Mod temperature controller for min and max volts

johnkenyon · Post by **johnkenyon** » Sun Jan 05, 2003 4:09 pm

Hi,

I am using the temperature based fan controller circuit from http://www.heatsink-guide.com/tempcontrol.htm to control my 3 case fans and a seperate controller for the CPU fan. I would like to update the circuit to allow a minimum and maximum voltage so that I can get the CPU fan spinning straight away on startup and limit it to a certain speed.
I have verified that 7V it enough to drive the circuit and the fans so that fixes the upper limit, but I am more unsure about the lower limit.

Anyone got any ideas?

Thanks,

John

NotSilentYet · Post by **NotSilentYet** » Sat Jan 11, 2003 9:44 am

I am not sure what you want to do but the way I understand it is that you want to be able to limit the lower and upper voltage the fan receives. I am assuming that you have some knowledge in basic electronics.

What you could do is to connect a diode from the drain on the FET to the 5 V supply. You could use a 1n400x to do this. It would be connected so that the side with the ring would be connected to the 5 V supply. Be sure to connect this the right way because if its connected the other way around it would basically short the supply. Doing this would give your fan a voltage range from 6.4-12 V. To limit the uppder voltage you would have to limit the supply to the circuit this could be done by putting a few diodes in series with the supply. Each diode will drop the voltage about 0.7V

I haven't tried this circuit myself so I can't guarantee you it will work. I have tested it in a simulator were it seems to be working fine.

Henrik

powergyoza · Post by **powergyoza** » Sat Jan 11, 2003 12:15 pm

I'm thinking about the same thing - how to modify the heatsink-guide circuit so that it always supplies a minimum voltage.

I found a circuit and posted a link in this thread in the fan forums: http://forums.silentpcreview.com//viewtopic.php?t=2511

The only thing is, I like the simplicity of the HG circuit. NotSilentYet, that's an interesting idea, but is there a way to not use the 5V line. I tried modifying the HG circuit with a zener diode, I have little knowledge of electionics. Anyone care to pitch in?

NotSilentYet · Post by **NotSilentYet** » Sat Jan 11, 2003 2:20 pm

That's a good idea powergyoza, using a zener instead. I think the circuit below might work. It seems to be working in the simulator atleast.

Again, I have not tested this ciruit in reality which means there is a chance something will fry. So test this at your own risk. If possible test it outside the case with another power supply so that you dont fry anything in your computer. I might fiddle around a bit with this myself but I don't think i have any suitable FET transistor handy. I might give it a go with a BJT instead.

The zener diode in the schematic is a 1.3 W part so should be able to cope with a normal fan up to 3 W I think. The lowest voltage the circuit will put out is 12-Zener voltage. So with a 6.8 V zener the lowest voltage would be 5.2 V.

powergyoza · Post by **powergyoza** » Sat Jan 11, 2003 4:29 pm

That's it! That's what I've been looking for! Thanks NSY, your schematics are very valuable. Are you using PSpice or something to do the circuit simulations?

NotSilentYet · Post by **NotSilentYet** » Sat Jan 11, 2003 6:19 pm

I'm using circuitmaker to do the simulations ( which uses the Pspice engine) and eagle layout editor to do the schematics. Have you tested it yet? I only had a 1k NTC thermistor which because of the current through it heated itself thus making the fan spin slower even if the thermistor got hotter defeating the purpose of the circuit. I managed to get it somewhat working by letting the fan blow at the thermistor and it seemed to work with the zener as well. If you try it let me know how it goes.

powergyoza · Post by **powergyoza** » Sat Jan 11, 2003 10:36 pm

NotSilentYet wrote:I'm using circuitmaker to do the simulations ( which uses the Pspice engine) and eagle layout editor to do the schematics. Have you tested it yet? I only had a 1k NTC thermistor which because of the current through it heated itself thus making the fan spin slower even if the thermistor got hotter defeating the purpose of the circuit. I managed to get it somewhat working by letting the fan blow at the thermistor and it seemed to work with the zener as well. If you try it let me know how it goes.

That's weird b/c with my 10k NTC, the fan keeps speeding up. And I thought it was because of the NTC heating itself. One of us must be using a PTC.

I tested with a 5.1V Zener. The zener works thanks! Just gotta troubleshoot the NTC...

powergyoza · Post by **powergyoza** » Sun Jan 12, 2003 3:15 am

UPDATE
I yanked out my old thermistor and put in a different one and the circuit works! Horray! My old thermistor must have been rated for too low a current, so it must of heated itself too much...

johnkenyon · Post by **johnkenyon** » Sun Jan 12, 2003 3:39 pm

Hi,

Thanks for all the tips. I'll try this lot on my CPU fan. The only worry I have is my case fans. I have 3 Panaflos running off one of these controllers. I think that it would be too much load for 1 zener diode. Is it possible to connect more than one zener in parallel to increase the handling? (My electronics knowledge is very rusty)

Cheers, John

Bat · Post by **Bat** » Sun Jan 12, 2003 4:01 pm

powergyoza · Post by **powergyoza** » Sun Jan 12, 2003 4:03 pm

Now the question is, where to get zener diodes on a sunday! I've got the itch to finish my circuit with a 7.5V zener!

NotSilentYet · Post by **NotSilentYet** » Sun Jan 12, 2003 4:41 pm

Johnkenyon

Actually I don't think it's recommended to connect more zeners in parallel because they all have a slightly different voltage. A 4.7 V zener for example may differ from 4.4 V to 5.0 V. This will make the current sharing unequal. Although I'm not sure to which extent. In this case it might just work. Give it a try and check if any of the diodes gets excessively hot. If not then it's probably safe.

Bat · Post by **Bat** » Sun Jan 12, 2003 5:48 pm

Oops, didn't think of that.
Will that be true even if it's the same model?

Alternatively, how about using a transistor along with the zener diode?
Or rather, just put the zener in parallel with the thermistor, instead of where it is?

NotSilentYet · Post by **NotSilentYet** » Sun Jan 12, 2003 6:31 pm

Well I did some experimenting with two 4.7 V zeners in parallel in series with a single 5.6 Ohm resistor. Then I measured the temperature of both zeners. They were fairly close although one diode got slightly hotter than the other. This indicates that they dissipate about the same amount of power. So if you use a few more than you usually would it's probably ok.

However the current draw of a panaflo L1A is only 68 mA. I did some simulations using a 58 Ohm resistor to simulate the fan load. With a 6.8 V zener diode this only came out to 0.6 W dissipated in the zener. So if you are using a BZX85 which is rated for 1.3 W you should be ok with a single zener.

If it were me I would just test it with 1 fan checking the temp on the zener then attach a second fan and check the temperature again and so on. The power dissipation reaches its maximum when the pot is turned so that the fan gets the lowest voltage.