Quote:
Originally Posted by Karen O
I've seen common audio amp ICs (e.g. LM386) used as supply splitters. They bias their output to about half supply...
|
Yes, that's all you need. It tends to be cheaper than buying a dedicated IC, especially as an audio company is already buying audio op-amps by the bucketload, so will be getting excellent prices... I'd prefer an op-amp to the LM386 because you're relying on the internal biasing of that to keep it at the mid-point rather than external circuitry - but obviously the LM386 can supply a bit more current.
It's literally 2 equal value resistors followed by a unity gain follower. Just about any op-amp will do, but check the current rating of the output stage. Having said that, there really won't be much in the way of DC currents in the earth circuit, so it's not important. Many years ago, Quad used just 2 resistors for some of their power amplifiers!
When doing this, remember that DC precision is not a big deal - audio circuits don't mind if the rails aren't perfectly symmetrical, especially if you are nowhere near the clipping point.
Some years back I built a super-simple version of this using just 4 or 5 transistors - details are at work, but I can look them up next time I go in to the office. An op-amp circuit would obviously work better, especially in terms of DC precision, but it was an idea I wanted to experiment with one lunchtime. Plus I wanted the circuitry to take almost no operating current, so the reading on the bench power supply was still fairly accurate, and we didn't have any low-current op-amps in stock at the time...
I mounted it in a small diecast box, with leads from the single bench PSU coming in one end, and three 4mm binding posts on the top. The box obviously acting as a heat sink, should there be significant earth currents. I arranged the circuit so that the two collectors of the power transistors (TIP31/32 IIRC) were connected to 0V, so no need to insulate them from the box. Nice and simple.
At the other end of the scale, Calrec did this with an op-amp and some beefy Darlington transistors - this circuit took 36V DC from a separate regulator and turns it into +/-18V. This would be a good one to copy if you wanted better DC precision.
Also, once the experimenting is done, and you're ready to implement a final circuit, it's usually easier to just modify your circuit if you wanted single-rail operation - you almost certainly won't need extra active devices to do that.
Cheers,
Mark