Just bought all the components I need... total cost of £6 from Maplin for them lol Maybe get it done today, maybe not!
A Darlington would have a 1.2v - 1.4v drop, limiting the maximum voltage. A Sziklai pair would be better. While we're at it, adding a capacitor would give a kickstart for fans that don't like start at lower voltages: OT: What program are you using to draw that circuit? Eagle is a bit of an overkill to fire up and doesn't have a motor or fan library (I just use an inductor and change the name for these diagrams).
Well, I knocked it up quickly and it works pretty well! Strange thing is, when I only had the one 555 the circuit did not work, wouldn't even vary my cheapo intel stock HSF, but when I popped the second in it controlled my Delta quite happily. May solder it all up tonight, not sure. Gotta make up a second one too so I can control the other fan (not same so want a seperate controller)
Well, that confirms that my soldering skills suck. Just going to tidy it up and shorten the wires and leave it on the breadboard me thinks. Will also go along to Maplin and buy another set of components to have a double controller on the one board. Will work for me!
Yeah, with heavier loads there is always bad dropout, and the kickstart feature is nice. Thanks for sharing the compound pair idea with kickstart - very nice. Re. drawing the circuits - I either use MS Paint, or (with the picture in question) nab somebody else's diagram! Nice to see you've caught the bug Soldering the controller is a different matter altogether and can be a royal pain in the proverbial donkey - took me several weeks to get a stripboard layout that I was happy with, and even then it was still tricky to put together. As for your single 555 problem, with breadboard it's very easy for a wire to become dislodged and the connections can be quite poor, so that might have caused it not to work. With two 555s the controller is much better anyway and will control several powerful fans. Now you can see why PWM server fans like your new AFC1212DE can be amazing for PC cooling - they are totally silent at low rpm but have a great airflow curve as you speed them up. Worth mentioning that the beastly Nidec that you bought from me has inverted PWM control, so turning the pot clockwise will slow it down! Also its lowest speed is not very slow and the motor is quite whiney, but then again it does go up to about 5000 RPM and sounds like a jet engine at full speed. Just remember that the Nidec used with the same controller will slow down when you turn the pot clockwise.
Noticed that already when testing them Quite impressed at how little this actually costs to build, Maplin is hardly the cheapest place and it all cost £6 from them. Hopefully get it all tidy and sorted in the next few days, maybe a simple box made to hold it all and protect it. And with any luck all my water cooling gear will arrive and I can perform some serious mods to my case to try and fit in 2x240mm rads along with these stupidly large 38mm fans!
Heh yeah man it's such an amazing wee controller circuit, dead cheap and reasonably easy to put together, and by far the best means of fan speed control. All you have to do now is start collecting crazy fast PWM server fans!
You sir, are exceeding skilled at M$ Paint. I've lost touch with it years ago. Another alternative is to use a cheap NPN transistor like the 2N3904 on the output of the 555. 4pin-PWM fans typically only require a maximum of 5mA of sourcing so the load on the 555 will be reduced easily with a simple NPN. Additional NPNs with base resistor can also be placed for driving multiple fans. Although there is enough current gain from a small NPN, multiple fans with their PWM pins being out of phase may cause issues with control of individual fans. If many units are being driven then a cheapo 74 type TTL buffer IC would be more cost effective and wouldn't need multiple base resistors. OTOH, the NPN will invert the output so pins 1 & 3 of the pot needs to be swapped. The pot can be swapped Pins 1 & 3 so that the control is more intuitive (clockwise to increase speed).
@ MS Paint, it's just a patience thing tbh. The diagrams for the 556 and linear voltage controller were a bit fiddly to draw up, but the benefit is you can use any symbol you want, as long as you can be bothered drawing it. Any chance you could put it into a diagram? I'm curious as I may have spare NPNs lying around and would like to try it with some of my more powerful PWM fans.
Don't have enough time to right now but it's basically connecting PWM out in your 555 diagram to a 1k resistor to the base of the NPN. Ground the Emitter of the NPN and flip the pot's pins 1 & 3. The Collector of the NPN then connects to the FAN's PWM wire. Basically, for the Nidec circuit, you're using the 555 to pull the PWM line low (4-wire PWM fans have an internal pull-up). When you use an NPN with grounded emitter, you're turning it into a switch where it conducts (from collector to emitter; thereby shorting to ground) when the base voltage exceeds 0.7v. Hence, you can regard it as inverting the input since the output is grounded when the base voltage is 'High' and the output is high-impedance when the input is 'Low'. Edit: Here's the diagram
Have been busy for the last hour, I cleaned up the circuit on the breadboard and shortened the wires on it to keep it clean, then managed to work out how to fit in two seperate circuits, and fan headers onto the one little board. Its a tight fit, I only have two rows free on the whole board but it works very well I have set it up so I will have to run wires off to the two pots, that way I can stow this inside the case without worry of short circuits, and then put the pots wherever suits me. TBH anyone who has PWM fans should give this a shot, I have a dual controller now that cost a total of £8 for components, and £1.50 for a breadboard, plus some wire I had going spare.
@ Dreamslacker, nice work and thank you! @ Tarka, very nice and neat. I think I'm gonna go back to breadboard and give Dreamslackers NPN mod a go! In the meantime, one of the benefits of this controller is the RPM sensing and (depending the fan) the ability to turn the fan completely off without any danger to the circuit. The RPM reporting can be a bit buggy and slow to respond, but it's fairly accurate for this 5,300 RPM fan
Slightly different variant: Part values of R1, R2 and C1 sets the frequency to 24.5KHz. Well within the accepted range for PWM fan control. R2 sets the minimum duty cycle to 20% because the specifications call for a minimum of 20%, below which, there is no obligation from the manufacturer that the fan will function properly. Duplicating R3, R4 and Q1 will allow for control of more fans. R3 can be scaled accordingly instead of duplicating. C4 is actually optional in this instance and in fact, this circuit may be driven directly from the 12v line. The use of the NPN transistor actually shields the PWM pin of the fan from the 12v line since it's sole purpose is to act as a switch to ground. The main difference from your circuit is that the RC circuit is isolated from the fan PWM pin. This pin contains an internal pull-up which will charge your timing capacitor when directly connected since it skews the charge/ discharge cycle by charging the capacitor through the diodes. That is why a 2nd 555 timer generally works better for you. It's not a current limitation (the output of the 555 can souce/ sink close to 200mA, far in excess of the 5mA max imposed on pwm fans). Rather, certain fans have a higher internal pull-up current which may charge the capacitor too fast and completely ruin the RC timing. The circuit I provided can actually be connected directly to a fan without any issues provided the pot is reversed. The discharge pin will only sink current and will rely on the internal pull-up of the PWM fan. The Threshold and Trigger pins are unaffected by the fan since they are not directly connected.
Awesome work Dreamslacker Re. PWM frequency depends on the fan - as the PWM fan document at formfactors.org is an Intel document, the 21-28KHz range is consistent with their fans and compliant PWM fan headers, but server fans can be very different, eg one of Nidec's 120mm monsters states 5KHz as preferred, whilst Deltas generally accept 300Hz-30KHz with 25KHz preferred. The min 20% duty cycle imposition is also an Intel-only thing - all my San ace, Nidec and Delta server fans are designed with specific speed at 10% duty cycle, sometimes even below. Fans which come to a full stop at 2% duty cycle are quite rare and IMO it's a really nice feature having a speed gradient across the full turn of the pot. I'm guessing R2 can be left out for lowest duty cycle? There is a 0-100% controller on OCforums which uses 2x 555 and a comparator (LM311), but it is most useful with the most expensive and best PWM server fans which have a lot of control circuitry for the duty cycle behaviour.
The frequency of the unit can be approximated by: Frequency = 1.44/ [ (R1+R2) x C1 ] or Frequency = 1/ [ ln2 x 2(R1+R2) x C1 ] If R2 is omitted, then calculate with R2 as '0'. If you omit R2, the value of R1 and/ or C1 has to change accordingly. The scaling is linear so values can be swapped accordingly. i.e. If you want to use a 100K pot instead of 1K, then reduce the value of C1 by 100 times. Also, you may connect a switch from the base of the transistor to ground. When turned on, this forces the transistor off. Thus, you will, in effect, create an 'emergency' bypass that allows the fan to spin at normal/ full speed. This allows you to temporarily override the low speed mode without actually changing the setting (desirable if you want to make a 'gaming mode' override function without having to tune the pot setting again). Another alternative is to pull-up the reset pin with a resistor and using a switch to ground it. Yes, I'm aware that the Intel specifications don't apply to all fans but it's a good starting point if anyone wants to work with fans that conform to those standards. This means brands that cater to the enthusiast crowd like Arctic Cooling, Gelid, Xigmatek etc. that produce fans meant specifically to be driven off the motherboard and would thus, conform fully to Intel specifications. Industry oriented fans like the Sanyo, Nidecs and Deltas are from a different generation and have different set of specifications where the automation products would conform to the specific fan used by reprogramming the PLDs or PIDs. Servers which use such fans generally have their own customized controllers with thermistor controls and LCD readouts. A little OT: There are some nice fans from the server market that use a variable voltage on the 4th wire for control rather than PWM. Those can be quite fun to play with too but are more rare (usually from the DEC Alpha and non-HP-Compaq Proliant days).
Awesome, thanks for the equation. Speaking of these rare non-PWM fans with four wires, someone at OC forums has one - it's a Papst with a 0-10v control wire and looks VERY old lol.
Works with my high-power fans pretty good, even with more voltage (with regulator) Here my 120mm Papst fan. -24V -5A -120W -570 m³/h -11000rpm And my other 172mm Papst fan. -48V -3.125A -150W -950 m³/h -9200rpm Greets
I am interested in making a pwm board but instead of manual rheobus style contol knobs I want to use a motherboard header for the speed control and seperate power input off a molex connector so not drawing power away from the motherboard.(since I need to control 14 120mm fans). I will have 4 triple 120mm rads so the fans will be in lots of 3 so I need 5 headers on the board. Will that be to much power draw for a single run of wire for the power( Think molex uses 18awg wire). Currently looking at yate loon fans. Also do the fans need to be pwm capable or can I get a silicon chip to convert 3 pin fans to pwm? I have seen sunbeam fan controler for 3 pin fans with pwm control and wondering how I can achieve the same result. I have no idea where to start anyone able to help me get started?
Tricky! Motherboard control of fans is usually pretty finnicky which is why I built my own voltage controller. I currently have a Rampage III Extreme motherboard which has 7 (yes, seven) PWM fan headers, but the disadvantage there is that only PWM style fans can be controlled. To control my Gentle Typhoons, I had to build my own voltage controller. The sunbeam fan controller, like others, uses PWM on the supply voltage which works fine with low powered fans but is not ideal for higher power fans, as I demonstrated in an earlier post in this thread. What fan headers does your motherboard have?
It has two CPU pwm fan headers. Was planning on connecting one for the pump control and the other for fans( the pumps are the mcp35x they have two connectors one for power the other for pwm. ) I also want to build a board for an led header/termination block that I can control via a dmx rgb led driver.
