Fixed Bias LTP with feedback

[nickb]

Ok, thanks, I just thought C18 was needed for filtering the HT like in Merlin's schem, but of course I see now what you mean with the NFB being disabled.

Is grounding the 220K really much different than my original scheme?

The difference should be negligible. It's just easier to understand IMHO. Call me picky if you like.

[Gaz]

Thanks for the clarification. So as it is, if I delete C18, then I'm not really filtering the HT at all with C19, and it only needs to be big enough to bypass the 2nd grid (around the typical 100n), correct?

I'm starting to wonder if the fixed bias is worth the trouble... Have any opinions? I think I'll have to try a 'regular' cathode biased LTP and see if I can measure any significant differences.

[Gaz]

Ok, I tried wiring it up as a cathode biased LTP (470R bias resistor gave the same bias point), and I could not measure or hear any difference. PI clips at the same input level, and same swing to the output tube grids. I think maybe because the fixed bias is tapped of the HT (which fluctuates with signal level), opposed to being regulated, there is not practical difference with cathode bias. I don't believe either Hiwatt arrangement regulates the bias voltage at all either.

[Helmholtz]

As already indicated by pdf64 (post #9), a fixed grid potential doesn't mean fixed bias as long as cathode potential is self-adjusting.
In fact fixed grid potential (typically 0V) is common to most cathode biased circuits.

[Chuck H]

Ok, I tried wiring it up as a cathode biased LTP (470R bias resistor gave the same bias point), and I could not measure or hear any difference. PI clips at the same input level, and same swing to the output tube grids. I think maybe because the fixed bias is tapped of the HT (which fluctuates with signal level), opposed to being regulated, there is not practical difference with cathode bias. I don't believe either Hiwatt arrangement regulates the bias voltage at all either.

And while it's an interesting thought experiment, how much advantage is there really even if it were regulated? For a guitar amp, especially if you're going to be clipping it, other larger voltage fluctuations would easily render any difference inaudible I think. The PI operates it's triodes in class A so ideally there is no bias fluctuation at the cathode other than those caused by voltage sag in other circuits. I'm sure that in reality and once clipping there may be a little. But probably not much.

But it's cool to have a thread that actually explores the difference on a bench and with a pair of ears because the consideration has come up before.

[Helmholtz]

I think that (apart from type and positioning of tone controls) the probably most sound shaping stage in many classic amps is the DC coupled cathode follower pair (could be considered a tube overdrive unit). It's a pity it can't be realized here.

[Chuck H]

I think that the probably most sound shaping stage in many classic amps is the DC coupled cathode follower pair. It's a pity it can't be realized here.

It does a lot of interesting things depending on how it's implemented. But that's not to say that you can improve any specific tonal quality by adding one or that a great amp can't be built without one. There has been a lot of discussion about the clipping properties of the CF, but something the driver/CF circuit does that gets overlooked is to isolate the driving stage from the following triode amplifier. That is, the driving stage IS a gain stage and it operates unencumbered into the CF. Whereas if you operated it into another triode gain stage at clipping levels that following triode influences the output of the driving stage a lot more than the CF would.

[Helmholtz]

IS a gain stage and it operates unencumbered into the CF.

Maybe I misunderstand, but I think "unencumbered" (had to look this up) is misleading. The point is that the cathode follower is typically operating at the edge of grid conduction which makes it look like a diode load to the driving gain stage before it.

Apart from that my comment above was meant with regard to reproducing the original Hiwatt amp's sound.

[Chuck H]

Maybe I misunderstand, but I think "unencumbered" (had to look this up) is misleading. The point is that the cathode follower is typically operating at the edge of grid conduction which makes it look like a diode load to the driving gain stage before it.

Apart from that my comment above was meant with regard to reproducing the original Hiwatt amp's sound.

Ah. Ok. I never knew exactly why the CF clips like it does before. Now I do and it makes sense. IMHE though the CF doesn't clip much. It adds it's own color to the tone, but doesn't clip as much as a typical gain stage would with the same input. What I was getting at was how the stage driving the CF isn't clipped off at it's output the same way it would be driving a typical gain stage. And then with the CF driving that following stage, with it's relatively low impedance, you end up with something very different from what you would get if you skipped using the CF. I'm sorry. My lack of tech is keeping me from explaining what I see on the bench well.

[Gaz]

The reality in the Hiwatt circuit the DC coupled doesn't matter much. Before adding an extra switchable gain stage to this amp (which is how I lost the extra triode), I've actually A/B'd both and can't hear the difference. I figure the reason is because it comes after the lossy tonestack and master volume. It's difficult to hit the CF hard enough in order for it to impart it's tonal quality unless the master is very high. Even then, I can't hear it because the power tubes are already clipping and the room is shaking.

Also, Hiwatt ditched DC coupled CF themselves and nobody seemed to notice. I've played around with this circuit a lot, and I can tell you you are correct that the tone control give Hiwatt's most of their sonic signature. The other majority would be their Fane speakers if you can find them. After that probably the 47n bypass caps on certain models and then maybe large amount of NFB and interesting Presence control. JM2C.

[pdf64]

I think the real (but non obvious) benefit of a DCCF is all about the way it shapes the waveform http://www.valvewizard.co.uk/dccf.html

[mhuss]

Another point to consider - the traditional LTP gets is DC grid bias from a tap on the cathode R string. This actually provides a bit of positive feedback (as the cathode current increases, so does the cathode voltage and therefore grid voltage.)

The Hiwatt PI keeps the DCV on the PI grid pretty much fixed (particularly since it's a low driving impedance), so there's only the NFB from the cathode current/voltage relationship affecting the PI.

This is still sort of cathode bias, but with zero/low grid feedback. The operating point is more precisely controlled, though.

[Helmholtz]

This actually provides a bit of positive feedback (as the cathode current increases, so does the cathode voltage and therefore grid voltage.)

...also known as bootstrapping .

[mhuss]

...also known as bootstrapping .

True, but I was primarily thinking in DC terms. And not that a triode stage really needs to increase it's input impedance.

[Helmholtz]

The small positive feedback provided by the bootstrapped grid resistor opposes the negative feedback/degeneration by the cathode resistor. So the Marshall type LTPI bias should behave a little more "fixed" than the Hiwatt type with a fixed grid potential.