CF

[catalin gramada]

Hello. Question for experts.please.
How useful can be a cathode follower as first stage? Or with other words. do you see any inconvenience to use a cf as buffer into first stage. Just thinking at pros. better adapted to cable capacitance and the chance to use a small value volume pot at the input.
There are any reasons to not recommend it in this position.please?thanks

[g1]

Lower down on this page (similar threads section) is a link to 'CF as first stage'.
There is some relevant info starting at post #11.

[catalin gramada]

Wow thanks. I can't see Steve's circuit. It was saved somewhere?

[Mike Sulzer]

Hello. Question for experts.please.
How useful can be a cathode follower as first stage? Or with other words. do you see any inconvenience to use a cf as buffer into first stage. Just thinking at pros. better adapted to cable capacitance and the chance to use a small value volume pot at the input.
There are any reasons to not recommend it in this position.please?thanks

Sometimes it is stated that a CF increases the noise. This is not true in nearly all guitar amp circuits because the resistor in series with the grid has much more noise than the tube.

The use of a differential pair of tubes as the input (often described as a cathode follower operating into a common grid stage) eliminates the Miller effect (as does the circuit usually referred as a cascade). Stability in an amp is a complicated subject, but often it is caused by feedback form a high output stage back to the input. It is my experience in this case that eliminating the Miller effect in the first stage decreases the overall phase shift and can help in getting rid of oscillation. Some well known folks dispute this, and even I would admit that the best cure for this kind of instability in a high gain amp is extraordinary care in layout and shielding.

[SoulFetish]

Sometimes it is stated that a CF increases the noise. This is not true in nearly all guitar amp circuits because the resistor in series with the grid has much more noise than the tube.

The use of a differential pair of tubes as the input (often described as a cathode follower operating into a common grid stage) eliminates the Miller effect (as does the circuit usually referred as a cascade). Stability in an amp is a complicated subject, but often it is caused by feedback form a high output stage back to the input. It is my experience in this case that eliminating the Miller effect in the first stage decreases the overall phase shift and can help in getting rid of oscillation. Some well known folks dispute this, and even I would admit that the best cure for this kind of instability in a high gain amp is extraordinary care in layout and shielding.

(It's like having deja vu)...
It just begs the question: Why hasn't anyone come up with an example of this in a low noise - hi gain amplifier yet?

[Mike Sulzer]

(It's like having deja vu)...
It just begs the question: Why hasn't anyone come up with an example of this in a low noise - hi gain amplifier yet?

http://music-electronics-forum.com/t41822/

[SoulFetish]

http://music-electronics-forum.com/t41822/

Mike, I said LOW noise!..


(I'm just kidding. You know I'm s big fan of the Cath-code)

[loudthud]

Side issue:

What makes an amp "pedal friendly" and does the CF input stage modify or destroy it? My theory has always been that driving the input stage to clipping on the positive side is the key component that unfriendly amps lack.

[Gingertube]

The usual reason for a cathode follower at the input is to bootstrap the input impedance to make it suitable for passive ceramic pickups which need 5 to 10 Meg input impedance.
However,
A 12AX7 triode will tolerate a 5M6 (or even a 10M) grid leak resistor if you keep its idle current to less than 1mA and the anode voltage high (say +200V or more) to minimise grid current.
You can also use a Hybrid circuit which looks like a cathodyne (concertina) phase splitter with a larger anode resistor than cathode resistor, which allows you to both bootstrap the input impedance AND get some gain with improved headroom.
Many ways to skin the cat (none pleasant for the cat).
As an Engineer I would say define your requirements - let that lead the design.
Cheers,
Ian

[SoulFetish]

The usual reason for a cathode follower at the input is to bootstrap the input impedance to make it suitable for passive ceramic pickups which need 5 to 10 Meg input impedance.
However,
A 12AX7 triode will tolerate a 5M6 (or even a 10M) grid leak resistor if you keep its idle current to less than 1mA and the anode voltage high (say +200V or more) to minimise grid current.
You can also use a Hybrid circuit which looks like a cathodyne (concertina) phase splitter with a larger anode resistor than cathode resistor, which allows you to both bootstrap the input impedance AND get some gain with improved headroom.
Many ways to skin the cat (none pleasant for the cat).
As an Engineer I would say define your requirements - let that lead the design.
Cheers,
Ian

A cathode follower as an input stage does present some tempting characteristics. Certainly not the least of wich is the the availability to increase the input impedace to well over 10M (without breaking a sweat), like you mentioned. If there's a worse sounding instrument than an piezo/acoustic guitar heavily loaded, then I haven't heard it. It's a disaster. It sounds like your throwing ping-pongs at a banjo. Plus it can provide a much needed low output impedance should someone want do drive tone controls-whatever they want.
But, i can also see the other perspective. Namely, you introduce noise voltages at the input of any amplifier, however, with a traditional voltage amplifying stage input, the noise is paid for in the process of amplifying the signal to noise ration. So, with a traditional cathode follower I guess you give up the ability to do that. However we're not tethered to tradition, and i like the circuit ideas that have been mentioned. Heres my thinking (I haven't had the chance to try this yet ), I'm wondering if bootstrapping a Mu Follower may be one of the solutions. The cathode follower folks (we'll call them "followers") get thier high input impedance AND nice low output impedance. While the voltage gain folks (we'll call them "engineers"), get the voltage gain/signal to noise ratio back to reasonable levels and can get back to prophesying the coming of the end of Moore's law.

[Mike Sulzer]

A cathode follower as an input stage does present some tempting characteristics. Certainly not the least of wich is the the availability to increase the input impedace to well over 10M (without breaking a sweat), like you mentioned. If there's a worse sounding instrument than an piezo/acoustic guitar heavily loaded, then I haven't heard it. It's a disaster. It sounds like your throwing ping-pongs at a banjo. Plus it can provide a much needed low output impedance should someone want do drive tone controls-whatever they want.
But, i can also see the other perspective. Namely, you introduce noise voltages at the input of any amplifier, however, with a traditional voltage amplifying stage input, the noise is paid for in the process of amplifying the signal to noise ration. So, with a traditional cathode follower I guess you give up the ability to do that. However we're not tethered to tradition, and i like the circuit ideas that have been mentioned. Heres my thinking (I haven't had the chance to try this yet ), I'm wondering if bootstrapping a Mu Follower may be one of the solutions. The cathode follower folks (we'll call them "followers") get thier high input impedance AND nice low output impedance. While the voltage gain folks (we'll call them "engineers"), get the voltage gain/signal to noise ratio back to reasonable levels and can get back to prophesying the coming of the end of Moore's law.

Traditional gain stages do not amplify the SNR. An input section consisting of a CF followed by a gain stage has about twice the noise power of of the stage of gain alone. Anyway, in almost all guitar amplifters, the input resistor has significantly more.

[SoulFetish]

Traditional gain stages do not amplify the SNR. An input section consisting of a CF followed by a gain stage has about twice the noise power of of the stage of gain alone. Anyway, in almost all guitar amplifters, the input resistor has significantly more.

I apologize for my misunderstanding and misinformation. I thought one of the fuctions of the first gain was that it amplified the input signal above the noise floor. Is this not a component of the signal to noise? Also, this probably obvious, but can you explain how a CF has twice the noise?
Thanks, for letting me know I was incorrect by the way.

[Mike Sulzer]

I apologize for my misunderstanding and misinformation. I thought one of the fuctions of the first gain was that it amplified the input signal above the noise floor. Is this not a component of the signal to noise? Also, this probably obvious, but can you explain how a CF has twice the noise?
Thanks, for letting me know I was incorrect by the way.

It has the same noise (roughly) so it and the stage of gain: both contribute and so it doubles. Details are more complicated of course, but that it the general idea.

One way to think of it is that the noise of a tube is a random voltage source in series with the input. Both signal and noise get amplified together.

[SoulFetish]

It has the same noise (roughly) so it and the stage of gain: both contribute and so it doubles. Details are more complicated of course, but that it the general idea.

One way to think of it is that the noise of a tube is a random voltage source in series with the input. Both signal and noise get amplified together.

Oh, yeah, yeah. Okay, it obviously makes sense that an amplifying stage would amplify the voltages present at its input noise and signal equally(for the most part). I see where My misunderstanding was confusing the difference in voltage that gets amplified between signal and noise being equal to the ratio. Which is not amplified.
More accurate?
So It makes more sense that the eliminating noise sources and choosing low noise topologies are the deciding factors in which voltages get either amplified or buffered and set the signalto noise ratio--which will ulimately be amplified or buffered?