Cornish Buffer Help

[mickeybellinello]

Thanks so so much to all for precious infos.

[yeeshkul]

Guys can you explain how R5 and C3 work? Or just please send me a link to bootstrapping, i don't get how it works.

[antonis]

Neither R5 nor C3 have to do anything with bootstrappibg..

R5/R6 set Base bias voltage where C4 is used for R4 bootrstrapping capacitor..
(it makes R4 apparent value much higher, hence it raises bias configuration input impedance..)

https://electronics.stackexchange.com/questions/77255/circuit-analysis-of-a-emitter-follower-with-bootstrap

[bonehead1972]

hey guys,
looks like i just made it. got 5.25V at the base and 4.7V at the emiter with a 9V supply .  are these voltages good enough? thanks!

[antonis]

 

[bonehead1972]

I would like to ask you for your opinion on voltages I got because i don't have any understanding on how this circuit  works. In order to get  as close as possible to suggested values in the schematic I used whatever old parts I had on hand with resistors in parallel at a few points. Thanks..

[antonis]

Just to have a quick reference..



Bias voltages are established as follows:

R5 & R6 form a voltage divider of about 5.6V..
(9 X 200k/(200k + 120k)
R4 connects 5.6V point to Q1 Base..
(with a small voltage drop across itsef - about 300mV or so(*))

Base-Emitter junction exhibits a voltage drop of about 600mV (**)

So, Emitter stands at 4.7 which is an almost ideal(***) quiescent voltage point..

P.S.
(*) It depends on Collector current, Q1 h_FE, R4 value and R5/R6 Thevenin equivalent resistance..
(**) It depends on Collector current, working temperature and particular device..
(***) "Ideal" Emitter voltage should be V_CC/2 (Collector supply midpoint) but practically must be a bit higher both due to transistor and Emitter follower configuration non-linear beahavior..

[bonehead1972]

Thanks for being so helpful Yamas!

[yeeshkul]

Thank you Antonis. And what is C3 for?

[antonis]

And what is C3 for?

For RF rectification prevention..
(it bypasses Base-Emitter junction at high frequencies..)

P.S.
IMHO, it isn't essential for CC amp (Emitter follower) but it makes no harm either..

[yeeshkul]

Thank you Antonis, i suspected that

[yeeshkul]

Actually, why doesn't C3 lead straight to the GND?

[PRR]

Because large RF at *either* B or E will rectify and interfere. But a cap *across* B and E reduces RFI all ways.

[Rob Strand]

And what is C3 for?

For RF rectification prevention..
(it bypasses Base-Emitter junction at high frequencies..)

P.S.
IMHO, it isn't essential for CC amp (Emitter follower) but it makes no harm either..

Interestingly I was doing some work on a wideband amp and playing around with various ways of controlling oscillation in buffer stages due to capacitive loading.     Adding a base-emitter cap can help prevent oscillation (although it's not good for bandwidth.)

You can see the effect of capacitive loading in replies #30 and #31 and other comments in this thread,

https://www.diystompboxes.com/smfforum/index.php?topic=119501.20

The addition of the cap across the base and emitter can help control oscillations without much effect on loading the guitar.  Notice how the Cornish buffer uses a low 150 ohm output resistor.  The low value promotes oscillations.  The post I linked showed oscillations even with somewhat higher output resistors.    The base-emitter cap on the Cornish may very well help remove oscillations with the 150 ohm output resistor.  I haven't analysed the Cornish circuit specifically.

[antonis]

The addition of the cap across the base and emitter can help control oscillations without much effect on loading the guitar.  Notice how the Cornish buffer uses a low 150 ohm output resistor.  The low value promotes oscillations.  The post I linked showed oscillations even with somewhat higher output resistors.    The base-emitter cap on the Cornish may very well help remove oscillations with the 150 ohm output resistor.

I'm not sure if Base-Emitter cap prevents oscillations or not 'cause as it's placed it actually "helps" to create a Colpitts oscillator with load capacitance and source series resistance & inductance..



I think it worths a simulation with real world item values..

[yeeshkul]

Also, what does that PI pad in the output accomplishes (apart from looking really sexy)?
And the 10M parallel resistor on the input?

[Rob Strand]

I'm not sure if Base-Emitter cap prevents oscillations or not 'cause as it's placed it actually "helps" to create a Colpitts oscillator with load capacitance and source series resistance & inductance..

I think it worths a simulation with real world item values..

Yes, I realized how matches with the Colpitts oscillator.  However,  a transistor already has an equivalent Cbe cap "inside" so to speak (as per hybrid pi model) and it does oscillate like a Colpitts oscillator when you place a capacitive load on the output.  Well, at least  when the input and output resistances aren't large enough to fend it off - like my old post I linked above.

I did do some sims just not with the Cornish circuit.   When you add the cap across BE it makes Cbe large and moves the circuit away from oscillation.     Of course if you make the load capacitance large again it pushes to circuit back towards oscillation.    So the oscillation depends on the relative values.   The output resistor also helps move the circuit away from oscillation.

I don't have much doubt adding the BE cap makes to circuit handle more capacitive load than without the added BE cap. 

The question to be answered is:   with a 150ohm output resistor is a 1nF BE cap large enough to prevent oscillation with *any* capacitive load.    My guess is probably not.     The next question is how much load capacitance can we handle before it oscillations, maybe that turns out to be reasonably large.

A sim would answer that.  We have to pick a source resistance.  Worst case is 0 ohms.

[FWIW adding resistance to the base is probably the best defense against oscillation.]

This note explains a few ideas in words without too much maths.   It does not present the BE cap method, probably because the other methods achieve better results.
https://www.hifisystemcomponents.com/downloads/articles/Prevent-Emitter-Follower-Oscillation.pdf

Also, what does that PI pad in the output accomplishes (apart from looking really sexy)?
And the 10M parallel resistor on the input?

10M is a de-pop resistor.

Not sure about the two resistors on the output.  You only need one.   The one on the transistor side is preferable but it's splitting hairs.

[Rob Strand]

I did a very basic simulation to ball-park the maximum load capacitance.   Boot-strapping removed.
Oscillation determined by looking at input impedance, not doing start-up pulses and transient analysis like my previous sims.

Worst case: Source resistance 0 ohms.   Circuit's 1k base left in circuit.

No BE cap:     maximum load capacitance  4pF
1n BE cap:      maximum load capacitance 11pF
1n BE cap, increase input R to 2k, increase output R to 1k:  maximum load 16pF

Clearly the BE cap helps.
Increasing the input and output resistances helps but not as much as I hoped.

As a reality check I need to see if this method of detecting possibile oscillations matches the previous method.

Very interesting:
Input 1k
Output 150 ohm
No BE cap
Add a 10pF cap from the base to ground
Max load capacitance 100 pF ----- enormous improvement!
[but there's  small catch because we now need to add a resistance in series with the 10pF cap]

[yeeshkul]

Max load capacitance 100 pF ----- enormous improvement!

Sorry for a noob question. You can induce oscillation just by connection a cap > 100pF across the output?

[antonis]

You can induce oscillation just by connection a cap > 100pF across the output?

More easily than you can imagine..

Just consider cap's impedance (capacitive reactance) at high frequencies..
(it actually shorts any signal source driving it..)