What methods are there of reducing volume between stages?

[Derringer]

I was messing with a design that I liked the tone of but the volume was just way to great. And maybe the tone that I thought that I liked had to do with the preamp of my amplifier being overdriven too .... but this brought up a question.

Aside from using a voltage divider between stages, what other ways are there to reduce volume ... especially without sacrificing tone?


Thanks

[John Lyons]

Pretty much either a voltage divider or use lower gain devices per stage.
Also depends on what kind of gain devices you are using. FETs, op amps,
transistors etc.

John

[Derringer]

is there any such thing as a negative gain stage?

or perhaps an amplitude reducing buffer?

[jacobyjd]

is there any such thing as a negative gain stage?

or perhaps an amplitude reducing buffer?

You could check out something with an emitter-based volume control, like the Beginner Project--just 'turn' it down with a fixed resistor set in place of the pot.

[earthtonesaudio]

Sure.  "Gain of less than one" is what you'd call it.  When you say "negative gain" it implies inverting the signal, rather than saying something about the amount of gain.

Take your basic op-amp inverting amplifier configuration, make the input resistor larger than the feedback resistor.  Voila, an active voltage divider.

Though I'd recommend reducing gain before attenuating, because that will reduce your noise levels as well.

[R.G.]

Aside from using a voltage divider between stages, what other ways are there to reduce volume ... especially without sacrificing tone?

What's not to like about voltage dividers?

Done properly, with parallel capacitances to swamp out parasitic caps, these can be made constant attenuation with frequency from DC up to well into the AM radio band, if not further. No tone loss there.

[Derringer]

is there any math to finding the "proper" parallel capacitance or just ears?

[R.G.]

is there any math to finding the "proper" parallel capacitance or just ears?

In this universe, there is math to everything.

Ever wonder how oscilloscopes did that signal division at the front end that lets them have flat response to tens of megahertz?

Resistors are naturally frequency-flat to the extent that they're resistors. But because there are metal leads at both ends of a resistor, there is some capacitance between those leads if nothing else. This few picofarads doesn't matter much if it's a 100 ohm resistor, but things get dicey if it's a 1M. What we want is something which makes the stray capacitance not matter. And that is - more capacitance. 3pF of stray capacitance doesn't matter much if you parallel it with 100pF. So you put capacitors in parallel with the resistors, and make the capacitive impedance be in the same ratio as the resistor impedance. That means that the total impedance of the resistor+capacitor divider goes down as frequency goes up, but the ratio stays the same.

For instance, if you want to do a 100:1 divider and have it be frequency independent (mostly), you might use a 100K resistor and a 1K resistor. Since these will physically be similar, they'll both have about the same parasitic capacitance; call it 2pf just to pick a number. We want the 100K resistor's stray capacitance not to matter, so we parallel it with a capacitance at least ten times the stray capacitance, or 22pF just to pick a standard value. But for the 1K, we have to pick a capacitance which has 1/100th the impedance at any frequency. Easy enough, use 2200pF to parallel the 1K.

Now the resistors dominate the divider up to F = 1/(2*pi*100K*22pF) = 72.4khz. Note that the time constant of the 1k/2200pF is the same, so the divider ratio is the same. And above 72.4kHz, increasingly the resistors don't matter and the caps do. But even at 10x or 100x the frequency, the caps keep the same 100:1 divider frequency. They only stray away from that when the frequency gets high enough that series inductance in the cap leads makes the ratio fail.

If you don't need a divider at DC => you don't even need the resistors <=; of course, the impedance of the divider gets big at low frequencies now, so maybe the resistors are a good idea to keep things in hand.

[Paul Marossy]

Though I'd recommend reducing gain before attenuating, because that will reduce your noise levels as well.

For the best results, I think it's a balance between the two. Especially in the earlier stages, because noise there will be further amplified by everything that follows.

[Derringer]

R.G. ... as always, thank you so much !

You are a true teacher!

[Mark Hammer]

Op-amp gains of less than 1 can be used in clever ways.

When you have a chance, take a look at the schematic for the PAiA Rocktave octave divider ( http://hammer.ampage.org/files/rocktave.pdf ).  The "bypass" switch consists of simply closing a switch in an op-amp feedback loop that replaces the 680k feedback resistor with a piece of straight wire.  Normally, with a 47k input resistor, and 680k feedback resistor, that stage would have a gain of 14.5.  When the 680k drops to zero ohms, the gain becomes zero and no signal continues onward.