What is a "buffer" ?

[Greg Ellis]

Can someone explain the concept of a "buffer" as it applies to a signal chain for guitar?

What is this thing, what does it do, what problem does it solve, and how, and where would you place it in the signal chain, etc?

Thanks

[Hamm Guitars]

A buffer amplifier is a device which can be used to interface two different devices (or stages of a circuit). The most common use is for impedance matching.

[zdogma]

Agreed.

If its done well, it will help to prevent loss of high end due to capacitance in long runs of cable or a whole bunch of true bypass effects. Trouble is, a lot of the buffers color the tone a bit, and when you put a whold bunch of them together it can color the tone a lot.

Also certain effects, like fuzz pedals don't work as well when the signal is buffered, they don't clean up well and they lose a lot of the "feel" that is so cool.

I like the buffers in certain pedals (like the visual sound and the Tonebone series) but I always try to stick the fuzz or wah on the front.

[mhammer]

We tend to think of the input to some pedal or amp as essentially "seeing" only the signal. Actually, the input "sees" multiple sources, some of which carry no signal other than thermal noise. For example, when you plug your guitar directly into the amplifier, there will be multiple paths to ground that the "input" sees. One of those is likely to be a resistor to ground at the input jack or some equivalent place on a circuit board. At the same time, there is:

1) A path to ground between the hot lead of the cord and the shielding on the cord (i.e., your cable is one big capacitor).

2) A path from the wiper (variable contact) of your volume pot to ground.

3) A path from the wiper of the volume pot through the tone control to ground.

4) A path from the wiper of the volume pot through every pickup connected in circuit to ground.

So, uh, just exactly what does the amp think is the "real" signal? Well, the amp doesn't think at all. What it does is lean towards the source whose impedance is most favourable.

What a buffer does is adapt the output impedance of the thing you want the next device in line (NDIL; whether amp or pedal) to favour so that the NDIL "sees" that source more than anything else connected to it.

Keep in mind, as well, that there can sometimes be summed effects that straddle several pedals. For example, in the early days of stompboxes, it was common to use a SPDT footswitch to select between the input jack of a pedal or the output of the electronic circuit. The actual circuit contained in the chassis always remained in contact with the input jack, so that if there was a 330k resistor tied to the input of the effect circuit, that 330k resistor was always there whether you wanted it or not. When we were back in the days of using a single effect, that was less of a problem. Once you plug in several pedals, though, and all their inputs are electronically connected even when in bypass mode, their combined parallel resistance starts to compete with the actual guitar signal. THAT is when you get tone-sucking, and THAT is why manufacturers started to switch to other bypass schemes than a mere SPDT switch; to disconnect the input resistors such that they would never inadvertently be placed in parallel with each other.

Of course, if the guitar's output impedance were low enough, even several input resistors in parallel would pose no problem. As well, if the guitar remained a low output-impedance device, then even long cables and their associated cable capacitance, would pose little problem. So, we saw the move towards having guitars with active electronics that would buffer the output and always remain the "preferred" signal source at the amp's input. I wrote about this back in 1979 (http://hammer.ampage.org/files/Device1-12.PDF) so you can see that it has been a source of concern for a while.

One of the related questions that perennially crops up is "What's the difference between a buffer and a booster/preamp?". Unless it is very poorly designed, ALL boosters and preamps have a buffering effect. That is, the input of the circuit has a very high input impedance that "sees" as much of the guitar signal as is electronically possible and available, and provides a low output impedance to enable the NDIL to see as much signal as possible too. That same basic circuit could be configured to add some gain, merely by adjusting some component values...or not. If it adds no gain, it will be "just" a buffer. If it adds gain as well as buffering, then we call it a booster or preamp. As for what makes the difference between those two, my own take on it is that if it brings the signal up to some expected or needed level, then we call it a preamp, and if it brings the signal level up beyond some expected level, we call it a booster. Feel free to choose another use of those terms.

Where do you put a buffer? That depends what is connected to it. A great many commercial pedals include a buffer at the input (and output too, sometimes) that is never off or disconnected from the signal path. They do this because they have no idea what in tarnation you will be feeding their pedal with, or feeding its output to. So, essentially the presence of buffers in all those Boss, DOD, and other pedals is a possibly superfluous attempt to deal with the unexpected and worst case scenario.

Enter "true bypass".

All those buffers, especially if you have a pedalboard with 6-10 pedals on it, can cumulatively add hiss and trim some of your high end and low end bandwidth. Not good. The move to use simple mechanical "true bypass" switching was motivated partly to eliminate the problems from vintage pedals that used only SPDT, and to sidestep all those buffers, but it was also motivated by a perceived need to completely remove the pedal electronics from the path in case your battery went on you. But here's the thing. If you plug your guitar into a 20' cord and then into the first of a bunch of TB pedals, and you run the last of those pedals into a 25' cable to the amp, when ALL your pedals are in bypass mode, your guitar is essentially plugged into a 45' cable to your amp. MAJOR tone suckage. BIG time. If you ever want a persuasive lesson in cable capacitance and its evil effects, plug your guitar into an amp set to ultra-clean through the absolute shortest cable you have (e.g., one of those colourful molded pedal patch cables), and then through the absolute longest cable you can find. You'll be shocked by what thelonger cable takes away.

Of course, if your guitar were to provide a buffered output then all that cable would matter much much less.

So, ideally, there should be nothing between guitar and amp that reduces the amp's sensitivity to the full bandwidth of the guitar signal. In a perfect world, that would mean a buffer built into the guitar. Of course, you can't always do that, so the next best thing is to at least use the shortest practical cable you can leading to the first pedal. That first pedal/device should provide an always-buffered output that remains in place whether you have that sucker set for gain or not. Once THAT high-to-low translation has taken place, then you can, in theory, stick all sorts of stuff after it on the way to the amp, including 20 true-bypass pedals with 8" cables between 'em and a 25 footer to the amp, and the amp will always "want" to see everything the guitar can give it.

Long and the short of it is that buffering should happen as close to the signal source as is practical, and need only happen once under most circumstances. The exception would be something like the case where you run a cable from a pedal-board to a remote mixer at a festival (the far end of Gage Park, Stanley Park, or Hawrlak Park, or some other equally beautiful equivalent). In those instances, you would need a high-current buffer so that the series resistance of a long cable does not erode the signal. Jack Orman has a design for one at his AMZ sight (www.muzique.com).

Any other questions?

[Greg Ellis]

WOW, thanks!

I'm starting to like this website a LOT!

[LowWatt]

WOW, thanks!

I'm starting to like this website a LOT!

If you spent a day just looking up posts by mhammer and Wild Bill you would no more about guitar and amp electronics than 99% of the musicians and techs out there. Those guys know their facts cold and are neevr shy about giving a thoughtful detailed explanation.

[JHarasym]

Sure, provided we comprehend what these guys are telling us...

[mhammer]

It's always difficult to come up with a single explanation that makes sense to absolutely everyone with ANY level of background knowledge. Feel free to boldface or underline anything that went pfffft over your head, and I'll be happy to explain it another way, either here or offline. I'm one of those doofuses who sees understanding as a basic human right.

Mark

[ENDITOL]

Great post! :food-smiley-004:

We tend to think of the input to some pedal or amp as essentially "seeing" only the signal. Actually, the input "sees" multiple sources, some of which carry no signal other than thermal noise. For example, when you plug your guitar directly into the amplifier, there will be multiple paths to ground that the "input" sees. One of those is likely to be a resistor to ground at the input jack or some equivalent place on a circuit board. At the same time, there is:

1) A path to ground between the hot lead of the cord and the shielding on the cord (i.e., your cable is one big capacitor).

2) A path from the wiper (variable contact) of your volume pot to ground.

3) A path from the wiper of the volume pot through the tone control to ground.

4) A path from the wiper of the volume pot through every pickup connected in circuit to ground.

So, uh, just exactly what does the amp think is the "real" signal? Well, the amp doesn't think at all. What it does is lean towards the source whose impedance is most favourable.

What a buffer does is adapt the output impedance of the thing you want the next device in line (NDIL; whether amp or pedal) to favour so that the NDIL "sees" that source more than anything else connected to it.

Keep in mind, as well, that there can sometimes be summed effects that straddle several pedals. For example, in the early days of stompboxes, it was common to use a SPDT footswitch to select between the input jack of a pedal or the output of the electronic circuit. The actual circuit contained in the chassis always remained in contact with the input jack, so that if there was a 330k resistor tied to the input of the effect circuit, that 330k resistor was always there whether you wanted it or not. When we were back in the days of using a single effect, that was less of a problem. Once you plug in several pedals, though, and all their inputs are electronically connected even when in bypass mode, their combined parallel resistance starts to compete with the actual guitar signal. THAT is when you get tone-sucking, and THAT is why manufacturers started to switch to other bypass schemes than a mere SPDT switch; to disconnect the input resistors such that they would never inadvertently be placed in parallel with each other.

Of course, if the guitar's output impedance were low enough, even several input resistors in parallel would pose no problem. As well, if the guitar remained a low output-impedance device, then even long cables and their associated cable capacitance, would pose little problem. So, we saw the move towards having guitars with active electronics that would buffer the output and always remain the "preferred" signal source at the amp's input. I wrote about this back in 1979 (http://hammer.ampage.org/files/Device1-12.PDF) so you can see that it has been a source of concern for a while.

One of the related questions that perennially crops up is "What's the difference between a buffer and a booster/preamp?". Unless it is very poorly designed, ALL boosters and preamps have a buffering effect. That is, the input of the circuit has a very high input impedance that "sees" as much of the guitar signal as is electronically possible and available, and provides a low output impedance to enable the NDIL to see as much signal as possible too. That same basic circuit could be configured to add some gain, merely by adjusting some component values...or not. If it adds no gain, it will be "just" a buffer. If it adds gain as well as buffering, then we call it a booster or preamp. As for what makes the difference between those two, my own take on it is that if it brings the signal up to some expected or needed level, then we call it a preamp, and if it brings the signal level up beyond some expected level, we call it a booster. Feel free to choose another use of those terms.

Where do you put a buffer? That depends what is connected to it. A great many commercial pedals include a buffer at the input (and output too, sometimes) that is never off or disconnected from the signal path. They do this because they have no idea what in tarnation you will be feeding their pedal with, or feeding its output to. So, essentially the presence of buffers in all those Boss, DOD, and other pedals is a possibly superfluous attempt to deal with the unexpected and worst case scenario.

Enter "true bypass".

All those buffers, especially if you have a pedalboard with 6-10 pedals on it, can cumulatively add hiss and trim some of your high end and low end bandwidth. Not good. The move to use simple mechanical "true bypass" switching was motivated partly to eliminate the problems from vintage pedals that used only SPDT, and to sidestep all those buffers, but it was also motivated by a perceived need to completely remove the pedal electronics from the path in case your battery went on you. But here's the thing. If you plug your guitar into a 20' cord and then into the first of a bunch of TB pedals, and you run the last of those pedals into a 25' cable to the amp, when ALL your pedals are in bypass mode, your guitar is essentially plugged into a 45' cable to your amp. MAJOR tone suckage. BIG time. If you ever want a persuasive lesson in cable capacitance and its evil effects, plug your guitar into an amp set to ultra-clean through the absolute shortest cable you have (e.g., one of those colourful molded pedal patch cables), and then through the absolute longest cable you can find. You'll be shocked by what thelonger cable takes away.

Of course, if your guitar were to provide a buffered output then all that cable would matter much much less.

So, ideally, there should be nothing between guitar and amp that reduces the amp's sensitivity to the full bandwidth of the guitar signal. In a perfect world, that would mean a buffer built into the guitar. Of course, you can't always do that, so the next best thing is to at least use the shortest practical cable you can leading to the first pedal. That first pedal/device should provide an always-buffered output that remains in place whether you have that sucker set for gain or not. Once THAT high-to-low translation has taken place, then you can, in theory, stick all sorts of stuff after it on the way to the amp, including 20 true-bypass pedals with 8" cables between 'em and a 25 footer to the amp, and the amp will always "want" to see everything the guitar can give it.

Long and the short of it is that buffering should happen as close to the signal source as is practical, and need only happen once under most circumstances. The exception would be something like the case where you run a cable from a pedal-board to a remote mixer at a festival (the far end of Gage Park, Stanley Park, or Hawrlak Park, or some other equally beautiful equivalent). In those instances, you would need a high-current buffer so that the series resistance of a long cable does not erode the signal. Jack Orman has a design for one at his AMZ sight (www.muzique.com).

Any other questions?