Quick questions about preamp for acoustic guitar

[David]

This (these) should be easy for someone who's done it before.

I have acoustic 6- and 12-strings.  Neither have pickups.  I got a Seymour Duncan "Woody" that I attach to whatever guitar I need to use.  Woody's output is weak, to say the least.  I had been using it with an old Barcus-Berry preamp -- which failed me last night and never provided much gain when it DID work.  OK, so I need a new preamp.  I can build one in my sleep.  I plan to add a simple tone control to the output.

OK, here's what I need to know:

Is it better to make the preamp gain fixed and put a volume pot after the tone control, or should I build a volume pot into the feedback / ratio resistors on the op-amp?  

Which arrangement, if any, plays nicer with a tone control?

Is gain of 20 enough for a weak pickup, or should I go for more?  How much before distortion starts?

Is this a case where a 5534 or TL071 is clearly better, or would either one work just as well?

Quick responses are greatly appreciated.  I need this thing on Sunday!

[Mike Burgundy]

Most of these questions/answers depend on what you want exactly, and what circuit you'll use.
The upshot of variable gain is less noise - this is why a lot of engineers set the faders (a volume pot, only subtracting) to 0dB (which is more or less full-on, don't get me wrong) and *mix* with the gain stages. If you use set gain but mix with the faders, in some conditions more noise will occur.
That said, I don't think it matters much (and I prefer mixing with faders), especially not here.
A pre with volume pots is that you can actually go to full-off (no sound, can be useful) which is something a gain knob won't do (no feedback resistance means unity gain).
Tone controls work best when fed with low-impedances. Doesn't really matter wether that's an opamp out, source follower or whatever.
Gain of 20 might be enough, migh not. My guess is it IS. Then again, if you work with IC's, it's going to be a breeze to mod. Sockets might be handy, esp. if you want to tailor tonal response as well.
There is no way to predict where distortion will occur in your setup with this info. Just try it.
5534/071? Well, you'll want a high input impedance - so an IC with a FET input is desirable. Other than that, a 5534 is a little less noisy. Not something you'll hear, but if you want to make sure go for the 5534.
you could also consider a discrete buildup, based on something like this:
http://www.till.com/articles/GuitarPreamp/
hih

[David]

Mike:

Thanks for getting back to me.

Let me just confirm what you indicated:

Gain of 20 is probably enough.

Variable gain in the feedback loop is more useful.

A 5534 would be less noisy.

Great!  I was hoping someone would give me some conclusive direction!

As for the impedance thing, I'm going to do that by sending the input signal to the noninverting input on the op-amp and setting the input impedance to 1M by using a voltage divider composed of 2 1M resistors in front of the noninverting input.

[spongebob]

I would favour the fixed-gain approach with the volume control on the input. From my experience, having a long feedback signal path (wires to pot and back) can introduce all kinds of problems (noise pickup, oscillation), especially with fast high bandwidth opamps. I never had stability problems with an OA that had the gain set by a fixed resistor. Just put the resistor in a socket and you can replace it easily if gain is too low.

As for the impedance thing, I'm going to do that by sending the input signal to the noninverting input on the op-amp and setting the input impedance to 1M by using a voltage divider composed of 2 1M resistors in front of the noninverting input.

For a better method look at figure 2 in this pdf:
Avoiding Op Amp Instability Problems In Single-Supply Applications

[Mark Hammer]

Take a look at the article on op-amps in one of the first couple of issues of DEVICE I have posted at my site (http://hammer.ampage.org ... scroll down a few pages).  Craig Anderton provides a nice coverage of many of the op-amps available in 1979.  Of course many more are available now, but the basic principles remain the same: noise levels will vary depending on the resistance values involved in the circuit, and are not based solely on the chip independent of the circuit.  In some instances a device like an NE5532/34 will provide superior noise levels, in other cases a different device like a TL071 or functional equivalent will.  For piezo pickups that lack any sort of active on-board electronics, however, the first device MUST be a high input impedance type.  Whatever comes *after* that can be other kinds, just as long as you design around what optimizes their noise levels.