555 based power supply getting too hot

[madan1]

Good day.
I have an AC/DC -> DC PWM (rectifier + 555 buck converter) power supply with MDF11n60, which is getting very hot.
The input is AC from a transformer and the output powers a 24v dc fan. The power supply is used to regulate the speed of the fan.
After several runs on the bench power supply I noticed that the mosfet is reaching 80+C*. The mosfet reads fine on component tester, the thermal paste is changed, the fan is cleaned and with freshly oiled bearings.
When powered directly from the bench psu, the fan takes ~0.7A on 12V. The current consumption is similar (~ 0.6-0.7A on MAX and ~0.3-0.4A on MIN) when the power goes through the power supply/regulator.
The board already has thermal damage and I'm not sure if this is the original mosfet.
Do you think there could be something wrong with the supply (bad IC, wrong mosfet, etc. ) or the cooling is inadequate? The fan is brushless and its bearings spin freely.

P.S. On min, it gets hotter than on max.

[PeteS in CA]

1. DC fan current is not necessarily a smooth DC current. I recommend looking at the current using an oscilloscope and a DC current probe. This is not a suggestion for a change, but it will give you a better understanding of the actual load current.

2. Your MOSFET is a 600V, 11A part. Because of the high maximum voltage, the R(DS On) is pretty high, .55 ohm on the datasheet. Since the input electrolytic capacitor is rated for 50V, I suggest trying a 100V part.

3. The MOSFET you used is in an isolated ("FullPAK") TO-220 package. Unless you actually need the heatsink to be isolated from the Drain, I suggest using a regular TO-220 part. The junction-to-case thermal resistance of a FullPAK is much higher than for a regular TO-220 package.

4. With thermal grease, more is not better. If you have more than a thin film of it only on the back of the device OR on the heatsink, it probably would increase the thermal resistance between the device and the heatsink.

Judging by the browning of that 3-pin connector, I think changing your MOSFET to a lower voltage part (with similar maximum current) would make the most improvement. Changing to a non-isolated package and proper grease application would just improve heat transfer from the MOSFET die to the heatsink. A lower R(DS On) MOSFET should produce less heat.

[CapLeaker]

Nah, that 11N60 isn’t original. It is a high voltage switching mosfet you find in a SMPS. What was in there is a low voltage switching high current. RDSon is like worlds apart.

100V N channel MOSFETs aren’t always that easy to find, but look in a range of 75V and 100V Pick one with low RDSon. The original MOSFET probably didn’t have that body diode in it either, like the 11n60 does.

[madan1]

Thank you for the replies!
I will start looking for a more suitable mosfet. Any ideas about easily obtainable ones?

[eccerr0r]

Looking at the design of this board, no, you cannot use bigger mosfets to use bigger fans. How big of a fan are you trying and what was the mosfet before the repair?

Staring at the design of the board I doubt there was anything bigger than perhaps an IRF510 MOSFET or something prior to the rework. Using the MDF11N60 will slow the circuit down so much that it will fry itself.

Also want to note that this circuit is shitty... at 24V input voltage (unregulated, it would be higher...) this is also running up to the limit of even high voltage mosfets...in fact IRF510 won't work here due to this problem. What was the designer thinking??? I don't think this board was designed to run at 24V at all... probably no more than 16V or so.

[madan1]

Looking at the design of this board, no, you cannot use bigger mosfets to use bigger fans. How big of a fan are you trying and what was the mosfet before the repair?

Staring at the design of the board I doubt there was anything bigger than perhaps an IRF510 MOSFET or something prior to the rework. Using the MDF11N60 will slow the circuit down so much that it will fry itself.

Also want to note that this circuit is shitty... at 24V input voltage (unregulated, it would be higher...) this is also running up to the limit of even high voltage mosfets...in fact IRF510 won't work here due to this problem. What was the designer thinking??? I don't think this board was designed to run at 24V at all... probably no more than 16V or so.

This is how I got the unit - board with burnt marks, lifted traces under the mosfet and dried thermal paste, the fan is the original one. The only thing I did was to test the mosfet and fix as much as possible its traces. I have no idea what has been the original mosfet.
By the way, still haven't tested the board on its original transformer, and I'm not really sure what's the original ACV. The 24V is guess based on the fan label.
The unit is a smoke-fog machine and the fan is used to dilute and blow away the smoke.

[stj]

do you have a schematic?
you need to make sure the fet is fully on and off.
if it's not then it will get hot.

[eccerr0r]

As a workaround, what's the capacitor between pins 1 and 2 of the 555?

If it's less than 100nF I'd suggest increasing it to something between 100nF and 1µF. If it's currently like 10nF or something, it would probably explain why it's overheating... that is if you're not frying the mosfet from using a voltage too high -- of which an IRF510 probably would with fry with a 24V transformer (though may work fine with a small capacitor). The MDF11N60 may work better if the capacitor is bigger.

Might be able to work it with a circuit topology change but how the pot is connected to the board would be interesting, though I think I have an idea...

BTW do measure the voltage of the transformer, need to know that constraint.

[madan1]

do you have a schematic?
you need to make sure the fet is fully on and off.
if it's not then it will get hot.

Oh, come on, everyone can repair with schematics and easily obtainable parts .
I will play around with the scope and see what is going on on the mosfet.

P.S. The exact unit model is QTX FH-1500.. I couldn't find any schematics.
The fan is PAPST 959 5414 200, it has 4 wires, but only the power ones are used. Couldn't find datasheet about it as well. I guess the extra wires are tacho and probably some kind of control signal?

[madan1]

As a workaround, what's the capacitor between pins 1 and 2 of the 555?

If it's less than 100nF I'd suggest increasing it to something between 100nF and 1µF. If it's currently like 10nF or something, it would probably explain why it's overheating... that is if you're not frying the mosfet from using a voltage too high -- of which an IRF510 probably would with fry with a 24V transformer (though may work fine with a small capacitor). The MDF11N60 may work better if the capacitor is bigger.

Might be able to work it with a circuit topology change but how the pot is connected to the board would be interesting, though I think I have an idea...

BTW do measure the voltage of the transformer, need to know that constraint.

The cap between pins 1 and 2 should be 0.1uF (104). Haven't actually measured it.

With the 555 supply connected, the transformer outputs between 21Vac on max fan speed and 23Vac on min fan speed.

I'm attaching a photo of the pot pinout and connector (the red wire goes in the middle).

[eccerr0r]

You might want to increase the cap anyway (or put some more in parallel) . Next thing is reducing that resistor near that cap but I can't read it. It's probably is risky too. TBH using the built-in PWM of the fan is the better solution if possible.

[stj]

you could give the fan full voltage and use the 555 maybe to drive the pwm pin
if the 555 is configured for fixed frequency with variable pulse width.

[CapLeaker]

I don’t get it. All that jazz just for a fan?

[madan1]

Unfortunately I can not find any documentation about the fan. The tacho-control wires is a guess based on other papst fans, but again - no real documentation about the pinout or how to connect and use the extra wires.
Just checked aliexpr. and found this https://www.aliexpress.com/item/33023396847.html .. I guess at the price of a decent mosfet, it makes all further interventions on the existing power supply pointless.

[CapLeaker]

If it doesn’t need to be variable fan speed, this circuit can be much simpler yet and you don’t need the PWM pin either.