I was thinking of using a TDA2030 audio power opamp to make a simple variable power supply, by arranging it for unity gain and just feeding a DC reference voltage to the non-inverting input.

However, looking at the data sheet,
http://www.farnell.com/datasheets/7918.pdf
it says on page 7 that the gain must be set higher than 24dB. Does anyone know why that is, and whether it applies to my application?

 

Why do you want to do that Even if your idea is not stupid at all. I think you you may be better of with voltage regulator. What kind of current and voltage output do you need?

 

It won't be stable if you try to use it at <=24dB gain. You won't be happy with the results.

Why don't you describe your requirements for your power supply?

 

Look closely at the schematic and realize that in the typical application the amplifier is AC coupled. It is not designed to be used in a direct coupled unity gain DC environment. I'm not sure I know exactly how it would misbehave without being AC coupled, but I think you need a different approach for a variable DC power supply.

My guess is that the open loop gain is so high (90 dB) that even very small offset voltages would cause the amplifier to go non-linear, saturate, and maybe let the magic smoke out.

The other clue is that the amplifier operates Class AB. Here is a quote from Wikipedia that summarizes this mode of operation:

http://en.wikipedia.org/wiki/Electronic_amplifier

"Here the two active elements conduct more than half of the time as a means to reduce the cross-over distortions of Class B amplifiers. In the example of the complementary emitter followers a bias network allows for more or less quiescent current thus providing an operating point somewhere between Class A and Class B."

At low frequencies, like near DC, having both active elements in the output stage on more than half the time has a tendency to let the magic smoke out. It is essentially a short across the power supplies.

 

I don't have a specific power supply requirement in mind, I just found a TDA2030 and thought I'd mess with it (unconventionally)!

According to the datasheet, with 8V across either output transistor it can deliver 3.5A peak, or I guess 2.5A continuous since it will shut down at 20W dissipation. I thought it might make an interesting regulator that could be varied continuously from positive to negative output.

I was just interested in why there was a minimum limit to the gain, unlike with most opamps. The PSRR graph on the datasheet suggested an outstanding figure at unity gain, way better than an LM317! But I guess I can just set it up for 25dB gain, I don't think it will mind operating at DC rather than with audio!

 

Funny but for about 10 years ago i did the same with a TDA2003. It was just some playing around on my work bench.
If you want to play with voltage regulator. A LM723 is an excellent chip to play with. With external power transistors you can have high current output.

 

The audio amplifier will oscillate at a high frequency (hundreds of kHz) if its gain is less than 24dB because its frequency compensation capacitor is a low value for high open loop gain at high audio frequencies to cancel distortion.

Class-AB amplifiers work normally and perfectly with DC output voltages since the output is always a steady or varying DC voltage.

 

Perfect, that's what I was after. Thanks AudioGuru!

 

I usta work designing with 723 chips. It's amazing what you can get one little chip to do. Variable voltage, constant voltage, constant current, foldback protection, huge amps (using external transistors), high voltages, and the feedback section is another place for "How did you do that?!" stuff. I once got a 28 volt supply to deliver 3 amps with .008 volts of sag error!

Have fun with your designs. There is plenty of fun to be had.

 

I usta work designing with 723 chips. It's amazing what you can get one little chip to do. Variable voltage, constant voltage, constant current, foldback protection, huge amps (using external transistors), high voltages, and the feedback section is another place for "How did you do that?!" stuff. I once got a 28 volt supply to deliver 3 amps with .008 volts of sag error!

Have fun with your designs. There is plenty of fun to be had.

Man I'd love to see some of your designs!

Feedback section?

 

I was thinking of using a TDA2030 audio power opamp to make a simple variable power supply, by arranging it for unity gain and just feeding a DC reference voltage to the non-inverting input.

However, looking at the data sheet,
http://www.farnell.com/datasheets/7918.pdf
it says on page 7 that the gain must be set higher than 24dB. Does anyone know why that is, and whether it applies to my application?

I was at farichild Semi back in 1978 which is when we developed the TDA2030 there. It was a second source of an SGS product.

It is an audio amp and is not stable at lower gains, that is to give a wider bandwidth and lower harmonic distortion.

You can use the part DC coupled at the output as long as you run it off dual supplies and use some kind of "input trimmer" to zero the output with no signal input.

I used a pair of these as bridge amps once upon a time and DC coupled to the speaker, but you do have to trim the output so there is no DC voltage across the speaker with no input signal.