Manually controlling 4 pin cooling fans

Buy a 5.25 inch drive bad fan controller.

If you connect a 4 pin pwm fan to a 3 pin header, you will lose the pwm control, and the fan will run at full speed.

If that 3 pin connection has a rheostat control, then you can fine tune the rpm.
For a single fan, the zalman fanmate 2 or similar will do the job.
More likely with 4 fans to manage a front panel controller will do the job.
Here is an example:
http://www.newegg.com/Product/Product.aspx?Item=N82E16811990008

You should be able to use a pot with a circuit board but your going to have a hard time finding one over 4 Watts so it would depend on the fans if you needed to use 1 or 2. You would only need to use 2 pins on the fans.
Depending on how low you need to cut the volts down would very the amount of resistance you would need.

Example this fan draws 1.68 watts.
http://www.coolermaster.com/cooling/case-fan/bc-140-red-led-fan/

That would work.
You will also need to do some wiring so all 4 fans can be ganged together and a psu molex lead to supply the 12v voltage.

Would you be adjusting all 4 fans in sync, or would you be wanting to adjust them individually?
In the latter case, you would need 4 such rheostats.

If this is a one time adjustment, you could use one or more fanmate 2 located inside the case:
http://www.newegg.com/Product/Product.aspx?Item=N82E16835118217

Hi
Just to add a little ; make sure you check the power rating of the pot you buy,standard carbon track pots ( which are the most common) will only typically handle 0.4 W which is insufficient for fans.

Standard 12v led wall dimmer switch , works fine with 12v fans - will look nice too.
You're plainly competent at wiring so connections won't be a problem for you

http://www.ebay.co.uk/itm/Eastlion-DC-12V-24V-8A-Touch-Panel-LED-Dimmer-Switch-for-Single-Color-LED-/281900248974?hash=item41a2906f8e:g:W3sAAOSwKtlWiVGm



Obviously this is UK but fairly certain you'll find one in your country.

What geofelt said about 4-pin fans on a mobo 3-pin header is wrong. 4-pin fans have a backwards compatibility feature so that they DO work under Voltage Control Mode on a 3-pin header. In such a case, the fan receives NOT a constant 12 VDC on Pin #2, but a voltage that is changed by the mobo port. It also does NOT receive any PWM signal to modify the voltage supply, so it just does whatever the voltage allows. Thus, you CAN alter the speed of a 4-pin fan by changing the voltage supplied to its Pin #2. (Pin #1 is Ground.)

The best way to do this is with a pre-engineered third-party fan speed controller because all the fancy design requirements have been done for you. But you want to DIY, so here's an outline of what you need to do.

1. Determine how many fans you will run together in PARALLEL from this one voltage source.
2. For each fan, get its rated current flow (or Wattage rating) at full speed with a 12 VDC supply. Calculate its individual current if necessary.
3. Add those currents because your several fans are in parallel. This is the max current at full 12 VDC supply when they are running. You should realize, however, that when they all are starting up from stalled, the momentary (1 to 2 seconds) current will be about three times that. However, this factor will not really influence a lot your power rating calculation later.
4. The reality of fan motors and supply voltage is this: at some reduced voltage (often between 5 and 6 volts, but could be higher on an older fan with bad bearings) the fan will stall and will not start up again until the voltage is raised enough. To start a stalled fan you need to supply at least 7 VDC for most units with good bearings. So, your design should somehow prevent you from giving the fans less than 7 volts. Maybe your variable resistor should have a mechanical stop affixed to it, or maybe you can simply buy a resistor whose max resistance provides this limit for you..
5. What you really plan to do, I believe, is install in the circuit in series with the motors (paralleled as as group) and the voltage source a variable resistor. The units often used for such things are commonly called "pots", short for Potentiometers, but a true potentiometer is a different circuit function and is wired and loaded differently. Still, it is the same type of component.
6. Start with the two extremes to do your calculations. We will assume that the voltage source you use (a 4-pin Molex output from the PSU) can supply well more than the total current needed by all your motors, so we will not worry about that supply voltage. First, the full-speed case of NO resistance from the variable resistor and full 12 VDC to the fans. You can calculate the total current through the fans (the sum of their individual currents), the total Wattage of power being consumed by the motors, and the apparent "resistance" of that group of motors from the Voltage and Current values. For subsequent calculations you can use that effective "motor resistance" even though, in truth, it will vary somewhat as motor sped changes. The variations will not be huge and will be ignored for purposes of our calculations. For this first case, you can assume that the resistance of the variable resistor is zero, and hence the power dissipated at that component also is zero.
7. Next extreme is the minimum voltage you will supply to the motors, say 7 volts. Thus the voltage drop across the variable resistor must be 5 volts. Using the "motor resistance" calculated above, calculate the current that must be flowing through those motors with a 7 volt supply. That same current must flow through the variable resistor. So now you can calculate two things for it: the resistance required to drop 5 volts across it with that current flow, and the power being dissipated (and converted to heat that must be removed) in that component. Now you have one limiting set of performance requirements for the variable resistor - its maximum resistance and a power rating.
8. Now, pick a few more voltage situations through the planned range - say, motor supplied by 6, 8, 9, 10 and 11 volts. Re-do the calculations as above to estimate the resistor value but, most importantly, the power being dissipated in it at that setting.
9. Now survey your results. You will need a variable resistor with a maximum resistance as calculated in step 7 (or maybe a bit more), and a power rating better than the max power being dissipated in those several scenarios you calculated.
10. Don't forget to use the maximum current flowing in the circuit (very likely the starting current, estimated to be at least twice the full-12-volt current) in choosing the gauge of the wiring you will use.
11. The variable resistor you buy probably has a power rating that assumes the heat will be removed simply by ambient air flow in the vicinity (but verify that) so you probably do not need a fan to blow over the resistor. However, you DO need to ensure it is in free air and not enclosed, AND that its mounting system can stand up to a hot component. You also need to recognize the physical size of the unit to mount it.

Third-party fan speed controllers do this job differently so they look quite different from simply a large variable resistor, and they dissipate less power (thus generating less waste heat) than your system will.

BIG caution for you. In a previous thread I was part of, others posted strong notes from the makers of 4-pin computer case PWM fans. Now, PWM modulation of power flowing to a motor has been used for some time. HOWEVER, those readily-available systems provide a pulse-modulated supply to a relatively simple motor. The so-called "PWM fans" for computer case ventilation are quite different. Inside the latter are printed circuit cards that accomplish at least two major jobs: they use electronic switching to provide a simulated commutator in a brushless motor, and they use a separately-provided (via Pin #4 of the connector) PWM pulse signal to do their own modulation of the constant 12 VDC supply in Pin #2. These motors should NEVER be supplied with the kind of already-modulated DC supply that traditional PWM controllers provide because they depend on correct input voltage and PWM signal (separately) AND their internal components can be damaged by pulses and noise in the incoming voltage supply that is used for ALL components (not just the motor winding). So, do NOT use a common PWM Motor Speed Controller to try to control the speed of a computer-style PWM fan. Further, do NOT use the common types of lighting brightness control units intended for room lights or LED's. These devices also provide very spiky interrupted voltages that are likely to damage a computer fan motor. Even 3-pin fans that operate on Voltage Control Mode and are designed for reduced DC inputs on Pin #2 still use internal circuits to simulate a commutator system, and hence are subject similarly to damage from noisy power containing spikes.

You're welcome, and thanks for Best Solution.