Temp-controlled SMPS fan supply (TCSFS?)


I find that some of my projects end up needing a fan.

My last amplifiers have run from +/-60V and the fans mostly run of 12V.
To keep them quiet I like run them around 10V, but in any case they cannot run from 60V.

I would like a temperature-controlled switch and I would like to avoid the massive loss from a linear regulator from 60V.

So we have two challenges: a buck-regulator that will drop 50V with low loss and a temp-controller.

The simplest fan-controller I can find on the www is from Mark Hennesy:
Micro-Amp › Circuit Design

I know that integrated SMPS-controllers exist that will do the job, but I did not have any and I wanted a solution that is widely availible.
A search revealed that a very simple circuit was maybe enough (Google: "dsicrete buck converter".
circuit analysis - DC-DC Buck Converter Issue - Electrical Engineering Stack Exchange

So the building-blocks look simple, - i "just" had to connect the two...
That proved a little complicated, and also I wanted to use a N-channel-FET as those are cheaper and already in my stash. luckily a discrete solution is easily mirrored to use N- instead of P-channels.

After a few iterations I ended up with the attached schematic.

A few notes:
  • If the circuit do not oscillate it can in some cases regulate linearily. It may produce an output voltage, but the dissapation is much to high
  • The feedback-capacitors C2 and C7 may need some trimming to optimize performance and achieve stable regulation
  • Two feedback-options are availible: throug C2 or C4. I still dont know whats best.
  • I have used a coil from the junk-box around 200uH. The value do not appear to be critical.
  • losses in the inductor may be enough to heat it up. Try different alternatives.
  • Be aveare that T2 and Q3 see the full input voltage. use 100V devices for 60V supply
  • Notice that the fan-output is NOT regulated relative to GND, but to Vcc. That should not be a problem for an insulated plastic-fan.
  • D1, D2 and D6 are 10V zeners
  • D3 is a low-voltage fast diode
  • D4 is a 100V fast diode
  • R5 sets temperature-trig point. May be replaced with R9 when the value is found.

Prototype draw around 40mA from 60V when supplying a fan around [email protected] Mosfet and coil is warm, but far from dangerous.

I'll share an single-sided PCB if there is any interest (and the prototype behaves)

What do you think?

Kind regards TroelsM


  • Schematic.pdf
    43.2 KB · Views: 42
Working prototype.


  • _20190928_225900.JPG
    539 KB · Views: 71
  • DSC_1960.jpg
    976.8 KB · Views: 70