 3-pin aswell)

[Martinm210]

Martinm210: For 3-pin you mean something like mCubed's FanAmp which regulates upto 20W (from PSU) similar to voltage it's fed from motherboard fan header? (and i might be wrong, but IIRC it passed rpm signal back to it's regulating 3-pin aswell)

I'll check into that one, sounds like the same thing..

[jalyst]

<SNIP

Thanks Martin!
I've PM'd my remaining Qns/responses, so as not to overrun this thread any more.

[jalyst]

Martinm210: For 3-pin you mean something like mCubed's FanAmp which regulates upto 20W (from PSU) similar to voltage it's fed from motherboard fan header? (and i might be wrong, but IIRC it passed rpm signal back to it's regulating 3-pin aswell)

Nice,

It'd be sweet if my MB could auto-speed-adjust 3pin fans daisy-chained off this*, based on the MB's temp readings!
Instead of relying on the device itself, and it's probes....

Also I guess you'd only be able to send RPM data from one of the fans daisy-chained off it?

*alas it's not using PWM, but then I guess you'd have to give up RPM!