Potenciometer for fan regulator

[Behemot]

I am thinking about making some fan regulators, for start some of the most basic ones with just transistor and potentiometer. Eventually some PWM and also PWM control for 4pin fans (found great circuit here http://forums.bit-tech.net/showthread.php?t=219273).

I am just thinking, since the transistor opening is basicaly more or less exponential, and also some experiences with driving it using linear potenitometers confirm that. Would logarithmic potenciometers be better? Price difference is neglible and if it can provide interesting revolution control on interval better than like 40°, it would definitelly add some juice to the product

[larrymoencurly]

Re: Potenciometer for fan regulator

Have the potentiometer control a MOSFET or an LM317 type adjustable voltage regulator instead, or use a 5W wirewound pot of maybe 50-100 ohms directly in series with the fan?

[Behemot]

Re: Potenciometer for fan regulator

Can't use regulator, that would need at least 14,5 V, or cut me output down to 9,5 V. Even ordinary NPN transistor for 20 cents would give me at least 11 V, more like 11,5 V in reality.

About controlling FET, would that offer me anything interesting to not buy 20 cents NPN transistor? Gonna source BD437 ATM.

Anyway, you did not really answered my question, you are aware of that, aren't you?

[ben7]

Re: Potenciometer for fan regulator

Can't use regulator, that would need at least 14,5 V, or cut me output down to 9,5 V. Even ordinary NPN transistor for 20 cents would give me at least 11 V, more like 11,5 V in reality.

About controlling FET, would that offer me anything interesting to not buy 20 cents NPN transistor? Gonna source BD437 ATM.

Anyway, you did not really answered my question, you are aware of that, aren't you?

lol

Can't you use an LM555 pulsing a MOSFET?

[Behemot]

Re: Potenciometer for fan regulator

Why would I do that? Sorry but can you understand written letters? Obviously not. I asked for something yet you push me in totally different things. I clearly said I will do PWM controller and PWM driving in future. Now I want to make simple transistor/potenciometer circuit.

Even many fan controller makers do this because yes, it eats some power, but even school child can make it work. So I would be glad if you answered for what I asked, and if you do not know, rather say nothing. Thank you very much.

[ben7]

Re: Potenciometer for fan regulator

Why would I do that?

Because you want to control the fan speed with PWM?

[Behemot]

Re: Potenciometer for fan regulator

Because you want to control the fan speed with PWM?

As edited, I DO NOT.

[ben7]

Re: Potenciometer for fan regulator

As edited, I DO NOT.

jeez ok ok ok.

Well, your gonna have a LOT of heat dissipated in those transistors then with the fan at low speed!

Or, do you think making an adjustable current source (using two transistors) would work well?

[Behemot]

Re: Potenciometer for fan regulator

Well, there are people out there in the wild using 500 mA transistors in tiny TO-92 package. I will have 4A SOT-32 transistors with heatsink so it is not a problem too;-) From my tests it is not even so dramatic, considering usuall current draw for a fan, we are talking about couple of watts.

[larrymoencurly]

Re: Potenciometer for fan regulator

Can't use regulator, that would need at least 14,5 V, or cut me output down to 9,5 V.

It's too bad there are no such things as low dropout adjustable regulators, like from the LM1117 series.

Anyway, you did not really answered my question, you are aware of that, aren't you?

A log pot traumatized me when I was a child, and I was able to supress that memory until you started this thread. Why not test by taking a linear pot and turning it into a log pot by wiring a fixed resistor between the center lug and the outer lug?

[mariushm]

Re: Potenciometer for fan regulator

There are LDOs like Micrel's MIC2941 which have up to 0.35v drop at 1A (they do max 1.25A) ... at about 50-100mA they can have as low as 0.05v drop.



An equivalent that's probably easier to get is lm2941 but this one needs 0.5v @ 1A and about 0.4 @ 0.5A:

http://uk.farnell.com/national-semic...220/dp/1564645


Depending on what resistors you select, you can get away with a linear 1kOhm potentiometer, but you could also go with a 10k one.

Linear regulators like these are super easy... they change the voltage. Thing is they generate heat, basically they dissipate (Vin - Vout) x current.

Most fans are 0.2-0.4A so at the worst case where you probably run the fan at 5v, then you'd have (12-5) x 0.4A = 2.8 watts dissipated as heat.

So for these ~ 3 watts you kinda need at least a small to220 heatsink, just to increase the surface of the metal tab on the back.... or solder the tab on the ground copper filling (tab is ground on these)

[Behemot]

Re: Potenciometer for fan regulator

Gave a try to log pot and it is nice, clearly the fan RPM drop is more linear as the pots resistance increases logarithmicaly. Only strange thing is that with linear, I could stop the fan completelly on maximum resistance. Now I have minimum output around 5,5 V at 11,4 V on +12 V.

The thingy is, I have changed transistor for a 45 V now (60 V before), is that it reacts sooner (=at lower) to voltage on gate?

[ben7]

Re: Potenciometer for fan regulator

Gave a try to log pot and it is nice, clearly the fan RPM drop is more linear as the pots resistance increases logarithmicaly. Only strange thing is that with linear, I could stop the fan completelly on maximum resistance. Now I have minimum output around 5,5 V at 11,4 V on +12 V.

The thingy is, I have changed transistor for a 45 V now (60 V before), is that it reacts sooner (=at lower) to voltage on gate?

If the voltage provided to the base goes below about 0.6v, then the transistor will start to turn off. If it turns off too much, then the fan will completely stop.

The transistor will react just the same, unless it is a darlington.

[Behemot]

Re: Potenciometer for fan regulator

So why does it give around 5,5 V minimum? Linear pot on the same resistance (100K) got, don't know how much, can measure, but the fan was long siting still by the time pot reached 100K…

[ben7]

Re: Potenciometer for fan regulator

So why does it give around 5,5 V minimum? Linear pot on the same resistance (100K) got, don't know how much, can measure, but the fan was long siting still by the time pot reached 100K…

I don't have a clue, never used log potentiometers before

Could you draw a circuit of what you have?

[budm]

Re: Potenciometer for fan regulator

Can we see the circuit you are using?

[Behemot]

Re: Potenciometer for fan regulator

Most simplest exists: NPN transistor's collector and the pot sits on +12 V. Other leg of the pot drives base. Load is getting voltage from emiter.

Basically transistor eats power I do not want and dissipates it as heat. The power wasted is not drastical anyway (few watts get lost easily and most likely are the same as with resistor, just I can change it now) and with 4 A transistors with heatsinks (assurance is machine gun) I do not care really OFC I will make it better, but all with time

[budm]

Re: Potenciometer for fan regulator

With that set up by having bias resistor (100K) between the Base and Collector, the transistor will never go into cutoff mode because it will always have bias current flowing through the load/ Base-Emitter junction/100K resistor back to the 12V source. And when the pot is at 0 Ohm, it will have a lot of current flowing through the Base/Emitter junction as well, may even exceed the transistor rating.
See simplest drawing I made that may suit your needs.

[budm]

Re: Potenciometer for fan regulator

"The thingy is, I have changed transistor for a 45 V now (60 V before), is that it reacts sooner (=at lower) to voltage on gate?" By the way the beta of the transistor will have an effect on the Ic current (Collector current). If one transistor has beta of 50, another one has 100, the one with higher beta will cause more Ic current to flow with the same bias resistor value.

[Behemot]

Re: Potenciometer for fan regulator

Base current is as high as 1 A for this transistor (BD437), that is not a problem. Emiter cut-off 5 V, base-emitter 5 V/10 mA, 1,2 V/2 A. As for the resistor, been already thinking about it, that's true, for limiting current just for case, the thing is, it will lower maximum output voltage. But at 1k it may not be so dramatic anyway…I think the zener is not even needed in there if you limit the current through base with resistor.

The cap after transistor is 0.1 uF/100 nF?