Input Overvoltage In Discrete Amplifiers

[N9]

The answer to this question may be blindingly obvious, but here goes:

I'm familiar with the input voltage limits for op-amps and have seen several protection strategies for them, the most common being reverse-biased diodes to VCC and GND.
However, I've realized that these protection strategies never seem to be employed on discrete amplifiers, at least not in the effects world. Is this type of protection not needed in typical discrete designs, or is it overlooked?

To give more context to this question, here's what got me thinking about it: more and more effects these days are using charge pumps and can, consequently, have large output signal swings. If someone runs a +-15V signal from their fancy overdrive into their lowly big muff running on an old 9V battery, can we expect the big muff to have a bad day?

[R.G.]

I'm familiar with the input voltage limits for op-amps and have seen several protection strategies for them, the most common being reverse-biased diodes to VCC and GND.
However, I've realized that these protection strategies never seem to be employed on discrete amplifiers, at least not in the effects world. Is this type of protection not needed in typical discrete designs, or is it overlooked?

It's overlooked, and often would be considered as sullying a "vintage" design if it was thought of. Which is a little sad, but there you are. Sometime discrete designs are more robust simply because the devices are larger and able to take more abuse without completely dying. One ramification of this is that old effects are often noisy. I personally believe that this comes from a long life of having the input base-emitters zenered in a non-immediately-fatal manner. This is a proven strategy for making transistors noisy.

To give more context to this question, here's what got me thinking about it: more and more effects these days are using charge pumps and can, consequently, have large output signal swings. If someone runs a +-15V signal from their fancy overdrive into their lowly big muff running on an old 9V battery, can we expect the big muff to have a bad day?

Probably no worse than normal. The fancy overdrive output probably has an opamp which has an internal limit on how much current it can put out, and it takes the product of the breakover voltage and current to overheat the input devices. Many older pedals also have a series resistor at the input, which further limits current and damage. However, it might be interesting to connect the input of an effect to the output of a good-quality hifi amplifier capable of, say, 400W (and hence a lot of current) and to then apply that same +/-15V to a pedal. Things might smoke. Or not. Depends on the effect being abus... er, tested.

[PRR]

Argh, R.G. types fast.

> runs a +-15V signal ...into their lowly big muff..., can we expect the big muff to have a bad day?

Look at a plan.

http://www.bigmuffpage.com/Big_Muff_Pi_versions_schematics_part1.html

Take the top one, "V1 66#5". All input must flow through R2 36K. Assume full 15V through 36K, we get less than 0.5mA. What is the blow-up Base current for a small transistor? Actually not specified for 2N5133 or for 2N5089; but the 2N5089 has a Vce(sat) test with Ib of 1.0mA, so it can surely pass that much. 2N4401 is about a 6X bigger part and has a test at Ib=50mA, however this is 2% pulses not steady.

Most transistors are at least as robust as a small-signal diode, and 10mA is not a big deal.

It is possible to pop an input transistor, but it hasn't been common.

That's when you pull the input up and make the transistor conduct like crazy. When you pull-down, at first the transistor Base junction blocks, then at -7V it breaks-down like a Zener. The current is still limited by the 36K, and if total power is not excessive there is no large damage to the transistor. *However* each such breakdown changes the junction properties in a small way. And one of thise ways is HISS. Greybeards have seen long-abused amplifiers with hissy inputs which clear-up with a new input transistor. We can't prove this was abuse rather than old-age or dirty Silicon, but it may be something to avoid.

[bool]

Argh, R.G. types fast.

... HISS. Greybeards have seen long-abused amplifiers with hissy inputs which clear-up with a new input transistor. We can't prove this was abuse rather than old-age or dirty Silicon, but it may be something to avoid.
...

Quick and dirty solution is to strap a SMT diode across the B-E of the input transistor (BJT). For NPN, K goes to B and A goes to E. For PNP, vice-versa.

[N9]

Thanks for your thorough replies! This has definitely given me some food for thought.

Quick and dirty solution is to strap a SMT diode across the B-E of the input transistor (BJT). For NPN, K goes to B and A goes to E. For PNP, vice-versa.

Is there a particular reason the diode should be surface mount?

[amptramp]

Thanks for your thorough replies! This has definitely given me some food for thought.

Quick and dirty solution is to strap a SMT diode across the B-E of the input transistor (BJT). For NPN, K goes to B and A goes to E. For PNP, vice-versa.

Is there a particular reason the diode should be surface mount?

A surface mount diode would probably fit without needing any new holes or wires.