0
\$\begingroup\$

I have a 12V 29.4W 3-wire Delta Electronics fan that I want to power using an EVGA 750 G3 PSU and control via one of the 4-pin fan terminals on an MSI MSI Z390-A PRO ATX LGA1151 motherboard.

Not considering the wattage of the fan, I originally plugged it into a (known working) fan terminal but it of course would not turn (and the BIOS correctly reported the RPM as 0). Connecting it directly to the PSU is simple enough, but this would result in it turning at full speed at all times.

I am curious if there is a "DIY" way that I can supply the fan with power from the PSU and control how much actually reaches the fan using the output of the motherboard (using something like an optoisolator to keep the motherboard safe).

The motherboard is able to control 4-wire fans using PWM, and 3-wire fans using DC voltage level.

Is there a product or combination of parts that will do this? I originally asked this on Super User, but it was deemed more appropriate here due to the involvement of DIY electronics.

Thanks in advance!

EDIT: The 4-pin connectors on the motherboard are indeed keyed to force correct connection of 3-pin plugs.

The BIOS includes a Fan interface that allows you to plot desired RPM vs temperature. Each fan output can individually be set to either PWM or DC voltage level control. The motherboard auto-detects what is plugged in (4 or 3 wire) and selects the appropriate setting, but it can be changed manually if desired.

The fan was not purchased through DigiKey, and came terminated with a keyed 3-hole connector (F). It also shipped with a 2-wire (but 4-wide) molex connector (M), unterminated on the opposite end.

\$\endgroup\$
  • \$\begingroup\$ Oops. I did not notice that the fan is a 3-wire fan. I assumed it was 4-wire. You can't do anything about its speed with a control signal because the motor driver buried inside the fan does not have a control input. You would have to actually change the supply voltage going to the fan. \$\endgroup\$ – DKNguyen Jan 22 at 23:02
  • 1
    \$\begingroup\$ Does this answer your question? How can I adapt a 3-pin PC fan connector to a 4-pin connector?. Although that post talks about adapting control signal, all of the suggested solutions there can be powered by external 12V \$\endgroup\$ – Maple Jan 23 at 2:52
  • \$\begingroup\$ Thanks for everyone's help! \$\endgroup\$ – Ctrl S Jan 23 at 3:18
1
\$\begingroup\$

The motherboard is able to control 4-wire fans using PWM, and 3-wire fans using DC voltage level.

I did not see anything about 3-wire fan support on that page, but if user manual says so then you don't need anything else but correct wiring on 3-pin plug. The PWM control pin will not be used, and the taho output from 3-pin fan is compatible with tach input on 4-pin fan connector. Furthermore, 4-pin connector is usually keyed in a way that forces correct connection of 3-pin plug.

Note, that if motherboard does indeed support both types of fans then most likely you have to configure fan type somewhere in BIOS.

I originally plugged it into a (known working) fan terminal but it of course would not turn

There is nothing "of course" about this. The fan should have been working at full speed all the time and motherboard should have been able to read its RPM. That's, of course, if it is configured for 4-pin.

From the link to a fan it seems it comes without a connector. So, check your wiring first.

UPDATE:

If your problem turns out to be insufficient motherboard output wattage, then one DIY solution would be to add OpAmp in voltage follower configuration, supplied directly from PCU and controlled by voltage output from motherboard connector. Depending on motherboard driver's design you might need to add a load resistor on the control line to simulate fan motor.

\$\endgroup\$
  • 1
    \$\begingroup\$ The OP's implication is that the fan is much higher wattage than a typical computer fan. a 30W @12V draws 2.5A almost 20x that of a typical fan. \$\endgroup\$ – DKNguyen Jan 22 at 23:01
  • \$\begingroup\$ @DKNguyen it is certainly on a high side, but we don't know what the motherboard driver is rated for. Without OP providing this info I can only assume that either wiring or BIOS configuration is incorrect. \$\endgroup\$ – Maple Jan 22 at 23:08
  • \$\begingroup\$ Might want to add that the opamp voltage follower will need an external output transistor at that level of current. \$\endgroup\$ – DKNguyen Jan 22 at 23:24
  • \$\begingroup\$ I am afraid @Voltage Spike found bigger problem - open collector output. I think by choosing this particular fan OP created nothing but problems. \$\endgroup\$ – Maple Jan 22 at 23:27
  • \$\begingroup\$ I don't see the issue. The fan having a tach output instead of a speed control input on the third pin is irrelevant to your solution of just providing supply voltage to the fan that tracks the mobo analog speed control signal. \$\endgroup\$ – DKNguyen Jan 22 at 23:30
1
\$\begingroup\$

Not with a 3-wire fan with a tachometer (TACH) output like the delta fan, the FG signal is for speed monitoring and it outputs a signal.

enter image description here
Source: https://www.delta-fan.com/Download/Spec/AFB1212GHE-CF00.pdf

One way would be to change the voltage which could be done with a variable resistor in series, but one problem with this is approach is the remaining power will be dissipated in the resistor. And with 24W you would have to be able to dissipate Watts of power in a resistor, which is not good.

A variable DC DC converter or voltage regulator might also do the job, the fan has a working voltage of 6-12V.

The best thing to do would be to get a fan with PWM control.

\$\endgroup\$
1
\$\begingroup\$

What you have.

enter image description here

What you need to make it work. A 20A Pch FET.

enter image description here

\$\endgroup\$
  • \$\begingroup\$ The only problem with this is one has to crack it open and figure out a way to mod the PCB to split "+" between hall sensor and motor. Unless, of course you modify the schematics to put the FET (and probably large capacitor) before power wire going in, leaving internal fan circuitry as is \$\endgroup\$ – Maple Jan 23 at 2:42
  • \$\begingroup\$ Also, I might be wrong but the FET will probably not open with the voltage level of the PWM signal \$\endgroup\$ – Maple Jan 23 at 2:50
  • \$\begingroup\$ It needs to have R pullup to 12V for OFF and low level to ON, so yes internal Mods \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jan 23 at 3:29

Not the answer you're looking for? Browse other questions tagged or ask your own question.