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

[jocanon]

The 12v supply says it goes up to only 1 amp. I just power everything on at once. Is there some simple way to make them start slower and then ramp up, or one start first? I was planning on putting together a temperature controlled circuit for the fans based on ()blivions circuit. I wonder if that will solve the problem because then they will ramp up in current slowly instead of drawing it all at once.

 

[ronv]

Let me rummage back and find ()'s circuit. For now you might try hooking them up in series. They may not start or maybe they will, and run at ~ 1/2 speed.

 

[jocanon]

I have a link to his circuit on the OP. I just got done putting it together, I will let you know how it goes.

 

[ronv]

First page..... should have known.
It looks like an on / off circuit so if you have one for each supply you would have to be pretty unlucky to have both turn on at exactly the same time and on power on. Even if they did the supplies are probably cool at power on?
If you only have one we could probably add a cap to keep it off until the supply is all the way on.

 

[jocanon]

The fan speeds up as the temperature rises on ()blivions circuit, so I think it will draw less current when it starts because the fans are running slower.

 

[jocanon]

The tempurature controll circuit seems to have fixed the problem.

 

[()blivion]

Putting an inductor in series with the fans *MIGHT* give you a starting current boost. Something like this...

View attachment 67214

!!WARNING!! I haven't actually tried this at all, so it *COULD* fry your fans. I don't think it will though.


EDIT: Or the temperature controller circuit fixed the problem, never mind P)
EDIT2: Actually... LOL.... this wouldn't work anyway. You would need a capacitor too, and you would need to hook it to only one fan.

 

[()blivion]

It looks like an on / off circuit so if you have one for each supply you would have to be pretty unlucky to have both turn on at exactly the same time and on power on. Even if they did the supplies are probably cool at power on?
If you only have one we could probably add a cap to keep it off until the supply is all the way on.

It's actually a linear controller that varies the impedance of an N-channel MOSFET to control current, and thus fan speed. Since it limits current, it's a no brainier that it would fix the before mentioned problem now that I think about it.

Meh... all's well that ends well.

 

[jocanon]

Yeah. With mains power I normally use a DPDT switch. This disconnects both the hot and neutral. Both will shock the crap out of you in reality, so I'd just as soon unplug them both.

()blivion, did you mean to say you normally use a DPST switch for mains?

 

[()blivion]

Yeah, DPST works fine. SPDT are more common. You CAN use a single pole single throw if you don't care. But you won't be isolating one of the two lines, so you will need to know which is neutral and which is hot. Both neutral and hot will happily electrocute the crap right out of you given proper opportunity. Though neutral is far more safe.

If you can't get or do the switch thing. You should just unplug it when you want to work on it. But with electricity it usually takes two [paths] to tango no matter what. So if you ABSOLUTELY have to work on it while it's plugged in, keep one hand in your pocket and you are MUCH safer, Also, remember that the capacitors they use in the corded end of most power supplies have ~300 volts on them at any given time, (voltage-doubler) and *CAN* hold that charge for several minutes after the device is unplugged. Needless to say they can kill you dead if you disrespect them.

 

[jocanon]

I am really not having much luck with the temperature control circuit on these fans. It works great on the smaller fans in the server PSUs, but I think the parts I am using are not well suited for the larger 24v fans. It is way too sensitive. The fan is either off or on, and then once it comes on I can't get it to slow back down and go off even when it cools. I think I may just put a pot in there to control the fans manually and maybe another volt meter wired to a temperature sensor like we did for the FETs so I can monitor it to make sure it's not getting too hot. The only time I will really care about slowing the fans down is when I am tuning the fan speed on the DUT so I can hear the fan better on the DUT because the dummy load fans are quite loud. Other then that, I will just leave the fans on the dummy load on full blast anyway.

 

[jocanon]

So, I was doing some tests and it appears that the fans are completely off when I have 500 ohms of resistance in series with the +24 volts from the power source and the fans positive leads. So, if I understand ohms law correctly, I would need a 500 ohm variable pot rated at at least 1.15 watts given that the power source is 24 volts. Then that would give me the ability to adjust the fans manually from fully off to fully on. Is that right?

 

[()blivion]

I am really not having much luck with the temperature control circuit on these fans.

Odd, I can't remember if I have used this circuit for 24 volt fans before, so you may be in uncharted water. But even with 12 it can happen that you only get a on/off control. When that happens it usually means there is too much gain in the temperature sensing section. So what I usually do when this happens is reduce the resistor that is parallel with the thermistor substantially, then readjust the pot. It's possible that you have to use as low as a 1K resistor to get it to work right. It could also be that some parts are broken, so swap things with fresh parts every now and then to check for this. If the FET got fried from static and the gate is leaky it can do all sorts of odd things. Also, under certain circumstances the trimmer pot will get damaged if you turn it to far one direction. This damage can make it really hard to get the voltage divider in just the right setting for biasing the FET. Finally, switcher FET usually have bad linearity, and some times it's so bad that all you can get is on/off from it.

Bottom line, I would say readjust and swap parts.

So, I was doing some tests and it appears that the fans are completely off when I have 500 ohms of resistance in series with the +24 volts from the power source and the fans positive leads. So, if I understand ohms law correctly, I would need a 500 ohm variable pot rated at at least 1.15 watts given that the power source is 24 volts. Then that would give me the ability to adjust the fans manually from fully off to fully on. Is that right?

Yeah, that sounds about right. You can always do this if you want to. My circuit wastes power just like a pot does. So there is no difference really. You can also set it up as a voltage divider instead if you would rather be able to control the fan with a varying voltage instead of a varying current.


NOTE: It's possible that your fans will not work by limiting the current going to them, though I have NEVER seen a fan that did this.

 

[jocanon]

It is still temperature controlled, not just on or off entirely, but it is so hyper sensitive that it is almost on/off. I tried lowering the resistor between the thermistor, I think I did go all the way down to 1K, I will try that again though because I don't remember for sure how low I went.

 

[ronv]

It may be the big fans. Quite a bit of current. When the FET heats up the threshold voltage goes down so it wants to turn on more. Try the fan in the source, maybe that will work better.
Don't know what FET you are using but it will want a heatsink. This one should be 8 or 9 volts minimum and 20 volts or so maximum. It's probably not good to try to let the fan just stall (or fail to start) things get kinda hot.

 

[jocanon]

It may be the big fans. Quite a bit of current. When the FET heats up the threshold voltage goes down so it wants to turn on more.

Yes, that makes perfect sense, it explains the way it was behaving exactly.

 

[jocanon]

Ronv, I want to make sure I am understading the schematic correctly, is R3 the fan, R1 the thermistor, and U1 the pot and it goes from 25k ohms down to .99 ohms? Is R2 just a normal 33k resistor that goes between the pot and ground? Btw, the FET I am using is IRFZ24N.

 

[ronv]

Right you are. Do you have the fet on a heatsink?

 

[jocanon]

no, I do not, but I should probably get one as the FET gets very hot.

 

[ronv]

I bet it does.
You will need one at about 12 sq. inches of surface area. (count both sides) 6 C per watt
It will still be toasty but ok.