Nifty offline converter

Thread Starter


Joined Nov 19, 2010

I had a fun experience recently after my Sunbeam heating pad suddenly continued to show an F on the LCD screen and did not heat. Figuring that it was unlikely that I could fix the unit, I took apart the controller to see what was what.

It appeared that the brain is a uC or a specialized control chip. (28 pin) What interested me was how the logic supply was derived from the 120V AC mains.

Tracing the circuit with a meter produced the circuit below drawn in LTSPICE. The circuit is a basic capacitive divider to drop the line voltage which then uses a zener diode and with a series diode to clean up the ripple using the 330uF cap.

What's not shown in the schematic is the series fuse with the input voltage. The varistor actually used in the circuit is a TVR07391 which I could not find a model for.

There is additional protection from overheating. I show R1 as a 100Ω resistor but in the controller, this resistor is actually two. A 1/2W 47Ω resistor in series with a 47Ω thermistor which I suspect is a PTC thermistor. (resistance increases if C1 or C10 get hot)

Am always nervous about line derived voltages, but seems like a clean design and is used in a commercial product.

My questions to the Forum are:

• Can anyone provide an LTSPICE model of the TVR07391 or an appropriate Littlefuse model?

• Is the series thermistor needed?

• Both C1 and C10 are quite large. (through-hole)If I were use this circuit, how to shrink the cap sizes for C1 and C10?

Screen caps show the circuit response while drawing 10 mA. Attached are the LTSPICE files including the Littlefuse varistor.

Any thoughts, comments, or help are greatly appreciated!






Joined Aug 7, 2020
You are very unlikely to find smaller versions of class-X capacitors, however, C10 is superfluous - you could remove it altogether.

Thread Starter


Joined Nov 19, 2010
Thanks for your response, Ian0. I just ran the sim with C10 at 1pF. You are right, output is the same as with C10 at .1uF in the circuit. It seemed that C1 and C10 were forming a capacitive voltage divider but that does not appear to be the case. Obviously, Sunbeam wouldn't place a component that wasn't necessary. Somehow for safety? They must make 100's of thousands or more per year- they wouldn't add a component for no reason. Agency approval: ETL, UL, CSU?

Seems odd.