Fan stuff Part I
. . I have seen a lot of schematics or pics on the Web about connecting fans by wiring them to power connectors for HDs/CD-ROMs (or to appropriate wires coming from PSU), especially if the fans were powerful ones, which could "kill" the MoBo header by demanding too much current for them & that headers are mostly enough for around 0,25A max. which means, that no more than TL of 2,5W fans can be connecter there (at 12V) per header or better less than that, because the starting movement of fans needs even a bit more of current than usual spinning.

That would not be a big problem as those fans usualy come already with end of wire with Molex connector for powering them & also some people with more fans than headers on MoBo connect them to HD Molex for example between black (2 or 3 pin of it) and yellow@ (1) for giving fans a 12V power or in very rare cases for extra silent operation on 5V (4-red), if the fan can start spinning there for good, since fans are mostly rated for 7V start voltage. Btw, I haven´t encountered myself yet a fan not starting at that low voltage, some smaller ones when new were starting at even less than 3V! Depends a lot of their inside electronics construction; yes they have some electronic circuitry inside in the middle of the fan under label sticker which drives them w/o brushes; thats why they are called brush-less opposite to simple DC ones.
  Mostly all of these days PC fans are brush-less type with two colored wires coming out of them, black means that has to be connected to negative potential (or Ground on PC circuitry) & a red one to positive potential. With these ones is not easy possible to have a RPM detected (exists a way - another mine idea will be later published here somewhere ), so you have there also a double priced ones with a third wire coming out for that and usualy is colored white or yellow, but can be also some other color.

Since some fans are really noisy running them with 12V, there are few ways to silencing them, mostly with supplying them with less voltage using serially connecting to them an appropriate resistor (fixed or variable), or some diodes, or some small electric light bulb or making a special "fanBus" with switching circuitry or even independent electronic voltage regulator and a potentiometer for linear regulating voltage for fans. There are a lot of combinations of that, but I like simplest ones like wiring fan for 7V on Molex connector like many guys did.

Since electrical potentials on that connector are +12 & +5V against the case (ground or minus) of PC with enough current on them, there is no problem to wire the fan directly to those sources with a black wire to lower (+5V red) one to get a 7V power to the fans!

The fan will be spinning at less RPMs (approx. 35% less & so much less also air pressure & air flow and so will do cooling worse!), BUT the noise will be more than one time lower (about 12dB lower*-see end note about dB levels!) according to my experiments. Some newer, especially for AMD CPUs MoBos need that Cpu HS Tacho fans to report to MoBo that are spinning Ok to prevent protection powering the machine off, but that could be also circumvented by connecting there another Tacho fan or disabling Bios protection for that, but that is not wise! So IMHO is better to implement such mine circuitry if wanting to run a Cpu fan powered at 7V.
7v Running Fan   This kind of wiring is also color coded drawn & shown on this slide, but here is also something new added, if having a 3-wire fan with RPM sensing, usualy used for CPU HSF "safe" Rpm sensing!
Is this additional wiring with additional 2 resistors & a capacitor needed? IMHO yes if wanting MoBo safety long term with Rpm sensing and fan connected like that! There have been reports on Web of wiring directly w/o those additional elements & not working in many cases & working in rare other ones (but how long?), so it looks that depends on MoBo HWmon chip circuitry design involved. Mostly are used Winbond chips there & their schematics involved in Rpm sensing looks like this on the right slide :
. . here is how fan gets power not shown on right slide
Since manufacturers these days take care much more on production costs & easier implementing stuff, they do certain things simpler with cheaper parts on cheaper MoBos & so that transistor there on left powering fan gets killed sooner or later if there is connected a fan with a greater current consumption than transistor rated (check the MoBo manual!). On the other side is a version of non protected input of HWm chip, which can also get killed if more than +5,5V gets there. Thats why on my schematics is a capacitor there for preventing that (of polarized ones can be also used just one if properly oriented-negative towards MoBo!); it stops DC current rated between +5 and +12V on Rpm wire with fan connected so, but "passes AC" Rpm impulses.

The 4,7k resistor lowers the voltage on chips input & thus making it more sensible since impulses get there with signal voltage rated lower than normal. The 2,2k resistor gives the open-Collector transistor inside fans Rpm circuitry a load to produce impulses with fan revolutions which works there like a kind of a switch shorting that resistor according to them (instead of shorting that 4,7k one on a Winbond schematics) to fans black (negative now +5V "virtual ground") wire once, twice or more per revolution (depends on fan electronics construction). The resistor has about half the value of ones on MoBo fan connector side to match a new total impedance . That´s it ; simple, effective & more follows on next site:

 

 

 © Zdenko Jerman-Spajky