Pete Cornish - Bypass buffer  [traced]

General documentation, gut shot, schematic links, ongoing circuit tracing, deep thoughts ... all about boutique stompboxes.
User avatar
andregarcia57
Cap Cooler
Information
Posts: 572
Joined: 16 Nov 2008, 15:42
Has thanked: 73 times
Been thanked: 109 times

Post by andregarcia57 »

bc239c will work well in this circuit?

which the correct value resistor to last? 51R or 51k

thanks

User avatar
monkeyxx
Resistor Ronker
Information
Posts: 279
Joined: 12 Jan 2011, 06:54
Has thanked: 52 times
Been thanked: 51 times

Post by monkeyxx »

Just use one of the verified layouts. They're a little different for some reason, maybe traced from different pedals?

I bet BC239c would work just fine

https://www.diystompboxes.com/smfforum/ ... 199.0;wap2

(scroll down to RG Keen post)

User avatar
andregarcia57
Cap Cooler
Information
Posts: 572
Joined: 16 Nov 2008, 15:42
Has thanked: 73 times
Been thanked: 109 times

Post by andregarcia57 »

I do not have 50k!

Here I put 47k or 51k? which would be better?

thanks

User avatar
marshmellow
Cap Cooler
Information
Posts: 469
Joined: 16 Aug 2007, 07:31
Location: Germany
Has thanked: 20 times
Been thanked: 127 times

Post by marshmellow »

Leave it out, it's not necessary if you build the buffer as standalone. If you decide you still want it either value is fine, it won't change a thing.

User avatar
clintrubber
Breadboard Brother
Information
Posts: 99
Joined: 08 Oct 2008, 19:21
Location: The Netherlands
Has thanked: 16 times
Been thanked: 5 times
Contact:

Post by clintrubber »

monkeyxx wrote: I would also love some insight to why Cornish would use two buffers in a row in a circuit like the G2, one for bypass, and one on the circuit input before the actual effect.
Wondering about that too! Any ideas about this meanwhile?

The second buffer seems redundant.

Thanks / Bye

User avatar
clintrubber
Breadboard Brother
Information
Posts: 99
Joined: 08 Oct 2008, 19:21
Location: The Netherlands
Has thanked: 16 times
Been thanked: 5 times
Contact:

Post by clintrubber »

clintrubber wrote:
monkeyxx wrote: I would also love some insight to why Cornish would use two buffers in a row in a circuit like the G2, one for bypass, and one on the circuit input before the actual effect.
Wondering about that too! Any ideas about this meanwhile?

The second buffer seems redundant.
OK, I saw there's a thread about this:

https://www.freestompboxes.org/viewtopic ... 10&t=23823

User avatar
Skreddy
Resistor Ronker
Information
Posts: 419
Joined: 11 Nov 2007, 04:22
Has thanked: 12 times
Been thanked: 228 times
Contact:

Post by Skreddy »

The redundant buffers are there because even with the buffered bypass, the input is still technically "sucking" tone unless its impedance is also raised with a buffer.

User avatar
clintrubber
Breadboard Brother
Information
Posts: 99
Joined: 08 Oct 2008, 19:21
Location: The Netherlands
Has thanked: 16 times
Been thanked: 5 times
Contact:

Post by clintrubber »

Skreddy wrote:The redundant buffers are there because even with the buffered bypass, the input is still technically "sucking" tone unless its impedance is also raised with a buffer.
That would indeed make sense if one was to address say 'second order' influences.

But if I'm looking at the correct circuits, then:

* the input-'Z' of the actual distortion circuit (after the second seemingly redundant buffer) is at least 39k

* the input buffer itself is loaded by 20k (R7) after the AC-coupling


So one could say that the second buffer could be removed, R7 could be increased to 39k as well and the loading of the input-buffer would be at least as high as before.

But I understood from other messages (DH?) that mr. C is a strict traditionalist so other reasons may come into play. I could agree with that to a certain extend, if it's proven & everybody is satisfied then all is OK - but it can sure confuse a proper understanding of a circuit :roll: :lol:

Sure interested to hear other ideas for this, I remain curious to the why - so far & in all respect I'd say it doesn't hurt to use two buffers (it's a distortion box!), but with this dimensioning it doesn't make sense either.

Best regards

User avatar
Skreddy
Resistor Ronker
Information
Posts: 419
Joined: 11 Nov 2007, 04:22
Has thanked: 12 times
Been thanked: 228 times
Contact:

Post by Skreddy »

clintrubber wrote:
Skreddy wrote:The redundant buffers are there because even with the buffered bypass, the input is still technically "sucking" tone unless its impedance is also raised with a buffer.
That would indeed make sense if one was to address say 'second order' influences.

But if I'm looking at the correct circuits, then:

* the input-'Z' of the actual distortion circuit (after the second seemingly redundant buffer) is at least 39k

* the input buffer itself is loaded by 20k (R7) after the AC-coupling


So one could say that the second buffer could be removed, R7 could be increased to 39k as well and the loading of the input-buffer would be at least as high as before.

But I understood from other messages (DH?) that mr. C is a strict traditionalist so other reasons may come into play. I could agree with that to a certain extend, if it's proven & everybody is satisfied then all is OK - but it can sure confuse a proper understanding of a circuit :roll: :lol:

Sure interested to hear other ideas for this, I remain curious to the why - so far & in all respect I'd say it doesn't hurt to use two buffers (it's a distortion box!), but with this dimensioning it doesn't make sense either.

Best regards

The input Z of the distortion circuit isn't just resistance to ground like the buffer's output stage; it's resistance to ground through an input cap, which would pull down the high end. The buffer at the input before the switch handles the input impedance for the whole circuit, and the buffer in front of the distortion stage prevents tone suck when in bypass.

Myself, I would prefer just simple true bypass and zero added buffers. I see buffers as a tool to use when needed, not something every circuit needs all the time. But that's just me. People seem happy with Cornish's stuff, and at least he gets a little more reliability out of his 3PDT switches by ganging up 2 switches for his non-true-bypass switching. So there's that.

User avatar
clintrubber
Breadboard Brother
Information
Posts: 99
Joined: 08 Oct 2008, 19:21
Location: The Netherlands
Has thanked: 16 times
Been thanked: 5 times
Contact:

Post by clintrubber »

Skreddy wrote: The input Z of the distortion circuit isn't just resistance to ground like the buffer's output stage; it's resistance to ground through an input cap, which would pull down the high end. The buffer at the input before the switch handles the input impedance for the whole circuit, and the buffer in front of the distortion stage prevents tone suck when in bypass.
Hi,

I'm referring to the 39k resistors on page 2 of this thread (R13 or R15, depending on the schematic you look at).

The load of the second buffer is higher than, but at least never no nothing below that 39k (as seen from the output of the preceeding ('second') buffer.

Bye

User avatar
Skreddy
Resistor Ronker
Information
Posts: 419
Joined: 11 Nov 2007, 04:22
Has thanked: 12 times
Been thanked: 228 times
Contact:

Post by Skreddy »

clintrubber wrote:
Skreddy wrote: The input Z of the distortion circuit isn't just resistance to ground like the buffer's output stage; it's resistance to ground through an input cap, which would pull down the high end. The buffer at the input before the switch handles the input impedance for the whole circuit, and the buffer in front of the distortion stage prevents tone suck when in bypass.
Hi,

I'm referring to the 39k resistors on page 2 of this thread (R13 or R15, depending on the schematic you look at).

The load of the second buffer is higher than, but at least never no nothing below that 39k (as seen from the output of the preceeding ('second') buffer.

Bye
Yep; I was looking at the G2 schematic. Without the buffer on the distortion front end, it would suck tone through the 47n capacitor, since the bypass would keep the input hanging onto the signal in bypass. Probably not horrible given the input buffer driving the signal, but still a tone suck nonetheless. Big Muff inputs are one of the 2 most famous tone suckers when not true bypassed (or buffered), 2nd only to wah pedals.

User avatar
clintrubber
Breadboard Brother
Information
Posts: 99
Joined: 08 Oct 2008, 19:21
Location: The Netherlands
Has thanked: 16 times
Been thanked: 5 times
Contact:

Post by clintrubber »

Skreddy wrote:
clintrubber wrote:
Skreddy wrote: The input Z of the distortion circuit isn't just resistance to ground like the buffer's output stage; it's resistance to ground through an input cap, which would pull down the high end. The buffer at the input before the switch handles the input impedance for the whole circuit, and the buffer in front of the distortion stage prevents tone suck when in bypass.
Hi,

I'm referring to the 39k resistors on page 2 of this thread (R13 or R15, depending on the schematic you look at).

The load of the second buffer is higher than, but at least never no nothing below that 39k (as seen from the output of the preceeding ('second') buffer.

Bye
Yep; I was looking at the G2 schematic. Without the buffer on the distortion front end, it would suck tone through the 47n capacitor, since the bypass would keep the input hanging onto the signal in bypass. Probably not horrible given the input buffer driving the signal, but still a tone suck nonetheless. Big Muff inputs are one of the 2 most famous tone suckers when not true bypassed (or buffered), 2nd only to wah pedals.
Hi Skreddy,

The 47n you mention, that's C8 in the G2-schematic and is in series with the 39k, correct? The tone suck you mention will indeed happen for a 'naked' Big Muff circuit (due to the mid-to-high output impedance of a (passive) guitar & the relatively low input impedance of the Big Muff circuit.

BUT since the first input buffer in the Cornish setup already completely isolates these two, the second buffer is in my understanding not necessary anymore in view of tone sucking.

That's what I meant by those figures: the 39k is ~double in value w.r.t. the 'dummy load' 20k resistor of the first buffer.

So imho it makes little sense [A] to have a first buffer, give that a 20k 'dummy-load' and [C] then to prevent any loading by a next circuit that has at least 39k input impedance (it's higher, but let's set an absolute lower limit here for now) one adds a second buffer. Other reasons than technical may exist, but it's not straightforward.

(Note that the 47n in series with the 39k could be considered just a thing that determines for which frequencies the 39k 'applies'. In other words: it doesn't lower the 39k in any way, since it's in series.)

Bye

User avatar
Skreddy
Resistor Ronker
Information
Posts: 419
Joined: 11 Nov 2007, 04:22
Has thanked: 12 times
Been thanked: 228 times
Contact:

Post by Skreddy »

clintrubber wrote:
Skreddy wrote:
clintrubber wrote:
Skreddy wrote: The input Z of the distortion circuit isn't just resistance to ground like the buffer's output stage; it's resistance to ground through an input cap, which would pull down the high end. The buffer at the input before the switch handles the input impedance for the whole circuit, and the buffer in front of the distortion stage prevents tone suck when in bypass.
Hi,

I'm referring to the 39k resistors on page 2 of this thread (R13 or R15, depending on the schematic you look at).

The load of the second buffer is higher than, but at least never no nothing below that 39k (as seen from the output of the preceeding ('second') buffer.

Bye
Yep; I was looking at the G2 schematic. Without the buffer on the distortion front end, it would suck tone through the 47n capacitor, since the bypass would keep the input hanging onto the signal in bypass. Probably not horrible given the input buffer driving the signal, but still a tone suck nonetheless. Big Muff inputs are one of the 2 most famous tone suckers when not true bypassed (or buffered), 2nd only to wah pedals.
Hi Skreddy,

The 47n you mention, that's C8 in the G2-schematic and is in series with the 39k, correct? The tone suck you mention will indeed happen for a 'naked' Big Muff circuit (due to the mid-to-high output impedance of a (passive) guitar & the relatively low input impedance of the Big Muff circuit.

BUT since the first input buffer in the Cornish setup already completely isolates these two, the second buffer is in my understanding not necessary anymore in view of tone sucking.

That's what I meant by those figures: the 39k is ~double in value w.r.t. the 'dummy load' 20k resistor of the first buffer.

So imho it makes little sense [A] to have a first buffer, give that a 20k 'dummy-load' and [C] then to prevent any loading by a next circuit that has at least 39k input impedance (it's higher, but let's set an absolute lower limit here for now) one adds a second buffer. Other reasons than technical may exist, but it's not straightforward.

(Note that the 47n in series with the 39k could be considered just a thing that determines for which frequencies the 39k 'applies'. In other words: it doesn't lower the 39k in any way, since it's in series.)

Bye



We're both repeating the same 2 arguments back and forth. Let me make clear, I am not talking about resistance to ground interacting with the guitar. I am talking about 47n through some resistance to ground creating a low-pass filter. It's not a very powerful low-pass filter, but it is a tone suck nonetheless. I already said that the buffer would likely minimize this effect, but it is still there. A low-pass filter is a low-pass filter. That's why we like to true-bypass Big Muffs. Yes, driving a low-pass filter with a buffer will minimize it (maybe you wouldn't even hear it?). But it is still there.

39k in series with a 100k to ground, 470k to negative feedback resistor, and 680R x ~300 Hfe equals about 97,721.22 ohms. Take any circuit and hang a 47n to 100k resistor to ground onto its output, and you've got yourself a (weak) low-pass filter. Myself, I would probably INCLUDE some amount of low-pass filtering IN MY BUFFER DESIGN to keep it from over-brightening the signal. But not that much. Not 47n to 100k to ground. I'd be working more like in the 1M region to just gently curve that high end down a bit and use smaller capacitors than 47n.

I'm not defending Pete Cornish's design philosophy; I have made the same exact criticism of it myself. I am just explaining why he did it this way.
Last edited by Skreddy on 31 Mar 2015, 14:45, edited 1 time in total.

User avatar
clintrubber
Breadboard Brother
Information
Posts: 99
Joined: 08 Oct 2008, 19:21
Location: The Netherlands
Has thanked: 16 times
Been thanked: 5 times
Contact:

Post by clintrubber »

Hi,

In all respect, I give up. I don't think your technical arguments are valid.

A 'filter' after 'just one' decent buffer instead of two cascaded decent buffers can not likely impart more 'tone suck' on the original source signal, unless a bizarre layout or some other unneeded non-technical influence has had its ways.

BTW, a coupling-cap followed by a resistor to ground is a highpass, not a lowpass.

Note these buffers are in series, so NOT connected to the same input-jack .... maybe that's what creating the confusion / different interpretations ?
(This was my last attempt, I'm out now. Like you say, we're both stubborn &^%'s who might perhaps better be spending time on playing guitar than spending time here ;-) OK, I stop. Bye)

User avatar
Skreddy
Resistor Ronker
Information
Posts: 419
Joined: 11 Nov 2007, 04:22
Has thanked: 12 times
Been thanked: 228 times
Contact:

Post by Skreddy »

clintrubber wrote:Hi,

In all respect, I give up. I don't think your technical arguments are valid.

A 'filter' after 'just one' decent buffer instead of two cascaded decent buffers can not likely impart more 'tone suck' on the original source signal, unless a bizarre layout or some other unneeded non-technical influence has had its ways.

BTW, a coupling-cap followed by a resistor to ground is a highpass, not a lowpass.

Note these buffers are in series, so NOT connected to the same input-jack .... maybe that's what creating the confusion / different interpretations ?
(This was my last attempt, I'm out now. Like you say, we're both stubborn &^%'s who might perhaps better be spending time on playing guitar than spending time here ;-) OK, I stop. Bye)
A coupling cap followed by a resistor to ground creates a highpass filter for the input of the circuit that follows, yes. But we are talking about a cap to resistance to ground's effect on the BYPASS, not the distortion circuit. The bypass sees that cap (to resistance) to ground hanging off it as a low-pass filter.

Build yourself a buffer. Then put a 47n to 100k to ground on its output and see if you can't hear the difference. Maybe you won't hear it? Maybe it'll be an improvement? At any rate, Pete wouldn't design a bypass that included that amount of tone suck hanging off of it--hence the extra buffer in front of the distortion circuit, even after the input buffer. Because, again, his bypass switching is not true bypass. It keeps the input of the distortion circuit connected when "off". So he buffers the distortion even after buffering the input.

Edit: You must be talking about the need for the 1st buffer, the "global" input buffer which is active all the time? Yeah, that's just Pete's design philosophy. Everything buffered all the time. Again, myself, I prefer simply true bypassing everything and then just using buffer(s) judiciously where needed.

User avatar
Skreddy
Resistor Ronker
Information
Posts: 419
Joined: 11 Nov 2007, 04:22
Has thanked: 12 times
Been thanked: 228 times
Contact:

Post by Skreddy »

Explaining the Cornish bypass buffer in pictures: Why the redundant buffer? (these portray the circuit when in bypass)

Image

Image

User avatar
lolbou
Old Solderhand
Information
Posts: 2617
Joined: 18 Nov 2008, 21:38
Has thanked: 308 times
Been thanked: 225 times

Post by lolbou »

So basically the first one is for cable driving in bypass mode up to the next pedal or even the furthest amp, and the second one is to make sure that the cable driving is not affected by any effect input?

I get your point Skreddy (and it's well explained! :thumbsup ), but I have to try your buffer + 47n-100k network to ground to hear. I have a LD-1 clone here to help... I really would have thought that the first buffer was strong enough to cope for both loss though...

I'll definitely give it a try and report!
- Are you a mod or a rocker?
- Uh, no, I'm a mocker.

User avatar
Skreddy
Resistor Ronker
Information
Posts: 419
Joined: 11 Nov 2007, 04:22
Has thanked: 12 times
Been thanked: 228 times
Contact:

Post by Skreddy »

lolbou wrote:So basically the first one is for cable driving in bypass mode up to the next pedal or even the furthest amp, and the second one is to make sure that the cable driving is not affected by any effect input?

I get your point Skreddy (and it's well explained! :thumbsup ), but I have to try your buffer + 47n-100k network to ground to hear. I have a LD-1 clone here to help... I really would have thought that the first buffer was strong enough to cope for both loss though...

I'll definitely give it a try and report!
The point of the input buffer is to prevent signal loading at the input. The point of the "redundant" buffer is to prevent signal loading at the output. You may not hear the loading effect of a Big Muff input after a buffer, but that's not really the point. The point is not to leave any messiness in the signal chain at all when in bypass, because if you had 5 or 10 Cornish pedals on a pedalboard, there should not be any audible effect to the signal when they're all in bypass. In reality the cumulative effect of multiple active buffers will have some effect; the only arguable point is how much of an effect, is it audible, and is it acceptable if so (maybe it's an overall improvement?). But to leave the inputs of vintage-type distortion/fuzz boxes hanging on when in bypass would not be very good design quality just on principle alone. And apart from the low-pass filtering effect of a non-true-bypassed Big Muff, the hanging input will also impart a bit of clipping artifact to the signal as well.

User avatar
clintrubber
Breadboard Brother
Information
Posts: 99
Joined: 08 Oct 2008, 19:21
Location: The Netherlands
Has thanked: 16 times
Been thanked: 5 times
Contact:

Post by clintrubber »

Hi,

I get all that. I see the mechanism of your point. But please do the math, look at the component values, and not just the topology.

The output impedance of the first buffer is say 38 (TR1) + 51 (R6) = 89 Ohm.

With that, the tone-suck attenuation would never be more than 0.020 dB ( 20*log(39k/39k089) )

My ears won't detect that. I bet yours neither.

And that's assuming the BigMuff-inp impedance is just that 39k, but it'll be even higher.

In addition, the corner frequency might be pretty low, so resulting in an even flatter attenuation for all guitar-frequencies.
(filter-curve transforms then in a flat attenuation)

Regards

User avatar
collingtech
Breadboard Brother
Information
Posts: 73
Joined: 08 Feb 2010, 18:52
Has thanked: 19 times
Been thanked: 25 times

Post by collingtech »

so you must put one buffer in every effect input ? if i have a 12 diy pedal this means 12 buffers? and about impedance correction? someone could explain better the placing ?
thanks

Post Reply