Understanding the Cornish style buffer (C-Buffer)

[Big Monk]

The missing context is maybe, again, that Cornish is designing things based largely on parts/methodology/design principles from the 50s-70s. You're getting a Cornish pedal not because it represents the apex of 21st century small signal amplification design or whatever, but because you want to buy the thing from the guy who's been making it the same way for 40yrs, progress be damned etc.

I think there is some nuance, at least in my implementation:

1.) Transistors and the corresponding parts seem to fit better overall in my PCB layout. Also, using BJTs means I can use the same transistors I'm using in my Fuzz and other pedals and get volume discounts.

2.) Transistor buffer stages do work well, which is a plus.

You might like the way the Cornish buffer design sounds with your guitar or how it reacts with your particular pedal design? If so cool, party onward. If you just need an input buffer for whatever reason there are much simpler ways to go.

I think his buffer is pretty complex parts wise. I personally look at it with interest mainly in the way that adding a bootstrap resistor can:

1.) Allow me to get higher input impedance without my current Darlington stage, which lets me use a transistor I already stock.

2.) Allows me to use smaller, more commonly available, and cheaper base resistors in my buffer, of which I already stock and can get a volume discount on.

I've been very conscious of cost recently and this is a natural side effect of ordering and seeing that certain posts simply cost more, etc.

 

[mzy12]

I'm not sure which schematic you're looking at, but I'm guessing that one of the caps you're referencing is the 1N between B-E? That's there to block RFI.

...

You might like the way the Cornish buffer design sounds with your guitar or how it reacts with your particular pedal design? If so cool, party onward. If you just need an input buffer for whatever reason there are much simpler ways to go.

Thanks for the response. I'm looking to add a buffer to an old Knight KG 389 diy-fuzz from the late 60's. Chuck discussed some changes here, which further expanded upon an older post by LucifersTrip on diystompboxes. What exactly am I looking from it? A consistent fuzz sound no matter where in the signal chain it's placed. After that, I don't know. I still need to breadboard the thing and try different buffers and different R1 values on the fuzz. I will also be adding a tone stack from AMZ, so I'll be adding an dual opamp to buffer the output of the fuzz and to buffer (with some makeup gain) the output of the tone control so that the volume control at the end doesn't load that down. I will, of course, be posting my results on the forum for "peer review" haha.

 

[finebyfine]

To be fair no one was discussing Cornish buffers in that thread (even though the OP meant well), so maybe it's working exactly as it should.

I tried, Robert, I tried!

Douglas Self’s high input impedance BJT stages are very cool, even if they veer into overkill territory for our applications. Those are from Small Signal Design the latest edition if I didn’t cite it in the other thread.

 

[Big Monk]

I pretty much have things integrated into my excel spreadsheet so I am going to fine tune things in the next day or two and run some comparisons.

 

[mdc]

What exactly am I looking from it? A consistent fuzz sound no matter where in the signal chain it's placed.

I guess the first question would be - have you built the stock circuit (or one of the modded versions for Chuck D B or Lucifers Trip) and discovered that you DON'T have that consistency?

 

[mzy12]

I guess the first question would be - have you built the stock circuit (or one of the modded versions for Chuck D B or Lucifers Trip) and discovered that you DON'T have that consistency?

Fantastic question - no. I really should. I have a tendency to get ahead of myself =P but it is a BJT fuzz with a low input impedance, so I already know it won't react well to having pedals placed before it. I just had questions about buffers and decided to ask them so that I could learn. If I have a question in specific to my ambitions re a redesign, I'll put it in a new thread to avoid toooo much off topic discussion.

 

[mdc]

Gotcha. With regard to the Cornish buffer and input impedance, per your original question, this DIYS thread might be useful to you?

BJTs buffers input impedance..

BJTs buffers input impedance..

www.diystompboxes.com

As an aside, one suggestion that's come up from time-to-time on other forums is in lieu of a unity(ish)-gain buffer, to use the Tillman preamp but replace the gate resistor with a 1M pot, and then put a voltage divider after it to help you achieve the amount of tone suck and volume drop that you need for a given application.

 

[mzy12]

Gotcha. With regard to the Cornish buffer and input impedance, per your original question, this DIYS thread might be useful to you?

I'll have a look at that now, thanks!

use the Tillman preamp but replace the gate resistor with a 1M pot, and then put a voltage divider after it to help you achieve the amount of tone suck and volume drop that you need for a given application.

Oooh that sounds very promising (and very interesting). I will have to try that out in addition to a clean buffer. Cheers

 

[finebyfine]

Gotcha. With regard to the Cornish buffer and input impedance, per your original question, this DIYS thread might be useful to you?

BJTs buffers input impedance..

BJTs buffers input impedance..

www.diystompboxes.com

As an aside, one suggestion that's come up from time-to-time on other forums is in lieu of a unity(ish)-gain buffer, to use the Tillman preamp but replace the gate resistor with a 1M pot, and then put a voltage divider after it to help you achieve the amount of tone suck and volume drop that you need for a given application.

Semi related to the Tillman Preamp is Gez Pelton's impedance matcher, at the bottom of this page: http://www.home-wrecker.com/wahmods.html

 

[mzy12]

Semi related to the Tillman Preamp is Gez Pelton's impedance matcher, at the bottom of this page: http://www.home-wrecker.com/wahmods.html

Also very interesting! Don't MOSFETs have higher noise floors than JFETs at lower frequencies? Would it make sense to build it with a JFET instead? Or is that just CMOS architecture? I think a 9v1 zener should also be placed between the gate and the source as protection.

 

[Big Monk]

If you want a simple buffer AND some signal gain, use a Darlington.

You should be able to go upwards of 1M input impedance AND get 10-14 dB of circuit gain.

Something like the MPSA13.

 

[Big Monk]

Okay folks.

Here is a little calculator I put together:

Loading Google Sheets

Shows Common Emitter Amplifier stages with and without bootstrapping resistor and capacitors. Also shows an Emitter Follower with and without the bootstrapping components.

I'm not sure I am accounting for the effect of the bootstrapping resistor correctly in the Emitter Follower, as I simply brought over the same calcs as in the CE Amplifier stage.

My research seemed to indicate that a bootstrapped Emitter Follower was subject to:

Input Impedance = Rbs/(1-Av)

but that seemed to yield ridiculously high input impedance vs. the Bootstrapped Common Emitter stage.

Anyway, pick it apart and maybe we can make something useful out of it.

 

[mzy12]

As an aside, one suggestion that's come up from time-to-time on other forums is in lieu of a unity(ish)-gain buffer, to use the Tillman preamp but replace the gate resistor with a 1M pot, and then put a voltage divider after it to help you achieve the amount of tone suck and volume drop that you need for a given application.

Sorry to come back to that, but where would I put the voltage divider on the schematic? Is it between R3 (now a 1M pot) and 9V? What value should I use for the extra resistor to make the voltage divider? Sorry if these are dumb questions haha

 

[mzy12]

Okay folks.

Here is a little calculator I put together:

Loading Google Sheets

Shows Common Emitter Amplifier stages with and without bootstrapping resistor and capacitors. Also shows an Emitter Follower with and without the bootstrapping components.

I'm not sure I am accounting for the effect of the bootstrapping resistor correctly in the Emitter Follower, as I simply brought over the same calcs as in the CE Amplifier stage.

My research seemed to indicate that a bootstrapped Emitter Follower was subject to:

Input Impedance = Rbs/(1-Av)

but that seemed to yield ridiculously high input impedance vs. the Bootstrapped Common Emitter stage.

Anyway, pick it apart and maybe we can make something useful out of it.

This is brilliant. Will be using it.

 

[mdc]

Sorry to come back to that, but where would I put the voltage divider on the schematic? Is it between R3 (now a 1M pot) and 9V? What value should I use for the extra resistor to make the voltage divider? Sorry if these are dumb questions haha

Sorry, by "voltage divider" I just meant your typical "volume knob." Per this schem:

R1 becomes a 1M pot - tie lug 1 to ground, lug 2 to gate.
OUTPUT goes to lug three of a 100K (or whatever) pot, lug 2 to the next stage, lug 3 to ground.

(You can remove R4 if you're adding a pot).

 

[mzy12]

Sorry, by "voltage divider" I just meant your typical "volume knob." Per this schem:

R1 becomes a 1M pot - tie lug 1 to ground, lug 2 to gate.
OUTPUT goes to lug three of a 100K (or whatever) pot, lug 2 to the next stage, lug 3 to ground.

(You can remove R4 if you're adding a pot).

Thank you for explaining. Much appreciated!

 

[Big Monk]

This is brilliant. Will be using it.

Definitely don’t lean fully on it until some other people review it.

It’s not fully vetted yet.

 

[mzy12]

Definitely don’t lean fully on it until some other people review it.

It’s not fully vetted yet.

Too late, it's already been used in a mission critical part for a satellite launch

 

[Big Monk]

Too late, it's already been used in a mission critical part for a satellite launch

In reality, the only part that’s not vetted is the bootstrapping resistor considerations for the Emitter Follower, i.e. Cornish style buffering.

I verified the Common Emitter stage, bootstrapped Common Emitter and the regular emitter follower against other calculators. I also verified the cutoff frequency calculations for Cbs1 and Cbs2 against known calculators.

I think you can probably feel good about using even the Bootstrapped Emitter Follower calcs because the input impedance calculation is conservative, i.e. it uses the same calcs from the bootstrapped Common Emitter. This were lower impedance than the calculations I found for the Bootstrapped Emitter Follower.

The one advantage my excel calcs seem to have over some of the online calculators, and big thanks to @Chuck D. Bones helping me to understand this (among many other things), is that you have a cell which lets you check if the transistor is saturating.

Many of the online calculators will give voltages that make no real world sense because they don’t error check on the input parameters, i.e. “garbage in = garbage out”.

This is particularly important in the common emitter stages because even if you are using something like a darlington to get high input impedance and some signal gain, you don’t really want the transistor in saturation and distorting.

 

[finebyfine]

Also very interesting! Don't MOSFETs have higher noise floors than JFETs at lower frequencies? Would it make sense to build it with a JFET instead? Or is that just CMOS architecture? I think a 9v1 zener should also be placed between the gate and the source as protection.

Don’t quote me on understanding this correctly but I believe you only need the gate protection diode on it if it’s the very first stage in a design, going off of this @Chuck D. Bones comment on a MOSFET AIAB: https://forum.pedalpcb.com/threads/dunwich-da-120-on-a-breadboard.13465/page-2#post-215875 But yeah MOSFETs are definitely noisier than JFETs.

I saw in a different thread elsewhere you mention SMD and some models of the SMD 2n7002 (I believe suffixes A and K? but varies by supplier) have this diode built into it