Fan stuff Part II
How did I set voltages for fans in my setup ...
There are many other ways to reduce voltages & thus power send to different fans in the setup described on many web sites except that mine direct connecting to Molex connector for 7 or 5V powering on part one article :
- using an independent electronic chip for voltage regulation; if they are "linear" type they heat & need a smaller heatsink; if they are PWM type & stronger one maybe there will not be any HS needed, but they are not so cheap anymore; such regulator needs basic knowledge of electronics. In this group we could put also different kind transistor regulators acting like an electronic variable resistor (mostly also needs additional HS) regulated by a cheap simple potentiometer or so. The benefit of such regulation is a possibility of powering fans by any desired voltage linearly.
- using resistors serially to fans (they get hot so need to be over-dimensioned (few Watts); have to be calculated for proper voltage drop for different (fan) currents but they are cheap.
- using a variable resistor serially to fans (a potentiometer), a bit more expensive but need to be rated also few W and have to be calculated for proper maximum voltage drop for different currents & is better if they are "wire" type since ordinary ones are not powerful enough & get destroyed (burned) sooner or later.
- using serially to fans a small electric light bulb acting like a resistor; which one is determined by experiments, the thin wire inside it should lit with dark orange color so few volts/less than a watt small one should be ok & there should be a voltage drop on it of 2 or 3V; so they do not lit full bright & when they are cold are having only few ohms of resistance, so for a moment they give fans a boost of full voltage to help them start spinning when powering up a PC & just than a second later get a higher resistance when start to emit light & they are cheap.
- if having for example two identical input air case fans close & are able to start spinning at 5V, than can be connected together serially to 12V power & so each having a half (6V) voltage on _
_ them spinning at less RPMs both. But if one fails, the other will spin full power or will not spin at all. This is the bad part, but occurs rarely!
- or ... using serially to fans connected an array of cheap Si Diodes also soldered serially; on each one there is a voltage drop of around 0,7 volts generally but rises a bit with higher currents to 0,8V. Nice ones are for example 1N4007, they are 1A rated, voltage drop using them with fans is around 0,8V each, so you need for 5V drop to power fans with 7V six of them (for 9V fan powering you would need 3 or 4 of them for example).
They have to be oriented properly - anode to higher positive potential, cathode (well marked with a colored ring on them) to lower (more negative) potential because they conduct electricity only this one way. The good part is that they do not add additional heat to the system & are small & cheap. Also the array does not change much its voltage drop, it is quite the same all time regardless how many fans are connected to it, but all fans current must not exceed single diode rated current. With a bunch of such diodes and a 1-pole few positions rotational switch can be made a nice "rheobus" - fan voltage controller (next page!) ...
Any serial reducing voltage stuff should be better inserted in positive line feed of fans, especially if having a tach type (3-wire) fans, since there could be a problem revealing fan RPMs if it would be inserted in a negative line & dropping the voltage more than 2 or 3V there and thus rising from ground (power minus pole) a zero point of tach impulses too much; this could lead to a non functioning revealing of fan RPMs.
Since I like simplicity, I used mostly Si Diodes connected serially to reduce voltages, but there is an additional transistor (mostly) acting like an electronic switch to reduce voltage further in special cases piloted by the voltage of a MoBo Cpu fan connector! The connection schematics for my setup looked like this before:
Like you can see on upper slide, I did not have a fixed (master) negative fan power line connected to negative (ground) potential directly but thru a transistor and 2 diodes. So negative feed for all of fans was connected to a higher bolded line "floating virtual ground" bypassed by a blocking (decoupling) condenser to reasure proper Rpm sensing of a Cpu tach type fan (the only 3-wire one I had inside). The voltage on MoBo Cpu fan header provided full voltage (current reduced by that resistor) for transistor B-terminal, so it conducted fully thus shortening those 3 diodes on its right side like they were even not there, so that virtual negative feed had in reality almost +2V there in reference to a ground ; the optional switch for high summer temperatures could reduce voltage drop even further if the contact was closed shorting that 2 diodes (that virtual negative feed become practically a real minus pole or ground) so providing even a bit higher voltage to fans so spinning faster if needed.
. . . The case fan & a chipset one had even more reduced voltage with additional diode; the first one because was the most powerful there & the second one to reduce its speed further to diminish its rotation speed preventing that nasty high pitched whine this small kind of fans usually have. Chipset fan is horizontally turned upside down to suck fresher air from MoBo to "greenie" (chipset HS) since I´m having a desktop case & Mobo is situated on the bottom of it where the temperature is lower than under the top of a case. Since I used an 6y old very weared & noisy P166 fan for chipset, I had to clean & oil it up; buy turning it "wrong" way I got it running mostly without laying its axis on weared bearings bottom but instead its rotor with blades is mostly spinning on its own "floating" rotational magnetic field thus preventing wearing bearings more & now its much quieter! Also I additionally reduced a Psu fan speed by inserting an additional diode to its positive feed line going to its thermal spinning control electronics inside PSU already existed there before.
© Zdenko Jerman-Spajky