Fan control circuit

[staze]

All,

I'm attempting to repair a faulty fan control circuit on a computer ATX PSU (Antec TP3-550). Attached is a schematic I did, but I'm pretty sure it's wrong. The transistors are guesses (pretty sure the top NPN is right, but the bottom NPN, I think, is wrong. The program I used didn't have a icon for a thermistor, so I just dropped in a pot. Anyway, the thermistor at "room temp" reads 5.5k. As the power supply heats up, the fan should increase in speed. Right now, the fan doesn't run at all (so I'm guessing one of the BJTs is dead), but I was trying to figure out the circuit before I just started pulling them. I also fully admit one of them (or both) may be a FET, but that doesn't seem right for a fan control (since BJTs are easier to operate linearly, especially with a "dumb" circuit like this.

Please, I need input here, because right now, nothing I do seems to allow me to control the fan speed with the thermistor (guessing because that top BJT is just saturated, and the fan provides an easier path to ground than the lower one, but that makes no sense).

Thanks!

[staze]

No wonder it wasn't making any sense... the lower BJT is actually a KA431AZ, or Programmable Shunt Regulator hooked to the thermistor. The top is an HI3669 (finally got good light to see it). it, at least, is wired how I indicated.

Why would it be involved in this circuit?!? The reference seems to be wired to the thermistor, and the other pins to the fan neg, and emitter of the BJT. So, it's serving as a "switch" of some type? I'll model it based on the block diagram from here: http://www.mouser.com/ds/2/149/KA431A-185801.pdf

[staze]

Pretty sure this is the circuit. Any thoughts? it seems to behave as it should by adjusting the temp on the thermistor (yes, I found the thermistor in the program).

Doesn't bring me a whole lot closer to figuring out what's bad. I do note that the 18K resistor measures as 7K in circuit... there must be some divider at play that I'm not seeing. It makes sense that it does, as well, since setting it to 18K makes the whole thing run current limited, but changing that resistor down to 7K makes things work properly.

[staze]

Last question: Should I be able to replace an HI3669 (http://www.hsmc.com.tw/pdf/HI3669.pdf) with a C5707 (http://pdf1.alldatasheet.com/datasheet-pdf/view/38689/SANYO/C5707.html)? Specs look similar from what I can tell... reason I ask is that I have the C5707's...

Thanks!

[richcj10]

I think it would work. it would be a better part.

[mij59]

The fan should run if the supply voltage is 8V, does ít ?

Don't make in-circuit measurements, their lead most of the time to more confusion.   

[staze]

The output of the circuit with no fan attached is only 5v. With a fan attached, it's only 2v.

I'm guessing the BJT overheated, or somehow got reversed, and the beta has diminished to nearly nothing. But I haven't desoldered to check. Was waiting to see if that c5707 could be used as an alternate.

[mij59]

The output of the circuit with no fan attached is only 5v. With a fan attached, it's only 2v.

I'm guessing the BJT overheated, or somehow got reversed, and the beta has diminished to nearly nothing. But I haven't desoldered to check. Was waiting to see if that c5707 could be used as an alternate.

Did you measure the 12V supply voltage ?

[staze]

Yes, it's fine.

[mij59]

Remove the transistor and test it.

Whats the current rating of the fan ?

[staze]

12v, 370mA. The question is still, would the c5707 make an okay replacement?

[mij59]

12v, 370mA. The question is still, would the c5707 make an okay replacement?

First you need to find the fault, than think about a replacement.

The thermal resistance of the transistor is really the only thing that matters.
The c5707 has similar properties, if you got it, try it, although  a TO220 transistor would be better.

[staze]

True. I just figured if I remove the BJT, and it's bad, then I've got to figure it out. :p

Thinking about it more, it could be the ka431, since I'm sure the failure is due to overloading the fan power "bus". But I figured since the BJT ultimately drives the bus, I figured it was likely. We shall see.

[staze]

Guess was right. Removed BJT, Hfe measured as 2! Found a C5707 that measured at 320 or so, which puts it in the range of the HI3669, and dropped it in. Fan fired right up. Yeah, a TO220 package would be better, but I'm not sure if a TIP121 or something would work in there (pretty high beta)...

Anyway, it's back to working, so I'll call it good for now, and remember not to overload that fan bus again (specs say don't put anymore than 2 other fans on the bus, I had put 3 since I didn't read the manual). =P

Thanks!