TL494 SMPS under developement

Thread Starter


Joined Apr 28, 2012
I am currently testing a TL494 based SMPS, loosely based on the attached schematics, but it is different in many details.

The transistor drive is the same however.

Cooling is done using a large VGA cooler! And the coils need cooling as well.

I have done tests at 12V, 4.8A, and will do more tests soon with 50W bulbs.

The TL494 has 12V external supply (electronic wall adapter).

There are various problems however, the VGA cooler needs mechanical padding since the power PNP is much thicker than VGA CPUs, so it is unjust. 2x TO220 components (burned) do the trick right now.

I can't really get efficiency more than 80% to some degree also my testing equipment is not that precise (DMM).

And current regulation does not have any effect whatshowever, maybe I connected it wrong. Having a PDF here from TI with explanation, so I will try to make this working.

Eventually I want to add digital voltage adjustment, as well displays for input voltage/current + output, temp. and efficiency.

If I will produce a PCB, eventually there will be economy option using smaller cheaper cooler, and only one coil. 5 Amps is a lot of current!

Using a 2 kW welding transformer for the input voltage right now!

I had a big MOSFET built into the circuit but destroyed it when I tried gate driving via transformer which wasn't working at all. MOSFET seem to be more sensible! For instance IRF9540 will latch up instantly, when 2x 50W bulbs are connected cold.

The TIP36C seems to be more robust. With 4.8A current output at 12V, two coils, there is no significant heat developement.

It's some effort to remove the power transistor/MOSFET, but I want to test again with big power MOSFET, having suitable coils in place now.

The coil system I have in place now has it's best efficiency at lower frequencies, audible and slightly above, if I increase it too much, current goes up a little and at some point, doubles etc.

700uH in original schematic, and 350uH now (2x coils in parallel), which slightly increases efficiency and helps to keep these coils cool.

One larger coil (also in the photo) with thick magnet wire is not matching, results in large currents, so maybe magnetic material is wrong.


Thread Starter


Joined Apr 28, 2012
Well I received the coils a few days ago, and I was lucky, ordered the right ones. Even if the wire is a bit thin, so one coil is good for max. 6A even with airflow.

I don't really have a schematic at all, and only some vague ideas how to do digital voltage adjustment. Current regulation I want to get working soon. And to buy more 12V bulbs...

Moved the diode under the VGA cooler as well, visible in the photo:

the TIP36C
the power diode
7805 used for LED display (from 12V AUX).

The padding is required since the VGA coolers have not been designed for this application.