OK, so for my 1000th [double] post I thought I’d try to contribute something that people will hopefully find useful (rather than arguing with Nexxo about porn). Since I’ve spent a lot of time in the past collecting and messing around with 12cm PC fans and making both simple and complicated fan controllers, I figured it’d be good to have a kind of “repository” of DIY fan controller approaches in one thread. So I’ll share all the controllers that I have built and tested myself and also other designs which I know to be very good but haven’t necessarily built. 1. DISCLAIMER Tampering with electronics is potentially dangerous -- if not for you, certainly for your PC -- so before you attempt anything in this thread please make sure you know what you are doing and employ common sense. A good rule of thumb is this: get an el-cheapo PSU so that you have all the standard PC voltages at your fingertips and don’t have to worry about frying your shiny new 256GB SSD when testing circuits for the first time. I fried my cherished £300 Foxconn motherboard before I developed this wonderful rule of thumb (and indeed before I knew what common sense was) so I have definitely learned my lesson the hard way – let’s hope you guys don’t have to! 2. UNDERSTANDING THE BASICS: PSU VOLTAGES As many of you will know, there are only three fixed and regulated voltages available from a standard ATX PSU: 12v (yellow), 5v (red), and 3.3v (orange). Only SATA and ATX power cables carry all three voltages; voltages below 3.3v are all regulated on the motherboard or on the components themselves. The 7V trick is very easy but I wouldn't recommend it because of the risks involved, so I'm not going to discuss it in this guide. 3. PC FANS AND 12VDC MOTORS There are principally two types of fans for PCs: two-wire fans (three-wire if they have RPM feedback/tachometer), and four-wire fans, which are usually called “PWM Fans.” Both types of fan are based on simple 12VDC motors, however PWM fans have more complicated internal circuitry and usually very high quality motors which are controlled in an entirely different way (more on that later too). The controller for four-wire fans will be discussed in part 2. 4. POWER AND CURRENT Before you attempt to regulate or control the speed of a fan, you MUST know what its power requirement is. Fans usually have an amperage rating or a wattage rating printed on the ticket. If the wattage is not printed on the fan, all you need to do is simply multiply the amperage by the voltage; if neither amperage nor wattage are printed on the ticket, you will need to use a digital multimeter to measure the current! Most 120mm PC fans require somewhere in the region of 100-500mA, however there are a few more powerful fans which require quite a bit more. Fans like the Ultra Kaze 3000rpm (600mA), and the high speed Gentle Typhoons (starting at 1.14A) have unusually high amperage, and server grade fans made by Delta, Sanyo Denki, NMB-MAT, Papst, Sunon, and Nidec, can draw in excess of 4A! 5. SPEED CONTROL BY RESISTANCE This is by far the simplest way to control the speed of a fan, but it has two disadvantages: it wastes energy and can create a lot of heat. The fan will slow down as much or as little as you like depending on the value of the resistor, but the power draw will remain the same, and the wasted energy will be dumped as heat by the resistor. Not a huge deal if your fan draws 2W, but a big problem if your fan is a server grade fan which draws 36W. Resistors are usually very small and have two values: resistance, in ohms (denoted by “R”), and power handling capacity, in watts. The small resistors are carbon or metal film and have varying capacities of 1/4W, 1/2W, and 2W; bigger resistors are called “wirewound” and have much higher capacities of 3W, 7W, 10W and above. In this picture I placed a 1/4w resistor next to three wirewound resistors: 6. CALCULATING RESISTANCE It is possible to calculate the resistance required to achieve a target voltage for your fan (there is a very nice calculator here), but I have found that trial-and-error is the best way to establish what resistance works best, especially when you are looking for a very specific speed or acoustic property rather than just a target voltage. Just make sure that the resistors you buy can handle the current that the fan will draw – like I said, most fans require less than 500mA and will therefore manage with a 1W or 2W resistor, but if your fan has a current draw of over 500mA you may need to use a 3W wirewound resistor. 7. VARIABLE RESISTANCE Variable resistors are called potentiometers and, like fixed resistors, they are also wirewound if they have a high power handling capacity. If you find that a 30 ohm resistor with 2W capacity slows your fan down nicely to about 30% speed, then a 30ohm, 2W wirewound potentiometer will allow you to regulate the speed of your fan between 30% and 100%. Be aware however that the potentiometer will get very hot if you constantly run the fan at its slowest speed! Here’s a pic of a typical 2W wirewound potentiometer: very small, but quite beefy. 8. CIRCUITS This is where the DIY part gets fun: building a controller circuit. Starting with the simplest and getting a little more complicated, there should be plenty here to satisfy any diehard modder. 9. TRANSISTOR CIRCUIT The simplest circuit I can think of is one that requires only two components, and it is reasonably effective for low powered fans (even though the transistor is rated for up to 3A). The transistor still has the problem of voltage dropout (discussed later) but is a quick-fix if you want to make a half decent fan controller on the cheap. 10. LINEAR VOLTAGE ADJUSTMENT CIRCUIT Before we look at the 5-11V circuit, we need to briefly discuss what voltage dropout is. Simply put, when voltage goes through an adjustable regulator (or transistor etc), the output voltage is not quite as high as the input voltage. The difference in voltage is referred to as voltage dropout and it is something of an inconvenience, because it means you will not be able to run your fans strictly at full speed. It is possible to get low dropout voltage regulators, but they are rare and can be very expensive. The most common adjustable voltage regulator is the LM317 but it has very high dropout (up to 2V) which is why this circuit uses the LM1084IT-ADJ, a more expensive and far better regulator. This circuit was designed by bing at overclockers.com forums and has a nice kickstart feature which means that the fan will start at full speed and gradually slow down to the speed set by the potentiometer. It is also a very high capacity controller capable of providing up to 36W of power. Here’s the circuit at breadboard, and a short video of the circuit with two fans (Gentle Typhoon 3000rpm and Delta FFB1212EHE):
PART 2: PULSE WIDTH MODULATION Now this is where it’s at: pulse width modulation. Most premade fan controllers use voltage regulation to control fan speed, which is why there are big heatsinks for the FETs or regulators on each channel. Even PWM-based fan controllers (there aren't many) waste heat, but that's because the way in which PWM is used is not ideal. Read on to find out how PWM can make your life very easy indeed. 1. TYPES OF PWM CONTROL There are two ways to implement PWM as a means of controlling a fan’s speed: modulating the 12v supply, which works for all fans but has drawbacks (clicking, loss of RPM sensor, wasted energy) modulating the fourth wire of PWM fans, which of course works only with four-wire PWM fans 2. BENEFITS OF THIS FAN CONTROLLER There is no wasted energy. None. The circuit uses only 5V for power, so, unlike all other fan controllers (including other styles of PWM controllers), this controller has nothing to do with the supply voltage for the fan; instead, it sends a signal to the internal circuitry of a PWM fan via the magical fourth wire, instructing it to switch on and off many times per second (16KHz, to be precise). Another benefit to this controller is the accuracy of it: the duty cycle range is exactly 2%-98%, and between these two levels there is very smooth and precise control of the fan’s speed. Unlike modulating the 12v supply, there is no clicking with this implementation of PWM, because PWM fans are specifically designed for this type of control. This is IMO the crowning glory of this controller: you can use extremely high powered server grade fans with this controller, and when turned down they will run as quiet as a mouse without creating any extra heat or wasting any energy. Most pre-built controllers have limits to each channel, but this controller doesn't carry the supply voltage to the fan so you don't have to worry about annoying power limitations. Just get a 50W server fan, connect it up, and watch it fly from 800 to 6000rpm! NB: For a very detailed guide on how to build this controller, please consult this article which I wrote for overclockers.com earlier this year. In this thread I will run through things quite quickly. 3. THE BASIC CIRCUIT So here you have it: you need only a handful of components and wire, a fairly basic understanding of electronics, and a breadboard (see below) to get you started. The values of each of the components are given in the article above. A more powerful pulse generator can be made from the original circuit – just add a second 555 timer as shown below, and the controller will be good for powerful server fans: Or, if you are particularly competent, you can try the 556 version – this IC has two 555 timers in one 14-pin package, so the circuit is slightly different: 4. BREADBOARDING This stage is really important if you want to be thorough (as you should) and test your circuits before you assemble them properly on stripboard or similar. Remember to use a spare PSU if possible – it’s very easy to cause a short circuit with all the exposed terminals, and you don’t want that happening when your PC is switched on! Here is the 556 circuit on my breadboard, fully operational: 5. FINAL BUILD And (after many, many incarnations) here is a two-channel version on stripboard. The two outermost four-pin headers are for the fans; the one on the inside is the main power header which is connected to a floppy drive power connector. Here is the controller sitting on top of a very nice Delta PWM server fan (AFC1212DE) which I currently use as my CPU fan. Thanks to the controller, this 4000rpm, 1.6A monster runs at about 800rpm and draws only 150mA of current without creating any excess heat! And finally, here is a video of my updated fan controller. I added a third channel, made a 5.25” drive bay case from black styrene modelling plastic (very easy to do), and threw this together. Note that ALL the fans in my PC at this time were PWM fans, so the controller was very well suited to my build (or, rather, the fan choice was pretty much centred on using a PWM controller). Well that’s it! If you have any questions or would like to suggest some more fan controllers to be included in this thread, please share your ideas! Before I go, I did mention the server grade fans that are rated at 4 amps. Well, here is this controller taming two of them!
I applaud you sir, and i move for instant sticky on this thread! I'll be giving this thread a more thorough read before the end of the month, when i decide how to control my next build!
Nice tutorial, I look forward to the next bit. I too have learnt the lesson while using the 7v trick after cooking 2x 2TB drives
on other forum one guy build fallout themed case and he try to hide/mask fan-pump controller. and i was SUDDENLY kicked by idea to create _lamp_ fan controller. but i don't have enough knowledges about lamps
Thanks for the feedback folks. There are of course many things you can do with simple DIY fan controllers - eg. using different coloured LEDs to signify a different level of resistance, or using slide potentiometers for a different look and feel. The options are endless!
This is awesome. Is this cheaper than buying a fan controller or just for bragging rights? The pmw one seems like it would be a better fan controller anyway
It was about time this made it to BitTech Lenny! I've been running my PWM controller for a few months now. It's still on the prototype board as I can't solder to save my life. Anyhow; this is my controller and the fans that I use it to "tame back". Controller: There is a small mistake on the wiring on that picture, but that has been fixed long ago. Fans: And a couple videos of the controller at work on the San Ace fan and then on the Nidec fan... The Nidec fans work on a different frequency than the San Ace fans so one of the capacitors on the circuit needs to be changed depending on which fan one wants to control. I keep the DYI controller configured for the Nidec fans simply because the San Ace fans understand PWM signal from the motherboard headers so I control those straight from there. And now for the videos... Nidec Servo San Ace Catch you at OCF Lenny... Sebas
It's not cheaper than buying a fan controller, because technically you may need to buy a soldering iron as well as all the components Plus, as Sebas said, maybe soldering is not your thang, in which case it can be quite a bit of work. But perseverance is key! And I would say that bing's PWM design is a much better fan controller than anything you can buy from a shop - no other pre-built controller can handle 50W server fans, and no other pre-built controller can control four-wire PWM fans properly. Haha I know mate - I dunno what took me so long I've been watching the PWM thread at OCF and it's gone completely mental lol.
Still to make mine Lenny But have to say, the videos just don't show quite how scary a 36W server fan is when spinning up at high speed, need to do another video with you throwing some carrots or something into them so people get the idea!
The 555 PWM is way too complicated I think, try; With this circuit depending on the FET used you could be controlling hundreds of watts of fans. Scale it how you want.
So true - the 6000rpm San Ace actually sounds like a jet engine, and I'm sure it would eat a few carrots for breakfast. Might try the carrot trick with one of the Deltas If you can, set aside a whole day to get the thing prototyped and boarded. It's a lot of fun, and once it's made, that's it - no more hard work lol. Thanks for sharing this - whilst it's a perfectly viable PWM controller, the drawbacks outlined above still exist: the use of a FET means there is a waste of energy in the circuit, ergo unnecessary heat output; and because it's the 12v supply being modulted, there might be clicking at low speeds and of course the RPM sensor is sacrificed. Not knocking it at all, but the 555/556 version although more complicated is far superior.
Sorry I may be wrong but why would there be clicking at low speeds? It's the same signal just a lot more current capability. The FET dumps almost no energy a small TO-220 heatsink is needed for about 100W of fan controlling. In all honest unless your controlling like one fan or want to controller fan speeds individually I would say the circuit I recommended is better.
Boscoe - I'll do my best to answer your questions. In short, this is more about PWM fans being superior than it is about bing's circuit being superior - I'm sure your circuit is very nice, but it does have its weaknesses, just the same as non-PWM fans have their weaknesses. First, here's a succinct explanation of the ticking noise from modulating the supply voltage, but it's pretty heavy stuff: So there you go - the ticking basically comes from the crude "start-stop" implementation of PWM. Admittedly, small fans are not such a big problem, but more powerful fans which have a high startup current (like my San Aces which need 4 amps, and the 100W load you specify) pose a big problem for a high duty cycle on the supply voltage (and can damage the fan's motor). Out of curiosity, have you tested your circuit with a powerful fan like my Deltas or San Aces? I'd be interested to know how it fares with 3 amps being modulated. As I said at the start of this post, the 555/556 controller is a better use of PWM mainly because of the quality of the fans that it is designed for. PWM fans are vastly superior to standard DC fans because they (especially server grade models) contain extremely complex internal circuitry. This smooths the torque profile of the square wave pulse sent through the fourth wire (PWM wire) and not only eliminates ticking but leaves the 12v supply entirely free from interruption; just smooth, steady current which increases or decreases relative to the duty cycle of the PWM signal. Check this pic of a monster DELTA PWM fan (TFC1212DE) from behind - this is no simple 12VDC motor! (The PWM wire has snapped off, but you can see the connection clearly, and you can also see how integral this method of PWM control is to the design of the motor).
I have used a very similir circuit to the one i posted (i think it had an extra resistor somewhere) driving an IRF450 which controlled a 200W bike motor quite easily with about 8A. Sorry I didn't realise you were using PWM fans I was on about regular DC fans, If you using a high switching frequency one way higher than the motor can manage to move at the same frequency you can beat the ticking.
I guess that's really it - you can use PWM on the supply voltage, but it will never be as efficient as using a pulse generator with a designated PWM motor which has a separate input for PWM. I'm interested in your PWM circuit and might throw one together for some of my non-PWM fans.
OK, I threw your suggested 555 controller together on my breadboard with a beefy MOSFET (IRF540A) and tested it with a couple of standard 12VDC fans. Although the controller works really well with a medium powered fan (700mA Y.S.Tech fan), when I tried it with one of my 3A Focused Flow Deltas there is a really bad pulsating whine noise at low rpm and the fan does not turn smooth at all. I think it is ideal for low powered fans because it generates a lot less heat than voltage regulation or resistor/transistor control, and the speed adjustment is extremely precise (even though RPM is lost, which might annoy some people). I wouldn't however recommend this controller for powerful fans, as the reason people want a controller is to make very loud fans quiet, and this controller certainly doesn't do that! The controller: And a video of the really bad noises at low rpm!
You're missing the connection to ground for the pot. there. Unless you intend to do current limiting by limiting the base current to the transistor rather than to do a voltage reference at the base (quite a hit and miss based on the actual hfe of the transistor and the value of the pot). Circuit should look like this for pass-transistor:
Lots of good work in this thread That fan sure doesn't sound too happy, maybe it's a little hungry or need it's nappy changed?
Grounding the pot isn't necessary (it's not "missing" ), but I was going for extreme simplicity - it is indeed very hit-and-miss, but it works fine with lower loads. There are a number of ways to modify and improve this circuit - eg. grounding the pot with a resistor allows the minimum output to be adjusted. A much better (but more complicated) transistor based controller uses two NPN transistors in parallel (or a single darlington). The additional transistor takes the potentially harmful base current away from the potentiometer track and increases the capacity of the circuit. The controller below can safely control fan loads up to 1A, and the 1K resistor limits minimum output to approx 4.5v. Hahaha A little hungry and a little tired, but only two weeks old!
