Direct Coupling with Zener Diodes

[jmaf]

There are discrepancies all over this thread

[J M Fahey]

which will have pulled its grid voltage down. That would explain the discrepency.

Agree.
Considering a 100K plate resistor and a 1M multimeter input impedance, if measuring from plate to +B , some of the actual plate current will go through it, substracting from what passes through the resistor, which will show *less* voltage than the actual one, so you will measure somewhat high (somewhat less than 10% error).
Vice-versa, if you measure from plate to ground, the opposite will happen and you will read a somewhat lower voltage , so the apparent *measured* grid to cathode voltage will look higher.
Some tube schematics carry the "do not measure" label close to some grids, or they simply state no chart voltages there, implying the same.
I said less than 10% because the triode itself is not a simple resistor but has a complex voltage to current curve itself.
Congratulations on your excellent work.

[woodyc]

This occured to me too after it was posted. I *thought* that when I measured the voltages I measured the voltage across the previous anode resistor and subtracted it from the HT. However, I may have accidentally measured the CF grid voltage directly, which will have pulled its grid voltage down. That would explain the discrepency. In reality I suspect the grid is about 0.5V negative of the cathode.

Hey Merlin, I don't know your book, but I really admire the candor and humility you demonstrated in not trying to bs your way out of a mistake.

By the way, the Rutt paper is interesting, but his data does not support his claim that the CF is a preferred way of producing distortion.

It seems to me that however the second triode is set up (as a follower, a cap coupled gain stage, or whatever) its going to contribute the same grid limiting effects. And the second stage is naturally going to be more effective than the input stage since the driving impedance is lower and the signal level is higher.

If you have two triodes available, my feeling is that two high gain stages are going to beat a gain stage and a CF and will provide more options for tailoring the harmonic content and timbre as well.

[kg]

It seems to me that however the second triode is set up (as a follower, a cap coupled gain stage, or whatever) its going to contribute the same grid limiting effects.

how would a traditional RC coupled common cathode stage contribute any grid current?

assuming it cannot, how would you say that the same soft limiting/compression effect on positive peaks could be achieved?

If you have two triodes available, my feeling is that two high gain stages are going to beat a gain stage and a CF and will provide more options for tailoring the harmonic content and timbre as well.

thousands of marshalls/soldanos say otherwise.

look, nobody wants to "waste" a potential gain stage by hooking it up to produce no gain.

the fact of the matter is that it is a different topology, with different characteristics, and different pros and cons.

even rather superficially looking at a simple direct coupled two stage CC/CF arrangement, steven has shown some interesting even order distortion artifacts. this doesn't even take into account the very different effect on the REST of the amplifier from having a driving impedance 100 times lower than before!

if you've only got 1 or 2 stages to play with, by all means, hook it up common cathode and get some drive out of it... you'll need it! but if you have a free stage, make no mistake: the direct coupled CC/CF can and will respond in ways that different topologies cannot.

if they didn't, you can believe me when i tell you the bean counters absolutely would have already cut it out of the design of hundreds of thousands of amplifiers.

[Steve Conner]

Yes! I used to think the same as what woodyc said above, until I did this experiment. Now I think the CF has some subtle mojo, and it may turn out to have a different tone flavour: I'd guess smoother and more modern-sustainy, if that is even a word.

Hopefully I can make my next amp build something along the lines of a 5F6A, Marshall or SLO, and learn a bit more.

When you cascade two triode gain stages, to be sure the second stage draws grid current and clips, but the output from the second stage plate can't really be compared with the output of a CF's cathode, it is quite different.

[kg]

When you cascade two triode gain stages, to be sure the second stage draws grid current and clips,

well, sure it STARTS to draw grid current, but how much actually flows? i mean, only 400 uA of current flowed in your extreme excursion case where the follower was being asked to run at Vak = 0

perhaps more importantly, where does it go?

[Merlinb]

If you have two triodes available, my feeling is that two high gain stages are going to beat a gain stage and a CF and will provide more options for tailoring the harmonic content and timbre as well.

That is true to an extent; certainly the opportunity for frequency shaping is greater with two gain stages. However, the GS>CF combination has the advantage of being able to produce a fair degree of two-stage overdrive with less noise and complete immunity to blocking distortion, when compared with two ordinary cascaded gain stages.

[raiken]

This occured to me too after it was posted. I *thought* that when I measured the voltages I measured the voltage across the previous anode resistor and subtracted it from the HT. However, I may have accidentally measured the CF grid voltage directly, which will have pulled its grid voltage down. That would explain the discrepency. In reality I suspect the grid is about 0.5V negative of the cathode.

Thank you! I was wondering how you got that -10V measurement in your otherwise excellent book, but I just figured you were smarter than me.

I always measure the grid voltage with respect to the cathode, this gives a more accurate representation, because the voltage differential is much smaller. I just measured the cathode follower in my 72 50W Marshall head that I have on the bench right now, and I got -0.35V between grid and cathode.

This is also especially true for long-tailed pair phase inverters. For example, I have another amp on the bench right now that measures +37.8V on the cathode and +23V on the grid, with respect to ground, using my Fluke 87V meter. However, if I put the negative probe on the cathode and the positive probe on the grid, I read the correct voltage difference: -1.08V.

And on the subject of direct coupling with zener diodes, this was how we designed discrete video feedback amplifiers back in the early 80's, when good video opamps with acceptable diff gain and phase specs weren't available. Video amplifiers have to be DC coupled, otherwise you lose the back-porch reference level, and you get compression on peaks when the average picture level shifts (the broadcast video equivalent of "blocking distortion" ). We would usually design a discrete 2-transistor, high bandwidth feedback amp, with an appropriate-valued zener coupling from the first transistor collector to the second transistor base to level-shift the signal back down to maintain the desired DC level at the output.

In the old days, they level-shifted with 0A2's.

Randall Aiken

[woodyc]

how would a traditional RC coupled common cathode stage contribute any grid current?...

Maybe I don't understand the question, but you get grid limiting regardless of how the stage is set up. Thats why Rutt didn't bother actually measuring the grid current in the CF, but used the stage in figure 5.

Everybody knows that the CF is present to provide a buffer or low impedance drive for a stack or effects loop et cetera (or because the design was copied from an amp where it was used that way) -- not because it has some distortion mojo.

Of course it adds distortion, but my feeling is that whatever it adds pales when you consider the other stuff you could do with the triode.

[kg]

Maybe I don't understand the question, but you get grid limiting regardless of how the stage is set up. Thats why Rutt didn't bother actually measuring the grid current in the CF, but used the stage in figure 5.

earlier you stated: "It seems to me that however the second triode is set up (as a follower, a cap coupled gain stage, or whatever) its going to contribute the same grid limiting effects."

and my answer is no, they do not have the same grid limiting effects.

why not?

because the circuit impedances are vastly different.

keep in mind some of the "specialness" of what we're talking about (or what i'm talking about at least) depends on direct coupling, as that is the only way to get substantial current flow from stage to stage. hence my question,

"how would a traditional RC coupled common cathode stage contribute any grid current?..."

should have read:

"how would a traditional RC coupled common cathode stage contribute any substantial grid current to the preceding stage's plate (or output) circuit?..."

... since that is what had really been discussed.

Everybody knows that the CF is present to provide a buffer or low impedance drive for a stack or effects loop et cetera (or because the design was copied from an amp where it was used that way) -- not because it has some distortion mojo.

whoah there.

yes, CFs have TYPICALLY been employed as impedance tranformers.

however, there IS some mojo going on with a direct coupled cathode follower: a positively biased control grid has a variable exponentially increasing conductance. that behavior is NOT something SIMPLY replicated with two RC coupled common cathode gain stages.

whether or not a typical implementation of a CF was for the original purpose of lowering source impedance is irrelevant to a discussion on whether or not they ALSO may have tonal characteristics.

Of course it adds distortion, but my feeling is that whatever it adds pales when you consider the other stuff you could do with the triode.

well, that's the beauty of DIY circuits... you're free to feel however you'd like, and design your circuit accordingly.

speaking for myself, i've ALWAYS had at least one of these stages in my preamps. my experience is more than a "feeling"--every time i tried to get rid of the "waste" of a stage, the tone and response suffered.

i suggest that anyone still reading this thread that is interested in what a DC CF "sounds" like, simply take the few hours to strap one up and try it out themselves. make sure to try a variety of cathode resistors on the CF, and possibly some different tube types. also be sure to generate sufficient signal swing on the input to actually get it to do some clipping.

[woodyc]

earlier you stated: "It seems to me that however the second triode is set up (as a follower, a cap coupled gain stage, or whatever) its going to contribute the same grid limiting effects."

and my answer is no, they do not have the same grid limiting effects.

why not?

because the circuit impedances are vastly different.

keep in mind some of the "specialness" of what we're talking about (or what i'm talking about at least) depends on direct coupling, as that is the only way to get substantial current flow from stage to stage. hence my question,
...

Saying it was "the same" was a really bad choice of words. My point was that you get SOME form of grid limiting in either case.

Think about it this way... If you are in the situation Rutt analyzes, namely you have two stages and no need of the low impedance output of the CF, and your purpose is only to produce tubey distortion, are the little effects (or "vastly different" or "substantial" in your terminlogy) of direct coupling really worth chucking 35dB of gain?

[kg]

Saying it was "the same" was a really bad choice of words. My point was that you get SOME form of grid limiting in either case.

first, please define "grid limiting."

second, do you find NO difference in #2 grid node voltage during overdrive vis a vis the two topologies?

Think about it this way... If you are in the situation Rutt analyzes, namely you have two stages and no need of the low impedance output of the CF, and your purpose is only to produce tubey distortion, are the little effects (or "vastly different" or "substantial" in your terminlogy) of direct coupling really worth chucking 35dB of gain?

obviously it is not worth it to you. i have no interest in convincing you otherwise.

as for me:

1) i have multiple tube sockets unused on my preamp chassis, so yes.
2) i have a huge box of spare, lightly used preamp tubes, so yes.
3) i have more than enough heater current to run any compliment i want, so yes.
4) i have more than enough b+ capacity to run any compliment i want, so yes.
5) i have more than enough chassis space to rig up supporting components, so yes.
6) i have tried to emulate the "dc cf" waveform distortion appearance (see steve conner's post for examples) without using a dc cf and it was not fruitful, so yes.
7) i have tried to emulate the "dc cf" sound (see marshall, mesa, bogner, soldano, VHT, etc for examples) without using a dc cf and it was not fruitful, so yes.

[J M Fahey]

Everybody knows that the CF is present to provide a buffer or low impedance drive for a stack or effects loop et cetera

Well, that sure was the original , and visible reason.

(or because the design was copied from an amp where it was used that way) --

Won't even comment on this.

not because it has some distortion mojo.

Ah, but it does !!
See for yourself: take an isolated cathode follower, +300V on its plate, +160V on its grid, 47K or 100K cathode resistor, load it with a 100K resistor through a .022uF cap and drive it with a perfect 100V RMS sinewave.
Scope the output.
Now to get things even funkier, replace the "easy" load I suggested with a typical tone control network.
Scope the output.
As a comparison, take a regular common cathode gain stage, more precisely the one driving the CF, and drive it with humble 2 to 4 V RMS, enough for it to *try* to produce 100V RMS on its plate.
Scope the output.
Any differences?
I bet so.
I think adding a CF is worth it; I "lose" nothing because I add another triode and problem solved.
Now, if the name of the game is "design the cheapest possible amplifier", I'd rather go the SS route and call it a day.
Good luck.

[Chuck H]

Ok... So the difference in the clipped waveform is clear. And the only way to get it is with a cathode follower. This seems to be the appeal.?. How does it sound better or different? What character does it impart in descriptive terms? Assuming we can't all look at a scope trace of a dummy loaded amplifier driven by a signal generator and know what it means adjectively to the tone.

[pdf64]

The Marshall 6100 design really favoured the gain stage / cathode follower arrangement; so much that there are 3 cascaded in the lead channel, and that's following the inital gain stage; yikes!
http://www.webphix.com/schematic%20h..._6100_100w.pdf
A friend had one of these but I only ever heard it in a domestic setting, not loud enough to overcome the low volume fizz.
I think it's fairly unique and hasn't been subject to DIY cloning, so, regarding the GS/CF, perhaps a little is enough of a good thing? Pete.