Advice required on Over-Temperature/ Fan Failure detector.

[pinkmouse]

Well, I've built a nice simple thermostatic fan controller using a LM317T and a thermistor for my Krell clone. No problems there. However, I would like to add a logic level out from the PCB if the fan fails, to drive a led or even cut off the power.

The most obvious way would seem to be to use a low ohmage load sensing resistor in the line to the motor, and use a schmidt trigger to drive an inverting flip flop, going high when no/low, or high stall current is flowing. However, before I go for it, I was just thinking I would consult the vast expertise of the EE gurus on the board to see if there is a simpler and more elegant way.

Any thoughts?

[megajocke]

Is it a brushless fan? If so, have you measured the current when the fan is stalled? The fan might shut down and try starting just now and then so maybe it doesn't draw much current when stalled.

[Arius]

Lots of IC's that do what you want now.

Here's an example of a chip that does temp sensing, fan PWM control and fault detection.
http://ww1.microchip.com/downloads/e...Doc/21444c.pdf

Microchip's TC642. Less than $2.

Only snag is the control method is PWM and not linear. Might generate some noise. Need careful layout/screening/shielding maybe.

[pinkmouse]

Megajocke, good point about the stall current, I'll have to check that out. I actually thought for a while about using a LED and a light sensor chip to actually detect fan rotation, but gave it up as too complicated.

Arius, Yes, I looked at some of the Maxim fan controller chips. That would be an easy way out, but I'm trying this as a learning experiment.

[AndrewT]

Hi,
how about a thermal switch into a latching circuit?

[pinkmouse]

That's certainly simple, Andrew, and will probably be the way I end up going. I just fancy playing around with this current sensing idea for a little longer. You know how it is when you get an itch?

[anatech]

Hi pinkmouse,
Try a diode with a resistor across it to detect "fan on". This is across the B-E of a transistor with a resistor in series with E. A thermal switch at 20° above the first one for a fail safe if you want.
Now you can have a pretty light that says FAN.
-Chris

[pinkmouse]

Thanks Chris, interesting idea.

I am away for a few days, but if anyone else has suggestions, keep 'em coming.

[mhennessy]

#1. Use a 3-wire fan (with tacho output)

#2. Power your fan via a small resistor, and use a scope to look at the current drawn by the rotating fan - you should see a significant AC component caused by the comutating of the coils. This AC component disapears when the fan is stalled, although you'd need to check what it does if/when it tries to restart.


Making use of these is the same as a conventional AC-fail circuit - I reckon that a couple of transistors will do you...

(Yes, this is an idea that I've thought about before - let me know if you'd like more help, meanwhile I'll try to find my old notes from the time)

[pinkmouse]

Just thought I'd pop in before I hit the road.

Originally posted by mhennessy
#1. Use a 3-wire fan (with tacho output)

Hey, you think I'm going to betray my Yorkshire roots by actually buying something!

#2. Power your fan via a small resistor, and use a scope to look at the current drawn by the rotating fan - you should see a significant AC component caused by the comutating of the coils. This AC component disapears when the fan is stalled, although you'd need to check what it does if/when it tries to restart.

Neat idea. I'll have to fire up the scope when I get back and have a look.

...meanwhile I'll try to find my old notes from the time

Might be useful, cheers.