Help with PSU (Temp control fan, load bank, & PWM circuit)

If I run both fans off the same circuit they will be in sync and I think it sounds better, plus I can save a little on parts. I will still need the single circuit for 12v PSUs, so I guess I will need different circuits for 12v versus 24v supplies it sounds like.

Actually, I think you might be better off with the op amp circuit. Have you built that one yet?

That one would run 2 fans without changes... I think.

It would still be nice to get the measurements with one fan just to confirm.

Still waiting to get the parts on order for the op-amp circuit.

I'll just report back the readings that changed with just one fan:
Drain=4.2v
Source=4.2v

I also measured the voltage going across the fan itself (one on the neg fan lead other in positive) with two fans it was 6.57v with one fan it was 8.23v.

There must be a poor connection there somewhere. To drop 4.2 volts across 24 ohms takes .175 amps. The fans say they are only .06 amps. This implies some additional resistance somewhere in the source to ground.

Well, I remember you said before that the resistor is only operating at 30% of capacity and will get a little warm, well it's getting really really hot, I mean so hot it might catch fire it seems like. I swear it is only 24ohms though .

I think I see what might be part of the problem, the fan is .6 amps not .06 amps. I am looking at one right now.

Wow! I need new glasses. That thing must scream!

I have to go now, but you might try it with 7.5 or 10 ohms in place of the 24 ohmer. Or worst case just the 2.2 ohm.

Lol, yeah, I could probably roast marshmallows on the resistor right now..., the fans do go pretty fast, that's why people complain about how loud they are and why I think there is a market for the PSUs with the temp control fan to slow them down in a smart way, versus just slowing them permenantly. I should probably call it a night now anyway. Thanks!

()blivion.

Seems like several little things. Evidently the big supply has a little bit more noise on the 12 volt line. This was made a little worse because a start cap was added fron +12 to the gate to make the fan start on power up. The circuit has a small linear region (.1 .4 volts depending on the FET) so the noise was enough to turn it on and off. We added a small resistance in the source to make the circuit less sensitive (~1.2 volts from off to on). This made it less sensitive to the noise but also had the nice effect of making the temperature band wider. Now instead of 3 or 4 degrees from off to on it is more like 25 or 35 degrees. I can send you the spice if you want to play with it.

Click to expand...

Sure, I'll take improvements to the design any time.

So, if my ohms law skills are working right...I think I would need a 1.764 watt resistor since I will be passing about 4.2volts through 10ohms, is that right?

Edit:

I belive with the op-amp circuit the power will all be dissipated in the transistor so I will not need a 1.764 watt resistor, right? However, I may need to heat sink the transistor, which I can easily do by glueing it to the case with non-conductive epoxy.

2nd edit:

I am very confused...I took the measurments last night and I was getting 4.2 volts from ground to source with one fan, now I am only getting 2.2 with one fan and 3.52 with two fans, hm, I am sure you guys are probably loosing confidence in my ability to measure correctly...I don't know why but that's what the meter says. If it's only 2.2 I would only need a 0.484 watt resistor with one fan and 1.239 watt resistor for 2 fans. Maybe be all irrelivant if the op-amp circuit is what I end up going with.

3rd edit
Ok, I feel a little silly...of course the voltage changed because I changed the 24ohm resistor to 10 ohms! Duh...sorry, if you just give me enough time I guess I will answer my own stupid questions .

We are kind of getting into a box here. What would be nice is to have 1 size fit all, but the one we are working on now will need resistors to match the number of motors. And the op amp circuit looses some voltage at maximum speed. A 2 or 3 volts out of 24 may not be to bad but 2 or 3 out of 12 may not be so good.
Your right the resistor will need a higher wattage in the long run. But for grins and giggles you might try it with 2 10 ohm resistor in parallel. We would like the drop across the source resistor to be a little over a volt.

I am sure you guys are probably loosing confidence in my ability to measure correctly

Click to expand...

Actually I was sorta almost losing confidence in my little circuit if anything.

I have never had the problem with this circuit that you appear to be experiencing. But then again, I have never used it for 24 Volts either I don't think. Kinda makes me wonder if it will work for your project at all. It really should, I don't see why it wouldn't. You CAN tune it for almost anything *IF* you know what your doing. It's less of a "build it exactly like this" type circuit and more of a "play with it till it works" type circuit. And I have had them be very picky, but they almost always work from room temp up to 100c with any configuration you build them in.

I came up with the original circuit via trial and error specifically for PC cooling fans. And with PC's, a few degrees C can mean a lot. CPU's now die at 70c so you really want to make sure you have the sensitivity. With power switches in SMPSs, the temp-of-death is a lot higher. So often times they let them run hotter.

ronv said:

We are kind of getting into a box here. What would be nice is to have 1 size fit all, but the one we are working on now will need resistors to match the number of motors. And the op amp circuit looses some voltage at maximum speed. A 2 or 3 volts out of 24 may not be to bad but 2 or 3 out of 12 may not be so good.
Your right the resistor will need a higher wattage in the long run. But for grins and giggles you might try it with 2 10 ohm resistor in parallel. We would like the drop across the source resistor to be a little over a volt.

Click to expand...

Actually, 12.5v less 2 or 3 might be fine on my set up. I mentioned it a couple posts back...many people that use this exact PSU to power lipo battery chargers end up shorting pins 4 & 8 which has the effect of slowing the fan down to standby mode speed permanently and they claim to have had no problems with overheating the PSU when doing so (actually, when shorting pins 4&8 the fan does speed up very slightly under full load, the change in speed is so slight that it is almost not noticeable). But I think that permanently slowing the fan could have a negative effect on the life of the PSU, so it's probably a good idea to put the temp control in there so the fan will speed up when it gets hot...anyway, the point I am trying to make is that if in practice the fan has been slowed to standby mode without reports of the PSU failing when used on a real life setting for our purposes, I think having a top speed at 9v or 10v would be just fine, in fact perhaps desirable because as ronv said earlier, at 12v that little fan screams=very noisy=annoying. I don't know what the voltage across the fan is when it's in standby, I would guess around 4v, I can check when I get home.

()blivion, I think the high frequency switching noise when I ramp up the current in my PSU (as seen in the youtube video in posts 512 & 516) makes it so I can't use the original fan temp control ciruit as is since every time I adjusted the current it had a dramatic effect on the fan speed independent of temperatur changes. That problem probably doesn't exist in a normal computer fan configuration.

On another note, I think we are OK loosing a couple volts. The fans ussually plug into a spot on the PCB inside the PSU, but I have spliced the wires and have them connected to the 12v rail (really 12.5v). So I measured the voltage going across the fan when plugged into the manufacturers normal spot on the PSU, at full on there was only 9.3v going across the fan. In standby there was only 4.6v going across the fan, as I state above, many guys run these supplies with only 4.6v on the fan. Conclusion...I think it is absolutely OK to loose even as much as 3v because of the way I am hooking it up, that would be 12.5-3=9.5v on the fan. If I can just use the op-amp circuit ronv built for the dummy load on the PSU as well, would that be a one-size-fits all for both the 12v and the 24v configurations? Meaning on the 24v I would power two fans off one circuit, on the 12v PSU I would obviously only power one fan. Is the op-amp circuit a constant voltage circuit so that no matter if I have 1 fan or 2 fans it will still have the same amount of voltage to the fans? If so then I think it would be a good one-size-fits all solution that would work.

OK.

Yep, same one for 12 or 24, 1 or 2 fans. I think you can probably eliminate the pot once your sure about the temperature. Let me look into the best way to implement the start pulse, unless you want to let them run at low speed all the time. I also need to check the lowest setting on both 12 and 24 to make sure that fits.

I think I know where the disconnect is with ()'s circuit. I think it is to be put in the air flow where 3 or 4 degrees might indicate a pretty big change in heat sink temperature.

The PSU fan starts at 5.33v and stalls at 3.3v.

OK, first of all, sorry for the long post.

Let me explain how I am thinking I am going to set this all up and let me know if it makes sense or if I should change anything...I am thinking for the 24v supplies (just 2x12v run in series) I will put the temp control fan circuit inside one of the PSUs and run wires to control the fan in the second PSU, so I will only have 1 thermistor. I think this should be fine as both PSU will share the current and they should be running at close enough to the same temperature. I mean as long as the fan is blowing on them within a pretty broad range I think it is ok, meaning, I don't think the fan speed has to be so exact that I need seperate temperture readings for each PSU. Another part of the circuit that we haven't been talking about much, but I want to bring up just to be sure we cover all the bases is the 555 timer. When I disconnected the fan for the first time I couldn't get any power on the 12v rail, come to find out, the PSU has a safety feature built in so that if it senses that the fan is off it shuts down power. So I had to trick it with a 555 timer circuit to send pulses to the tach to make it think the fan is on. I think I will be able to only have one 555 timer circuit as well and send wires from it to the second PSU. Here is the 555 timer circuit I am using (picture attached below, I am using the first of the two circuits on the page with penciled in values for resistors and caps), let me know if there is an easier way, but this circuit is already pretty simple.

As far as a start pulse, that would be good if there is a way to make them start smoothly. I was thinking I would leave them running at a slow speed all the time because I didn't like they way they sounded when they were starting, they make a lot of faint clicks for a minute or two until it finally heats up enough to start. If there is a way to make them start without making any of those clicks first then that might be the way to go, otherwise I think it would be best to leave them running a low speed all the time.

As a note of interest, I don't think it will effect anyting we are doing here, but there is a guy in the other forum from Australia that works for a company that had a ton of servers with this model of PSU in them. A few years ago when the PSU was new they bought them and put them in there servers and left them plugged in but the servers were not turned on for a couple weeks. So the PSUs were in standby mode and the fans did not come on at all in standby mode. They came back a couple weeks later to turn the computers on and the PSUs had all failed due to over heating, they sent them back and got new ones, but this time the PSU's fan came on slowly even when in standby (so HP's original version they fans did not come on in standby, but now they do). Moral of the story, when I was talking about making it so the fans don't come on at first he relayed this to me and said it might not be a good idea, BUT I think my situation is different for a couple reasons. I don't think most hobbyists are going to leave it turned on for a couple of weeks without using it. Secondly, the original HP PSUs did not have a temperature controlled fan in them, with the circuit we design, I think if you leave it sitting long enough it will heat up and turn the fan on eventually. Anyway, I just wanted to give you all the facts so you can help me make a wise decision. I think it is safe to make the fans not come on at all until it reaches a certain temperature (assuming we can make them not click for 2-3 min), but what do you guys think?

Hmm. Here is my 2 cents worth.

I would run the fans at low speed (5.33 volts). I would leave the pot in so I could make sure both fans just started at room temperature. I would then add the start cap to make sure they start even as they age. I would then leave the tach pulse going to the supply so that if a fan goes bad the supply shuts down and you don't have to build another circuit to fool it.

I tried leaving the tach pulse from the fan (the blue wire) plugged in instead of using the 555 timer, but I couldn't get the fan to power on since it is connected to the 12v rail which needs a tach pulse to turn on...Maybe if I leave the fan plugged into the PCB's original spot instead of powering it from the 12v rail, I could do that, but then I run into the problem of not having enough voltage to sacrafice 2-3volts for the op-amp circuit.

Come to think of it that may not work with the supplies in series unless you have a fan circuit for each supply. I don't think they will click if tou make sure they will run at 5.3 volts. I don't think they would have shut down the supply for a bad fan if they didn't need to. The fact that the fans are so high speed says they need them.

Sorry, we posted at the same time. What is the voltage on power up at the fan plug?

Ahh, I think I see it. 4.6 volts.