Rules of thumb for gain of each stage in multi-stage amps ?

[Vivek]

Are there any rules of thumb that help one roughly decide on the amount of "gain" and amount of clipping of each stage of a multi-stage amp or pedal ?

[ElectricDruid]

No.

These are the things that you choose when you design a overdrive/distortion pedal, and they're some of the things that give one pedal a slightly different flavour from another, in as much as they are different at all.

[antonis]

No.

+1..

P.S.
Of course, you have to calculate each stage IN-OUT impedance, take into account "loading" and choose overall tone responce..

[GibsonGM]

^^  This.   There IS a logic in these things, but no hard rules.   

You can have an idea of what your input signal is, maybe 1V p-p.   You can set your amount of gain accordingly, can take into account various input levels and what they'll do (humbucker vs. single coil etc).  You know your power supply voltage, and can know the limits of the device to get near them.    You can set if you want a stage to clip or not based on how you set the gain, whether you come near the rails or not; and with how you bias it (symmetrical vs. asymmetric).   You can learn what things sound like, opamp vs. BJT vs. FET clipping - some are 'nicer' than others.   Some are GREAT for fuzz, not so great for playing dynamics.

You can consider if we're going to amplify cleanly and then use clipping diodes? How hard do we want to hit them (experiment with all things...).

We want a healthy level of amplification for the first stage, to get above the noise floor.  But we may not want to add 'grit', so we stay inside the bounds of what the PS can deliver...things like this.    Looking at the work already done for us is VERY helpful, and nothing takes the place of experimenting with the ready-made designs we can play with.   And when it sounds good, it's right!

[Vivek]

Can anyone share the gain data for the circuits they analysed or designed ?

For example, in BSIAB2

@1Khz
First Stage gain = 19.22x
Second stage with gain pot max = 65.26x
Third stage gain 2.66x
Total gain from input till output of stage 3, 1Khz = 3333x

[antonis]

Are we talking about DIY effects or computerized mass production pedals..??

[Vivek]

I'm looking for per stage gain data on Amps, Amp-in-a-box and distortion pedals for jazz, blues, rock

[antonis]

Simulators are your friends, then..

[marcelomd]

Can anyone share the gain data for the circuits they analysed or designed ?

For example, in BSIAB2

@1Khz
First Stage gain = 19.22x
Second stage with gain pot max = 65.26x
Third stage gain 2.66x
Total gain from input till output of stage 3, 1Khz = 3333x

Don't forget to account for losses/attenuation between gain stages. The output of these specific three gain stages is not the same as one from one single big gain stage.

[iainpunk]

either simulations or experimentation!

i suggest you start with a single gain stage and start tone shaping,
then add a gain stage and tone-shape again, and just play with the gain and its relationship with order and tone shaping.
i once tested a 2 gain-stage circuit, with a gain knob controlling both gain stages, turning one up and the other down at the same time, the first gain stage was fat and had hard clipping this went into a thinner sounding soft clipper. changing the gain structure really gave a big plethora of morphing tones between fat fuzz and focused overdrive.
get to know what changes the tone in which way by experimenting a lot.
filtering between clipping also changes the character. overdoing low-pass might muddy the signal like a fuzz and when high-pass is done generously enough, you start to distort the generated harmonics, instead of distorting the original wave even more, resulting in in-harmonc distortion, which is generally harsh and nasty to the ear. what i want to clarify with this is that for a smooth tone, remove bass at the very beginning, and remove treble at the very end.

on a more pragmatic note,
when it comes to overdrive and light distortion, i like a buffer, followed by an opamp gain stage which shapes the frequency spectrum, followed by some sort of discrete style gain stage, like a BJT or Jfet for a 'natural clipping character'.
because the gain stage at the end, which is used for clipping, sometimes has ample gain on tap, i think its wise to have the opamp gain stage to be inverting, so it can both boost and cut level while shaping it. also, most discrete gain stages are inverting, so if you want to have Phase Continuity with the bypass signal, an inverting gain stage is necessary.
to have an inverting opamp behaving right, use a Jfet buffer to drive it.

cheers

[Vivek]

Thanks to everyone for sharing your experience and wisdom !

[ThermionicScott]

At least in the tube amp world, there is a thought that touch-sensitivity is improved when you pay close attention to the headroom at each stage.  In a non-master amp, the output tubes clip first, followed by the phase inverter, followed by the preceding gain stage, and so on as you work your way upstream to the input stage.  If your output stage is at the clipping point but the stage before it doesn't have a lot of headroom to spare at that point, you can influence whether several stages distort or clean up just based on your pick attack.   

Some vintage amps do this better than others since it wasn't a design goal back in the 1940s/1950s, but boutique amp makers have learned from the amps that did it best.  The idea is that you don't drive any particular stage too hard, since that can sound bad.

...

So when you model/measure the gain at each stage of a pedal, you may also be seeing the result of adjusting the clipping points of each stage so that they are not too far apart.

[Vivek]

I know nothing about tube amps

Thanks for this knowledge that output tubes saturate before preamp tubes

So if I understood you correctly, if we slowly increase signals, the tubes closest to cabinet saturate first

At slightly higher signal, the penultimate tubes and the output stage will saturate


So cleaner tubes drive more saturated tubes !

I have been looking at few pedals and it seems stages closest to the guitar saturate first.

[ThermionicScott]

So cleaner tubes drive more saturated tubes !

Yep!

I have been looking at few pedals and it seems stages closest to the guitar saturate first.

I think that comes down to the order of operations in a distortion pedal.  There might be a buffer up front, but then we get right into the gain/clipping stuff before applying post-distortion EQ and buffering.

Usually the clipping section does something to limit the signal coming out so it doesn't overdrive the later stages.  As far as the gain/clipping section is concerned, I think it's still true that the last stage will clip first unless there's some attenuation or headroom-limiting going on to force it the other way. 

[POTL]

keep in mind that tube amplifiers work differently. If you look at Marshall amplifiers, old models without a master volume were saturated by the power amplifier tubes and a huge volume was required to get overdrive. Eddie Van Halen used a Variac (voltage reducing device) to turn the volume down and get overdrive at low volume (this is the Brown Sound). later models such as the JCM800 2203 had 1 step more gain (in the preamp) and the master volume gets overdriven from the preamp section. I already wrote about hot and cold restrictions in another topic.

The first stage of amplification in the preamp gives a clean boost, the rest is up to the designer. Some produce symmetrical distortion, some asymmetric. Use a cathode follower for vintage sound and feel, or discard it. The number of amplification stages indirectly affects the level of overload. The famous clear sound of the Fender Blackface 65 reverb consists of three stages of amplification, the famous heavy overdrive of Diezel and Friedman also consists of three stages of amplification. The position of the EQ in the chain is also important, it can be at the beginning, middle or end.

[ElectricDruid]

The point about "gain staging" in tube amps (any amps, or pedals too, really) is a good one.

Generally, you want the amount of distortion to increase gradually (smoothly) as the input volume increases. This makes the design "responsive" and "dynamic" and other marvellous marketing words like that.

If the first stage the signal hits has a gain control that goes from x1 to x400, it'll just saturate by the time the control is barely half way up and all the other stages will see a signal that's exactly the same level (the clipping level). This is why it's best to arrange things so later stages clip earlier and work backwards towards the input with more stages clipping as the input increases.

Of course, this isn't the only way to do it, or the only way that it's done (so it's definitely not a "rule"), but it's probably important to know about.

[iainpunk]

i think the bass mid and treble controls shouldn't be in one tone stack, but divided over the stages.
if i were building a fancy high power tube (or solid state) amp, i'd go with this 'Order Of Operations'
bass > gain 1 (with gain control) > (active) mids > gain 2 > treble > volume
i think the bass mid and treble controls shouldn't be in one tone stack, but divided over the stages giving greater control over the final tone of the preamp.

cheers, Iain

[POTL]

i think the bass mid and treble controls shouldn't be in one tone stack, but divided over the stages.
if i were building a fancy high power tube (or solid state) amp, i'd go with this 'Order Of Operations'
bass > gain 1 (with gain control) > (active) mids > gain 2 > treble > volume
i think the bass mid and treble controls shouldn't be in one tone stack, but divided over the stages giving greater control over the final tone of the preamp.

cheers, Iain

I don't think it's worth generalizing. amplifiers for pure sound can have an EQ anywhere in the chain and work just fine. The fender has an EQ at the start of the chain (models with reverb). Equalizer in the middle of the chain (non-reverb models). I'm talking about the Blackface series. Vox, modern Supro or Fender Vibro King have an EQ at the end of the chain, they give a good clean sound, but less headroom and more distortion. Nearly all mid to high distortion amplifiers have EQ at the end of the chain, where it is most effective. An exception to the Mesa Boogie Mark i, where the EQ goes in the middle of the chain and is not very effective, the Dumble at the beginning of the chain is also not famous for flexibility. There are amplifiers with two equalizers, at the beginning and at the end, these are Mesa mark ii and newer.

EQ at the end of the chain was introduced to fix problems in the first generation, with an EQ at the beginning of the chain. Adjusting the treble at the start will generally increase the gain level. Low frequencies are of little use and generally they are put to the minimum position. The mids also raise the volume and thicken the sound a little, but do not feel like mids.