• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

FEATURED Building PWM Controller for 4 wires PWM fan

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Brutal, that's why I also ordered some wire fan guards with the Big Beast Denki. I refuse to run it unless I have fan guards on it on the suction side especially. You won't believe the amount of suction the fan has, believe me. Anything you have messed with until now is just a tinker toy piece of trash compared to these fans.


You have obviously never seen a Bus Condenser or Evaporator Motor. :D There is a large cover that is about 2'X6' and 2-3mm thick steel that is held on by the suction from two of the condenser motors.

I was considering one of the Driver's Blower fans, but I imagine that it vs. this Beast would be a draw, and the otherone pulls way more power, and its 24V and not 12.

Fastest Fan in the World, Moooo Ha ha ha! I will order a fan grill for it, just gonna test it out tonight. :D wish me luck, I will make sure I count all my fingers before going to bed.
 
Last edited:
Anything you have messed with until now is just a tinker toy piece of trash compared to these fans.

LOL ... :D :D

My Denki Long Life 4800 rpm with grill, even its not comparable with yours, they're still scary to me.
View attachment 82309


Actually what I'm interested is to see how this beast performs when hooked up on mobo's PWM, since it is designed at 16Khz PWM frequency while the mobo PWM signal is the standard 25Khz.


Brutal, if you mobo is capable of bringing down only to 10% pwm duty cycle, that will be around 1400 to 1500 rpm, to have lower than 1000 rpm, looks like the only way is using controller.

Another interesting data from this chart is, once the the duty cycle reached approx 95%, it will be at it's full speed at 6000 rpm.
So I believe the PWM controller circuit is capable of providing full sweep from silent up to max 6000 rpm for this beast. :D

View attachment 82310


@Mud, would you please change this thread title to "Building PWM Fan Controller for 4 wires PWM fan", of course with Brutal's permisson.
 
Last edited:
Yes please, Mud.

Yeah I figure I will probably end up building the controller.

1st, it also gives a good evaluation tool. I may end up later also getting a Flow Meter, so I can calculate CFM. Then When I test Heatsinks, I can use this and actually give details by RPM/CFM.

And if what you say is true, depending on the design of the fan and how it scales noise wise, I may need it to go lower than 1400 RPMs.

I am pretty sure I will still be building this controller.
 
Thread title changed for you, Brutal.

Just be careful around that fan when you fire it up and make sure nothing is close around the suction side of the fan. I really couldn't believe what that beast of mine was trying to suck in. Be sure and watch your shirt.:rock:
 
Thanks Mud ! :thup:

Hopefully this will attract more viewers or better more participants. :)
 
Ok, running prime 95 right now. Gonna do a 15 min. run. So far after 2 min. temps are 10 degrees lower than the two Panaflos in Push/Pull.

CPU fan is running at 5750 RPMs.

Running @ 1.4 Vcore, 181X22 - 4000 MHz

Ambient temp is 23 degrees Celcius.

Temperatures dance between 66 degrees and 72 degrees.

No doubt the fan is loud, but thats running at full speed. After I run Prime, I am going to work on its scaling using Fan Q Control.

10 min. Temps are still between 66-72 degrees. Highest a core went up was to 79, but it looks like it was a glitch.

Ok, Initial Screenshot.

Screenie%20Beast.PNG


Ok, ignore the 79 degrees, that was a glitch. The temps are actually what your seeing in the screenshot. This was taken while running Prime at 15 min. with 1.4V Vcore and 4000 MHz. I dropped about 10 degrees off of my temps that the Panaflos were producing, which is about 30 degrees lower than before I put the Panaflos on.

Ok, for noise. The PWM allows me to go from 20% to 100%. Idle RPMs are 2350ish and Full speed is 5700 RPMs.

I will most likely be making a controller although this fan is perfectly tolerable at 2300 RPMs, but since I am aiming for 1000 RPMs, I will build a PWM Controller. I really don't even see where you can buy a good Fan Controller, so this will be my project.

Edit: By going into bios, I can lower the PWM to 0% so the fan runs at 0 RPMs. This is obviously silent, and the Venom X idles at about 45 degrees with no fan. Not ideal, but just a FYI. I will still be building the fan controller. The Fan will kick in now and then from 500-600 RPMs. Above 1000 RPMs it becomes audible. I also have the option in bios to limit the upper fan speed percentage, so I can just go in when I get ready to bench and raise it to 100% but for 24/7 I can set it between 0% and 60% which is about 3500 RPMs which is about how loud the Panaflos got at max RPMs.
 
Last edited:
I would say sound wise its tolerable. Since I can turn it down to 0, obviously its not noisey. Running at 2400 RPMs is not deafening either. I think though the fan controller is going to be the way to go, because even despite being able to control the fans, sleep mode screws it up.

I like it, but it will take some getting used to when it fires up.

I would say at comparable RPMs (up to 2500) it is about as loud as a Panaflo. Once it goes past that, of course its going to be loud. Not as loud as the typical vacuum cleaner. Also a large portion of the noise is wind noise and not fan noise. It is pretty clean sounding, but not for the weak at heart. As long as you can control it with a controller or bios, it is perfectly acceptable.
 
Last edited:
Sounds like fun.....I really wish I had a controller for my Delta when benching :(
That thing is crazy loud and it's 100%....100% of the time :eek: I much prefer benching with my water cooling loop compared to my Megahalems with the Delta :p
 
Ok, ordered TLC556 at Digikey. When it gets here I will make a trip to radioshack and pick out some circuit board supplies/Potentiometer/Capacitors/Resistor/Diodes.

bing, just for clarification:

C1 - 470 pF Capacitor

C2 - 0.01 Micro Farad Capacitor

C3 - 10 Micro Farad Capacitor

C4 - 0.1 Micro Farad Capacitor

D1 & D2 are Diodes with the part number 1N4148?

R1 - 1K resistor

R2 - 10K resistor

P1 - 100K Potentiometer

A PCB ? recommended size, quality? Do they make a build kit somewhere?

4 Pin PWM Connector - I am thinking 6 of these for long term.
 
Yours will be better:



If that was directed toward me, the problem is my fans are not PWM ;)

Yes mine will be better, because I made it!

Yours isn't PWM. Uh, Ok thats why you don't like it. Oh yeah, being controlled PWM is the Bee's knees.

While I like the automatic control, I think that ultimately, the selectable controll will be nicer. I think its kind of the same as trying to have fan controllers automatically adjust by temp, it never works out well because it doesn't know what our noise tolerance is. And fans have different noises at different speeds, some times 2400 RPMs is nerve grating, but 2450 RPM is just fine.
 
Hey, that is a really good performance, 10 degrees reduction from double panaflos ! Good for benching session ! :thup:

Looks like you like this fan, don't you ? Great to hear that ! :D


Ok, ordered TLC556 at Digikey. When it gets here I will make a trip to radioshack and pick out some circuit board supplies/Potentiometer/Capacitors/Resistor/Diodes.

bing, just for clarification:

C1 - 470 pF Capacitor

C2 - 0.01 Micro Farad Capacitor

C3 - 10 Micro Farad Capacitor

C4 - 0.1 Micro Farad Capacitor

D1 & D2 are Diodes with the part number 1N4148?

R1 - 1K resistor

R2 - 10K resistor

P1 - 100K Potentiometer

A PCB ? recommended size, quality? Do they make a build kit somewhere?

4 Pin PWM Connector - I am thinking 6 of these for long term.

Dont forget the 680pF capacitor, this will make the pwm freq at 16Khz at that fan's specification, though its working at your mobo pwm at 25khz, its worth to try if at 16Khz makes any difference, especially at the noise profile, anyway this cap is dirt cheap and its easy to swap them.

I recommend you do this circuit 1st at breadboard level, once its working , then transfer it to PCB, cause at PCB level, trouble shooting is harder.

Also at breadboard level, if you have trouble, you just take a shot at it and share it here so all of us can help you on the circuit just in case you missed some connections or wrong cables routing at the circuit.

Some examples like HERE or some corrections HERE or circuit is working once correted HERE. :)

Another suggestion is, buy everything at least 2 complete sets or even more, since these are relatively cheap components, this should iron out defective components if any. Trouble shooting DIY circuit without adequate measurement tools when one of the component gone bad is a nightmare, especially at the semiconductor or that IC.

Great, happy to see you've started to build the controller, the fun part begins... :D
 
Last edited:
Sounds like fun.....I really wish I had a controller for my Delta when benching :(
That thing is crazy loud and it's 100%....100% of the time :eek: I much prefer benching with my water cooling loop compared to my Megahalems with the Delta :p

Miah, I think you should consider to invest on good and beast like PWM fans like this one, equipped with controller, imo they're really handy at benching session where you can rig up multiple of these to blow off the condensation.

Anyway, remember, only single controller needed to tame multiple beasts like this one. ;)
 
Last edited:
bing, I have been looking around at fan controllers and I am somewhat confused.

There are fan and DC motor controllers that say they function with PWM. But they are not 3 wire fans or motors that they are controlling. Are they creating a pulse width on the 5-12V Motor in order to obtain PWM? If so, how does this compare to 3 - wire PWM?

So far I see a PWM Motor/Fan as having a 12V power source and a ground source. The motor is switched on and off rapidly with a 5V digital square wave signal. Does the PWM on a two wire system function the same? But the 12V source is switched on and off with a 12V square wave signal? I would think this would create alot of heat due to the constant 12V being applied and removed.

Then there is the Hz frequency, which is the time from peak to peak, correct? This means that the pulse of the signal is faster at a lower frequency? I.e. the 16Hz freq. you were recommending to me. Basically, means it pulses the 5V switching signal faster, and like bumps in the road, will help smooth out the fan noise. Or is a higher frequency when the bumps get closer together, but the peaks come closer and closer together as the frequency increases. Is higher frequency more desirable, or lower? Did you only recommend 16Hz because that is what is on the Data sheet or is a higher freq. actually more desirable?

I also want to be able to switch from 0-100% because, the lowest noise would obviously be 0, but if the fan stops, the Heat sink will only be passively cooled. Also fans develop different noises at different RPMs, so maybe a 5% signal would be a low RPM, but there may be a whine associated with running it at 5% that might not be there at 7%.

If you could give me some clarification between 2-wire PWM and 3-wire PWM, I would appreciate it.
 
Last edited:
Never, mind. Found it.

There are 2 approaches to using PWM for fan speed control:
1st approach uses 2 or 3-wire fan (3rd wire is speed or alarm sensor). The fan Ground or
Power line is controlled by N-MOSFET or P-MOSFET.
Pulse width modulating (PWMing) the fan directly involves turning the fan’s power
supply ON and OFF at a fixed frequency. Duty cycle adjustments are then used to control
the speed of the fan. Choosing the appropriate frequency for this method can be
somewhat tricky. If the frequency of the PWM signal is too slow, the fan’s speed will
noticeably oscillate within PWM cycle resulting in undesirable fan noise. The frequency
can be also too high. But PWMing the fan and therefore the internal commutation
electronics too quickly can cause the internal commutation electronics to cease
functioning properly. Remember these electronics were designed to run on anything but
DC supplies. In the case of a 3-wire fan, powering the fan supply voltage up and down
will affect the accuracy of the sensors because the speed and alarm circuitry is also being
powered up and down rendering the speed and alarm sensors useless.

A better approach to using PWM speed control is to use a 4-wire PWM fan.
The PWM and the tachometer line are connected to the PWM controller, leaving the fan
power and ground lines uninterrupted, as shown in figure 3. In this way, the circuit inside
the fan is working normally, sending a valid speed signal and accepting PWM control to
change motor speed accordingly. As a result, a simple automatic closed-loop speed
control system is formed.
With closed-loop speed control, the tolerance of the fan PWM duty cycle vs. the speed
curve can be very wide. The controller can command the fan to achieve a desired speed
(RPM) goal by adjusting the PWM duty cycle. If the speed is below the goal, the PWM
duty cycle will be increased, and vice versa. The speed goal will also be maintained when
there is voltage variation or load variation on the fan.
 
PWM%20Circuit.PNG


Picked up the components, now working on the layout.

DSCN0555.JPG


DSCN0557.JPG


DSCN0558.JPG


DSCN0561.JPG


DSCN0562.JPG


I am pretty sure the IC I got from Radioshack won't work, but I already have the right one from Digikey on order. I also have an IC Socket instead of soldering it directly to the board. In case I have to change it out later. :)

Also, I couldn't find a 680 micro Farad, so I will use the 470 for now and order a 680.
 
Last edited:
Back