LM317 - voltage limits

Question

I have a couple of active PA monitors with only unbalanced inputs. I want to be able to use the balanced monitor outputs from my mixer, so I am planning to build in a balanced to unbalanced converter in each monitor. I have build the boards using NE5532s. These works fine, then my next step should be to make a power supply taking power from the power lines for the amplifier and regulating it to approx +/- 15V using LM317 / LM337.

When I measured the power supply of the amplifier, it turned out to run on +/- 58V. As the NE5532s are rated to run at up to +/-22 V (recommended up to +/-15) and the LM317 / LM 337 can take a voltage drop of 40V, this should just work, but I do not like it - it is a bit too close to the limits. As I measured the voltage of the amplifier across the filter caps without any signal, the voltage will probably be slightly lower under normal use, but anyhow...

At the moment think I have these possibilities:

• Run it as planned, since the circuit is not pulling much current (< 100mA), it will probably work fine.
• Try to find other power lines in the amplifier with lower voltage. This may be more difficult and a bigger risk of somehow creating problems for the amplifier.
• Running two sets of LM317 / 337s in tandem, i.e. using the first stage to get from 58 to 30V then the next from 30 to 15. - I have never seen this been done, a slight waste of components, but I cannot see any reason why it should not work.
• Use some other voltage regulators, integrated or discrete, that can take a higher voltage drop. The only reason to use the LM317/LM337 is that I have a few of them in storage. But are there any recommended voltage regulators that can take a voltage drop in the order of 40V? The monitors are old Laneys, so I do not think I want to put a switched voltage regulator too close to the active circuits.

What would you recommend? Maybe I even have overlooked some possibilities...

Answer 1

This looks good on paper until "something happens". Example schematic from the net with random resistor values:

If this is connected to +58V then LM317 will only survive if it never experiences excessive voltage, and that can happen in several ways:

• Input voltage is connected while output caps are discharged (this can happen if there is a loose or intermittent connection).

• Short on the output

• Output cap too large so LM317 can't ramp up output voltage fast enough on power-up

• Etc...

If you put two LM317 in series I wonder what would happen in case of a short circuit at the output. The LM317 which goes into current limit mode first would probably exceed its maximum voltage...

Personally I would rather put a Zener of adequate power in series on the input to drop enough voltage to make it safe for LM317. If the zener gets too hot you can beef it up with a transistor, or use several zeners in series to spread the heat.