Hi guys...needed a little help with designing this circuit I have in mind - info below. Note - the 12V input from mobo is from the CPU fan header.
There's a few ways to go about this. the easiest would be just to buffer the output of the CPU fan header with an NPN transistor. The only problem, if you want to call it that, is that the voltage to your fans will be approx. 0.5V to 0.8V less than the output of the cpu fan header. If youre worried about that 0.5V to 0.8V voltage drop things will have to find a rail-to-rail op-amp that is capable of driving enough current to power you fans.
Don't care about 0.8V. So is it as simple as that for each fan? FAN (red) | NPN ----- mobo 12V (red wire) | 12V PSU (red) As you can see I don't know much about circuits/components.
Yup...that easy. I'd put a resistor inbetween the mobo and the base of the transistor just in case a fan fails there wont be a direct short between tha cpu fan header and ground. A 100ohm would do just right. I'll try to post a diagram of what i'm talking about.
That would be excellent - any particular npn transistor - 2N2222A/ PN 2222A H/Sp Sw NPN - would that do the trick? Have just thrown up something - let me know if it is correct. Thanks for the help btw.
YMMV, but I don't think it's quite that simple. The motherboard fan speed control is generally by PWM, an on-off pulsed supply, often with a small capacitor to roughly smooth it at a typical single HSF current draw so the speed signal functions OK. Somebody at SPCR forums tried the transistor boost approach and ended up with a dead header. An oscilloscope would be handy to see just what's going off on your board when a single fan is being controlled.
the diagram looks good. Just to note... the collector of the NPN goes to the +12V from the psu and the emitter goes to the red wire on the fan. The 2n2222 is too small, something like an MJE3055 would be better. Also you might need a small heatsink on it as it may get a bit warm.
Ya what cpemma said. From what you've told me the black on the cpu fan header is 0v always and the red ranges from 7V to 12V. If your mobo does use pwm, usually the switching element is on the low side. If this is the case then the circuit will change slightly. Double check your readings. Put the black lead form your multimeter to the psu ground and then measure the voltages on the black and red lead from the cpu fan header. I have my suspicions (sp?) that you'll find the red lead is always 12V and the black will range from 0V to 5V. Anyway just double check before we go further.
Yes it is the red mobo fan header wire that varies. The black lead is always 0V and the red lead varies from 7V to 12V. How do i determine if it is using PWM? alvin
Sorry to keep you waiting, but with all the problems other people seem to be having doing this I figured I better put down my beer and design something a little better. So here's the schematic. This design doesn't care if the mobo uses PWM, linear, open drain, open source, or any other kind of output. Also you should be able to get about 11.9V from the output. Yes its more complex than the original transistor that we discussed, but I figured that you could build this once and then transfer it from mobo to mobo without having to go through this again. Near the inputs from the cpu fan header is a resistor drawn in dashed lines. This is a dummy load that may be required for some mobos if it uses open drain or open source fan control, since the input to this circuit is relatively high. The 200 ohm value i gave this is just an arbitrary value, at 12V it will put have a draw of about 0.75W. You may be able to raise the value of this resistor to waste less heat or if you're still going to have the cpu fan running off this header and just tap into it for this circuit you wont need it at all since that fan will be the load. The picture turned out a bit fuzzy for some reason... i'll see if i can fix it up. Let me know if you have any questions.
The transistor approach should work even if the motherboard IS giving out a PWM signal, however one would neet to place diodes over the fans. In dutch 'vrijloopdiodes', dunno what this means in english. Then there shouldn't be a problem.
That is excellent!# I will try this out over the weekend and let you know how it goes. Again thanks for your help. I'm looking forward to this. Cheers, Alvin
Just to give you a heads up... I designed this circuit using a simulator, I haven't actually built and tested it yet. I just ran the circuit through every possible loading scenario I can think of and I thought I'd pass along some additional info. First off, that 47Ω resistor, you can make that a 1/4W resistor no problem. Secondly, I found that the maximum wattage that 3904 transistor will dissipate is about 310mW which is pretty much exactly half of what the datasheets say that transistor can handle, but anything above half of a maximum makes me nervous... I like big safety margins. Now that transistor should be just fine but it could get quite warm to the touch, maybe enough to tattoo the part# in your finger tip. If you like you could try it cause the worst that could happen is you smoke the 2N3904 and at 5 cents or so a transistor who cares. But I think you should sub the 2N3904 for an MJE182 just to be safe, maybe even put one of those cheapie 25 cent heatsinks on it. Lastly... make sure you give that MJE2955T some good heatsinking. A rule of thumb is that it will dissipate 1/4 of the maximum output wattage so size the heatsinking accordingly. If I get a chance tomorrow i'll actually build the output stage and let you know what I find.
I just built it and all went just as planned. To my surprise the 2N3904 wasn't as hot as I expected. It was quite warm but never got to the point that it was uncomfortable to touch. The 100ohm was getting to the point of being uncomfortable to touch though, but I only had a 1W handy. The closest thing I had for a load to simulate the fans was a 3.33ohm resistor. With this load the circuit was outputing 10.7v, which works out to about 35W. But my 12V power supply actually only puts out 11.5V so that works out to a 0.8V drop. At any rate it handled it just fine. The limitation of the output of this circuit isn't with the output transistor but with the base drive for it so voltage drops in the output transistor start to rise after the 25W to 30W mark. Well good luck building yours and let me know how it goes. I think i'll continue to tinker with this circuit to over come some of its limitations but it will do what you need it to do as it sits.
Building it as we speak. Just a little thing - my electronics supplier didn't have any 2W resistors so instead I changed it for a 5W - is that ok? And instead a 200ohm 2W I have a 220ohm 5W. Hope all goes well. Well let you know a little later. cheers, alvin btw. did I say thanks!
The higher wattage resistors won't be a problem at all, they'll be better if anything. It just means they wont get as warm. Changing the 200ohm for a 220ohm wont be a problem either, its just to give the fan speed control on the mobo some kind of a load that it might need to operate properly if it is a open source or open drain design.
Another question - I got tantalum capacitors for the 0.1uf's (I should have got ceramics). They are polarised and the schematic doesn't specify which way I should put them in (I'm not educated enough to figure out) - which way should they go?
There's no problem replacing the cap shown just below the op-amp with a tantalum. It's called a by-pass capacitor and is just there to improve power quality to the op-amp. Just connect it positive to positive and negative to negative. The capacitor in the output stage is going to have to ba a non-polar one though. This cap is there to help with stability. Most amplifiers are unstable by nature and they can break into oscillation at any moment. This capacitor doesn't "have" to be a 0.1μF. If you have something close hanging around you can use it. Anything inbetween 0.1μF and 0.33μF should work just fine, but it does have to be non-polar.
Built the circuit last week - gruelling hours getting the circuit together. Was in the process of testing and I blew the circuit up - I was shattered =\ . The MJE2955T just touched my case and bang. Rebuilding this weekend hopefully.
