Hmm, will it?
When the right tube is run in pentode mode it will have a very high Rp compared to the left one which is acting more like a triode with low(ish) Rp.
I have difficulty understanding all aspects of this output stage topology, but I have tried to read up on this earlier thread on the subject: Beyond SET and ParaFeed.... Complementary Current Triode started by our smoking-amp Don.
SveinB.
But when the right tube is current-slaved to the triode, the Zo of the stage
is 1/2 that of the triode alone (and the Po is 2X), assuming equal operating points.
IOW, the slave tube is forced to provide equal signal current swing to the
master tube, and the feedback is from the output (both tubes) to the grid
of the master tube (triode or Schadeode, doesn't matter).
I have validated it through measurement of prototype amplifiers also.
Cheers,
Michael
The currents are equal and opposite for previously described reason.
The voltages are equal and opposite (even though pentode doesn't
have Mu of its own) because of the transformer coupling...
Therefore the power assist from the Pentode is equal to the Triode.
And if one considers "plate resistance" as the slope of a VI curve,
if forced to be exactly equal (in quiescent state) and opposite of
the controlling triode when so slaved.. Doesn't matter how was
strapped to make it act that way, only the end effect. Its now an
Anti-Triode.
The voltages are equal and opposite (even though pentode doesn't
have Mu of its own) because of the transformer coupling...
Therefore the power assist from the Pentode is equal to the Triode.
And if one considers "plate resistance" as the slope of a VI curve,
if forced to be exactly equal (in quiescent state) and opposite of
the controlling triode when so slaved.. Doesn't matter how was
strapped to make it act that way, only the end effect. Its now an
Anti-Triode.
Last edited:
I did a few mods yesterday. Firstly i took out cap C5, which just confirmed that that was a bad idea as now it has quite a bit of hum. It has at least confirmed that the +B could do with a bit more filtering which I will try.
I then inserted a 18K resistor between the 6AU6 plate and the feedback resistor and coupling cap. I also increased the feedback resistor to 82K. All hint of the harshness has gone now so it was obviously a problem of overloaded driver.
My next steps are to try reversing the polarity of the OT to see if that cleans up the sound a tad. I have decided the easiest way to deal with the excessive gain is to voltage divide down the input - which I will do today.
I like the sound, clean and detailed with good bass extension. I will do another test up against my benchmark Tabor clone.
shoog
I then inserted a 18K resistor between the 6AU6 plate and the feedback resistor and coupling cap. I also increased the feedback resistor to 82K. All hint of the harshness has gone now so it was obviously a problem of overloaded driver.
My next steps are to try reversing the polarity of the OT to see if that cleans up the sound a tad. I have decided the easiest way to deal with the excessive gain is to voltage divide down the input - which I will do today.
I like the sound, clean and detailed with good bass extension. I will do another test up against my benchmark Tabor clone.
shoog
Interesting. I suppose the gm nonlinearities of the 2 similar tubes cancel each
other out through the common cathode signal voltage.
Lars, can I get a copy of your .asc to play with this?
Thanks!
Michael
Thanks. I wasn't sure if you were originally referring to a triode LF topology generating high levels of IM distortion. I hadn't seen that with my 6SN7/801 version. I see now that it was applied to something different.
Sheldon
Its a great idea, but not by itself. Won't do any good until both sides float
the same hum. If grid of pentode on the right remains ground ref'd, hum will
only be aggrivated, exactly as you observed.
If you re-utilize your C5, to bypass G1 of rightmost pentode up to B+ hum...
Now you got a situation where hum can be fully balanced and cancelled.
I may try that, as there is a slight amount of residual clean power supply hum left. At this stage I have things working well and am reluctant to mess about to much as this usually cascades into multiple other problems (you should see it in there - not a lot of room).
I will draw up a revised schematic for completeness sake. After that anyone who wants to place improved versions for discussion would be more than welcome. At the moment the design is still elegantly simple and performing well so I am happy.
Shoog
I will draw up a revised schematic for completeness sake. After that anyone who wants to place improved versions for discussion would be more than welcome. At the moment the design is still elegantly simple and performing well so I am happy.
Shoog
Well, that was unexpected, but makes perfect sense...
Easy enough to cascode a fet under the anti-cathode to bump up Gm.
What else would you expect if you feedback the sum from both toobs on an output transformer?
Yeah, about what I thought. LW's ultrapath dominates over the cathode to
cathode Pfeedback... Can't cancel out input cathode resistor's Nfeedback
this way. Not much different than sampling hum directly from the top rail...
Gonna be even more the case here... With anti-triode as an ultrapath cap
multiplier. The node between output cathodes should closely follow the high
rail hum in terms of a voltage swing. And the input signal in terms of current
swing.
Anyways, CHB shouldn't be necessary once all real triodes are out of the
picture... High plate impedance flips the hum to the high rail, and no need
to do an injection at the cathode.
Not sure if you're talking about a Schadeode here, but if you are, won't the
plate-grid feedback impose equivalent AC current from B+ ripple as the low Ri
of a real triode would?
Remember, your plate to plate Schade network is folding back to the top rail
(assuming C5 wasn't there). And thats why we want to let the entire output
stage float on that same hum...
The output stage cathodes sitting on CCS won't mind...
And its easy enough to bypass the Anti-Triode's grid to that same B+ noise.
To be clear, the Loftin White PSRR treatment only applies if the input stage is
a Triode with low plate impedance. I am suggesting it is more trouble than its
worth to fix the Triode's PSRR probs. A pentode simply works better here.
(assuming C5 wasn't there). And thats why we want to let the entire output
stage float on that same hum...
The output stage cathodes sitting on CCS won't mind...
And its easy enough to bypass the Anti-Triode's grid to that same B+ noise.
To be clear, the Loftin White PSRR treatment only applies if the input stage is
a Triode with low plate impedance. I am suggesting it is more trouble than its
worth to fix the Triode's PSRR probs. A pentode simply works better here.
Last edited:
Re: Lars:
"Hey Don,
Did a sim using two 6GW8 models vs. one 6GW8 and one DN2540 and can confirm this. The SE signature is lost using a Schaded pentode together with a pentode. But it still acts like a PP triode."
Very interesting. Did the schaded pentode with equal gm pentode model show any even harmonics at all? After Michael mentioned the triode/pentode dissimilarity of effective gm (plate gm cancelling some grid 1 gm), I've been thinking there would be some SE signature, but not a strong SE signature that way. I'm away from my usual computer, so I can't do any sims here.
I think Shoog's approach will soon have us all building models.
Don
"Hey Don,
Did a sim using two 6GW8 models vs. one 6GW8 and one DN2540 and can confirm this. The SE signature is lost using a Schaded pentode together with a pentode. But it still acts like a PP triode."
Very interesting. Did the schaded pentode with equal gm pentode model show any even harmonics at all? After Michael mentioned the triode/pentode dissimilarity of effective gm (plate gm cancelling some grid 1 gm), I've been thinking there would be some SE signature, but not a strong SE signature that way. I'm away from my usual computer, so I can't do any sims here.
I think Shoog's approach will soon have us all building models.
Don
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.