Methods for limiting input

[edvard]

What's the best way to protect a SS input stage (non-inverting op-amp set for about 5x gain, in my case) from overdriving by overwound or active pickups, or a boost pedal? 
Things I've thought of:
- Make the integral gain adjustable via the (-)input leg to ground as a "drive" or "trim" control, and attach a clipping detector circuit with an LED indicator.  I figured out pretty quickly that idea is probably a bit too complicated. 
- Run the op-amp at a higher voltage for maximum headroom.  ±15v isn't that much higher headroom, and if I ever experiment with active pickups or overdrive pedals, eek! (I went with ±15v because rated max voltage for most common op-amps is ±18v, but IMO, that's pushing it kinda hard)
- Replace the op-amp with a MOSFET or JFET buffer/booster, so when it clips, it's not harsh.  However, simulations in LTSpice show that contrary to popular belief, MOSFETs and JFETs do indeed clip hard.
- Wire up a bare-bones limiter circuit so the signal never goes over a certain level.
- Instead of an op-amp chip at the front end, build a discrete unit, so I can run it at even higher voltage, because eventually I'll be putting a power amp on the back end, and it'll be running on ±24 or 35 volts.

Or do you simply design it for maximum headroom and let it ride?

[J M Fahey]

In principle, lower first stage gain, there is only so much swing you can get out of a piss poor 9V supply.

For high gain humbuckers or active pickups, often no more than 2X, go figure.

A regular Op Amp, say TL072, can only swing within 2V of 9V or ground so 5Vpp tops.

Rail to Rail Op Amps such as LM833 can swing the full 9V pp

IF you increase supply to 18V or higher, up to +30V single supply, headroom improves dramatically.

[edvard]

The pre-amp I'm designing is running off of ±12V, and boosts over 5x already start clipping transients, and that's from some overwound mini-humbuckers.  I hesitate to think what full 'buckers or active pickups like the EMGs that can output up to 3V will do. The goal is clean boost at the front end, so all the overdrive comes from everything after that.

So, basically, just design for the most headroom and hope for the best?  There aren't too many common op-amps designed for high voltage, unless I want to use a poweramp like the LM1875 as a common op-amp; that one goes up to 60V (I assume that means ±30v), or new devices like the MAX74810 that can handle 50V.  The cheapest, highest voltage through-hole DIP I could find was the LF356 that can handle 44V, but you have to get the right one.  All others in the LF35x line are 36V, which is still respectable, but that's what most common op-amps will handle, so why look for specialty devices?

I was thinking also some simple soft-limiting up front might be ideal, which would leave the potential for some clipping at input, but then again, there's not many tube amps that are completely clean, and overdriving them is kinda the point, so maybe that would be OK.  Back to the drawing board... 

[g1]

The most standard, tried and true method is a pad/attenuator at the input; either using 2 input jacks or a switch.  Usually labelled Lo and High, or passive/active.
It is still the most common method you will find in use today.
Is there a particular reason you do not want to use this method?  Perhaps trying to address any shortcomings would be easier than trying to 're-invent the wheel' so to speak.