So you want PWM control of your 3-pin fan?

[Vario]

All I have is a digital multimeter. Could you suggest a way to test it? I could use a battery for the 12v source but not sure where to get a pwm input, maybe my dead h100i?

 

[Lazzer408]

Connect it to your 12v source. PWM lead should have ~5v on it. With PWM lead disconnected, the fan should run full speed. With the PWM lead grounded, the fan will stop.

I'll look over your pics tonight.

Velict's diagram looks fine other then the diode's anode connection doesn't have a dot on it.

 

[Vario]

I plugged in a 12v trickle charger and was able to run a single fan on it. Haven't done any more tests at the moment. I think the fan was running full speed. I'll try messing with the PWM lead.

 

[Vario]

My DMM measured a bit over 4 volts (4.10ish) on the PWM lead. When grounded, the fan turned off. I also ran this with a few Deltas and San Aces, 8 fans total, they all seemed to run full bore with no PWM connected.

I haven't measured the voltage of the trickle charger.

I am pretty excited.

 

[Lazzer408]

Sounds like it's working fine. Good job.

 

[Vario]

Sounds like it's working fine. Good job.

Thanks for the help, I am waiting for a 4 pin male-male fan cable to hook it up.

 

[OneMoar]

yall realize this thread is from 2010 right ?

 

[Lazzer408]

What's that have to do with anything?

 

[OneMoar]

a little something called thread necromancy

 

[Lazzer408]

Which hurts who in what way? Information doesn't have a "sell-by" date. If people are still benefiting from the thread, it's not dead.

 

[Vario]

I am glad the thread was brought to the front. Its very helpful, learned a lot.

I hope to solve some fan control issues I am having with this circuit.

Update:

I am running it now. My San Ace 109R1212H1011 will barely run at idle, and only if they are alone in the circuit. If I run two high powered fans they turn off. A Delta WFB1212HH won't spin at all. A Koolance 12038HBK will run very low voltage but at higher pwm, they start to sing in a high pitched whine that sounds terrible.

I am using the parts list shown here:
http://www.overclock.net/t/1505404/guide-convert-3-pin-fan-to-pwm-56k-warning-courtesy-of-lazzer408

edit: if it runs bursts of 12v, and it isn't spinning up, then maybe the PWM circuit for the motherboard isn't compatible with the power demands of these fans because it doesn't switch on frequently enough?

edit2: Different mosfet?

should i try constructing your minimum speed version?
http://www.techpowerup.com/forums/t...ontrol-of-your-3-pin-fan.115752/#post-1779591 Does that switch more frequently?

 

[Lazzer408]

Are you using a logic-level mosfet?

 

[Vario]

Yes I am using an N Channel NTE 2985

 

[Lazzer408]

Not a very good component but RDS(on) at 4v is .07 ohms so you 4.1v is plenty of gate voltage. NTE parts are junk (for future reference) but not the cause of your problem.

The switching frequency is a function of the motherboard and should be around ~20khz. What do the fans do if you pull the connector from the motherboard?

 

[Random Murderer]

What do the fans do if you pull the connector from the motherboard?

Could also be that his "trickle charger" is not supplying enough current on the 12V in for those high-current fans.
@Vario, have you tried powering this directly from a 4-pin Molex?

 

[Lazzer408]

He mentioned in #79 that he tried the charger... "with a few Deltas and San Aces, 8 fans total, they all seemed to run full bore with no PWM connected". I think it's in his computer now.

 

[Vario]

Could also be that his "trickle charger" is not supplying enough current on the 12V in for those high-current fans.
@Vario, have you tried powering this directly from a 4-pin Molex?

In observing this behavior, it was plugged into the 4 pin molex and the motherboards PWM. My "trickle charger" was only used for testing the circuit on my work table. I don't see why it needs to be in quotes unless you aim to make me look backward. Also a trickle charger should have plenty of current, I believe this one can do 8 amps.

Its possible its less than 4.1v now that its running on the power supply.

Not a very good component but RDS(on) at 4v is .07 ohms so you 4.1v is plenty of gate voltage. NTE parts are junk (for future reference) but not the cause of your problem.

The switching frequency is a function of the motherboard and should be around ~20khz. What do the fans do if you pull the connector from the motherboard?

If I pull connector from the motherboard they run full speed.

I could measure the voltage on the PWM lead with it running on my power supply and no 4 pin input. However, i believe the gate is modulating because the smaller lower powered fans do spin.

-It also seemed like running the h1011 on the single tach reporting 3 pin and the koolance 12038hbk on one of the 7 non-tach 3 pins, the h1011 wouldn't spin at all but the koolance would. The h1011 alone would spin, but very slowly.

-Koolances seemed better at managing the lower speeds but at higher pwm frequency they sang (whine), while the San Aces did not make this whine. The single San Ace barely spun at the lowest setting.

-So its possible this motherboard's fan curve sucks.

 

[Lazzer408]

I think it's in quotes because we have no idea of knowing what type of charger it is. If it's a trickle charger for lead acid batteries, then it's likely much higher then 12v and it's not DC either.

If the fans run full speed with the PWM unplugged, your gate drive is ok. What happens with you go into the BIOS and set fans at a constant speed?

 

[Vario]

I think it's in quotes because we have no idea of knowing what type of charger it is. If it's a trickle charger for lead acid batteries, then it's likely much higher then 12v and it's not DC either.

If the fans run full speed with the PWM unplugged, your gate drive is ok. What happens with you go into the BIOS and set fans at a constant speed?

Yes its a trickle charger for a battery. It read as DC current on my multimeter.

I'll try setting fans for constant speed.

 

[Lazzer408]

Put your meter on AC and measure that charger. Read anything? If so, that's the ripple in the DC caused by the chargers lack of filtering (they don't need it).

Next, place a capacitor in parallel with the charger and measure the DC voltage. Is it higher? If so, that's the peak DC voltage coming from the charger.

 

[Vario]

Put your meter on AC and measure that charger. Read anything? If so, that's the ripple in the DC caused by the chargers lack of filtering (they don't need it).

Next, place a capacitor in parallel with the charger and measure the DC voltage. Is it higher? If so, that's the peak DC voltage coming from the charger.

I'll check this out next.

I think it's in quotes because we have no idea of knowing what type of charger it is. If it's a trickle charger for lead acid batteries, then it's likely much higher then 12v and it's not DC either.

If the fans run full speed with the PWM unplugged, your gate drive is ok. What happens with you go into the BIOS and set fans at a constant speed?

So far, voltage mode = fans run full power. disabled fan control = fans run at full power, pwm mode = they operate as described above. These are the fan control options I have with this mobo.

Thank you for the help with this.

edit:
I am presently thinking the fan curve is too weak for the h1011 to spin. In my testing, they work as PWM should, where they spin up at boot, but then stop completely.
I am running two san aces and nothing else, with one of the san aces in the tach reporting 3 pin.
Is it possible to increase the switching frequency to counteract the motherboard's weak fan curve?

edit2:
-car battery trickle charger has 13.1v dc and 4.32v ac, didn't test with cap.
-I tested circuit on computer with five 3 pin cougar fans and it worked fine, the 5 cougars varied speed properly. These cougar fans are very wimpy, in comparison to the San Ace and Deltas.

 

[Lazzer408]

The PWM frequency is a function of the motherboard. All the board does is ground the PWM pin at varying widths. If you don't have complete control over the fan curves in your BIOS, this circuit may not work for you.

The "correct" way to control a fan (using Intel's PWM spec) is a more complicated circuit. My attempt was to simplify it as much as possible but it may not be 100% compatible with every fan.

One thing about PWM and power switching circuits is the layout of the PCB is very important. Wires jumping all over the place isn't good. What you see as a wire with 0-ohms is actually a resistor and when switching high current loads very quickly will product all sorts of voltage transients and noise that can interfere with the circuit functioning correctly. This may be the case when you start paralleling high current fans. Some suggestions would be to put a capacitor as close to your molex connector as possible. Also place the mosfet so the source pin is as close to the capacitor's negative terminal as you can get it. Try and keep all your grounds connected at the same point at the capacitor's negative terminal. That's about all I can suggest at this point.

 

[Vario]

Thanks for the advice. I would put the capacitor between the mosfet's source and the molex ground?

I think you are right about the fan curves in bios. Gigabyte's PWM curve isn't very curvy.

Would there be a way to just increase the duty cycle?

 

[Lazzer408]

The capacitor should be placed in parallel with your supply voltage.

Duty cycle (pulse width) is controlled by the motherboard. If you can't adjust it in the BIOS then there's nothing you can do about it.

 

[Vario]

The capacitor should be placed in parallel with your supply voltage.

Duty cycle (pulse width) is controlled by the motherboard. If you can't adjust it in the BIOS then there's nothing you can do about it.

Okay thanks. Damn.