Dave, you have said in post 1132,
I have asked you what this method is and how it can be applied to circuits in general. Your reply was to state that I was erecting a “false dichotomy”. I erected nothing - it’s your method. And your lack of a clarification of it confirms my belief that it doesn’t exist.
Of course. I have never implied otherwise.
Circuits consider our intentions and their engineering contexts and reply, “Bwaa-ha-ha-ha!” The Cathodyne cannot escape the fact that supply noise is an SE signal and that it responds to that signal. Knowing that response demands knowledge of SE impedances.
I think you meant “unbalanced signal.”
Of course. I have never implied otherwise. Whereas you, in post 1138…
I agree completely with your post 1139. I will add that people can (and will) argue about impedances. But impedances are not a matter of opinion – they are matters of fact and have singular, not multiple, values in a given situation. These are easily demonstrated in a number of ways using well-established, proven and time-tested techniques which do not grant special dispensation to Cathodynes.
I have asked you what this method is and how it can be applied to circuits in general. Your reply was to state that I was erecting a “false dichotomy”. I erected nothing - it’s your method. And your lack of a clarification of it confirms my belief that it doesn’t exist.
Of course. I have never implied otherwise.
Circuits consider our intentions and their engineering contexts and reply, “Bwaa-ha-ha-ha!” The Cathodyne cannot escape the fact that supply noise is an SE signal and that it responds to that signal. Knowing that response demands knowledge of SE impedances.
I think you meant “unbalanced signal.”
Of course. I have never implied otherwise. Whereas you, in post 1138…
I agree completely with your post 1139. I will add that people can (and will) argue about impedances. But impedances are not a matter of opinion – they are matters of fact and have singular, not multiple, values in a given situation. These are easily demonstrated in a number of ways using well-established, proven and time-tested techniques which do not grant special dispensation to Cathodynes.
I think the false dichotomy was directed at me not claimed by me....
OK... a traditional method for measuring output Z is to measure the OC voltage and then apply a load equal to the output impedance and you will get 1/2 the OC voltage.
I am simply saying that in the case of a cathodyne when you want to measure the differential output z wrt ground then you need to apply the load to both the anode and the cathode to do your measurements otherwise you are measuring the single ended impedance.
I know thevenin will allow you to choose the plate and cathode as your two points and give my expected results, however i do not know if it is "legal" to manipulate the Zkg when you are measuring the Zag to get results that match 1/2Zak (i know i am swapping between c and K for cathode but Zac looks wrong)
yes.
I think we both agree that the SE and balanced behavior of the split load need to be treated differently.
The word "here" in post 1136 is a link to a full schematic of a Fisher X-101-B. Is clicking on that out of the question for some reason?
Dave, OK, we'll drop the "false dichotomy" thing since it didn't come from you.
The traditional Zout measurement is good - for one pair of nodes at a time, not two different pairs simultaneously, agreed?
I can measure a differential Zout between the plate and cathode. I connect a load between them to do so. I don't know how to do this w.r.t. ground. That's the three node problem again - impedance measurements involve two and only two nodes. Now, you can measure anything you want, but something involving three nodes cannot be a measurement of what impedance is understood to be.
The Cathodyne provides an interesting coda to this discussion. Since the plate and cathode have equal and opposite voltages, if you sense the AC voltage at the middle of the test load, you would find it to be zero. So if you AC coupled this point directly to ground, nothing, neither voltage nor current, would change in this circuit, and you would obtain the same results as a standard, two node impedance test with no test load path to ground.
I agree that it is "illegal" to manipulate Zkg if you wish to measure Zag. I don't understand the part about "get results that match 1/2Zak".
Yup, we agree that the balanced and SE behaviors need to be treated differently.
The traditional Zout measurement is good - for one pair of nodes at a time, not two different pairs simultaneously, agreed?
I can measure a differential Zout between the plate and cathode. I connect a load between them to do so. I don't know how to do this w.r.t. ground. That's the three node problem again - impedance measurements involve two and only two nodes. Now, you can measure anything you want, but something involving three nodes cannot be a measurement of what impedance is understood to be.
The Cathodyne provides an interesting coda to this discussion. Since the plate and cathode have equal and opposite voltages, if you sense the AC voltage at the middle of the test load, you would find it to be zero. So if you AC coupled this point directly to ground, nothing, neither voltage nor current, would change in this circuit, and you would obtain the same results as a standard, two node impedance test with no test load path to ground.
I agree that it is "illegal" to manipulate Zkg if you wish to measure Zag. I don't understand the part about "get results that match 1/2Zak".
Yup, we agree that the balanced and SE behaviors need to be treated differently.
The Fischer design is a bit difficult to analyze because of the feedback applied from the signal going through the output stage and transformer.
I disconnected and ignored the feedback and sim'd the circuit otherwise as shown, with the 50K pot centered.
The 330k resistor to the Cathodyne cathode has a negligible, worsening effect on the balance. Suffice it to say that I don't understand its purpose.
So if, as in this case, the gain triode's plate has less than half the supply noise on it, I recommend compensation of the form I mentioned in an earlier post in this thread.
And, contradicting my 4AM "revelation" that I mentioned earlier, a resistor from the supply to the Cathodyne cathode could solve the problem if a pentode forms the gain stage and more than half the supply noise appears on its plate. A caveat is that the compensatory resistor may need to be so small as to render a balanced output signal problematic.
I disconnected and ignored the feedback and sim'd the circuit otherwise as shown, with the 50K pot centered.
The 330k resistor to the Cathodyne cathode has a negligible, worsening effect on the balance. Suffice it to say that I don't understand its purpose.
So if, as in this case, the gain triode's plate has less than half the supply noise on it, I recommend compensation of the form I mentioned in an earlier post in this thread.
And, contradicting my 4AM "revelation" that I mentioned earlier, a resistor from the supply to the Cathodyne cathode could solve the problem if a pentode forms the gain stage and more than half the supply noise appears on its plate. A caveat is that the compensatory resistor may need to be so small as to render a balanced output signal problematic.
Signal balance? Well, sure, but can the pot compensate for the imbalance? Does it improve balance of PS noise so that it can be better cancelled by the output stage? If not, it seems that the Fisher company wasted a good bit of money on 330k resistors that they could have kept in their pocket.
Personally, I own one of these amplifiers and I found that it always had a faint buzz on the output. The cause was the shared cathode resistor at the output stage and associated current imbalance between tubes. I cured that by using four separate bypassed CCSs as cathode resistors on the output tubes.
So even if the 330k resistor helped eliminate some buzz, they still botched the overall design with their "clever" DC heater supply/power tube cathode resistor that made the design dependent on perfectly matched output tubes.
The reason for the 330K is to keep a 50V reference for the balance in case the bias of 100V raises or cathodyne saturates.
It is not necessary to have the 330k but please put a grid stopper resistor if you remove it. I have no clue if it would have a noticeable effect on sound.
On the Fisher schematic the R83 Trim-pot is what I was referring too.
The purpose is to balance perfectly the signals to match the rest of the amplifier if you adjust that trim-pot with a good distortion meter.
(and yes SpreadSpectrum, the output stage bias is unorthodox to say no more)
It is not necessary to have the 330k but please put a grid stopper resistor if you remove it. I have no clue if it would have a noticeable effect on sound.
On the Fisher schematic the R83 Trim-pot is what I was referring too.
The purpose is to balance perfectly the signals to match the rest of the amplifier if you adjust that trim-pot with a good distortion meter.
(and yes SpreadSpectrum, the output stage bias is unorthodox to say no more)
Both the name and the function that I can suss out of R83 lead me to believe that it can compensate for the levels of the desired audio signals presented to unmatched, or matched, output tubes. However, it cannot simultaneously equalize the supply noise in the cathodyne outputs. For that, the best choice is a pot in series with a resistor and a cap connected between the gain stage cathode and the supply. Leave R83 in, go back and forth between tweaking both pots, and you can address both concerns.
But even after gabdx's comment, I don't understand the purpose of the 330k resistor.
But even after gabdx's comment, I don't understand the purpose of the 330k resistor.
Hi
In my opinion, the 330k is used to inject additionnal current in the Rk in order to raise the cathode potential and so the grid potential og the cathodyne.
In this situation, the voltage at the plate preceeding the cathodyne can be higher, allowing more room to choose the polarisation point for this stage.
Jacques
In my opinion, the 330k is used to inject additionnal current in the Rk in order to raise the cathode potential and so the grid potential og the cathodyne.
In this situation, the voltage at the plate preceeding the cathodyne can be higher, allowing more room to choose the polarisation point for this stage.
Jacques
Last edited by a moderator:
Well, years ago I came here and asked what the designer could have possibly been thinking and the best answer I got was maybe ripple injection. Looking it over myself, my gut tells me the ripple injection would be really weak so I question whether it is worth the resistor. Apparently, the designer really liked it though, because they used it in a lot of Fisher amps.
I removed it in my own personal X-101-B and suffered no ill consequences. That was after I rebuilt the power supply on a separate chassis and eliminated pretty much all ripple, though (the reason for starting on that project was the meltdown of the original power transformer and lack of replacements that would fit on the original chassis).
I hear you about your output stage mod. It equalized the bias currents of the output tubes. But their transconductances were not necessarily balanced as a result, right?
It probably improved the gain balance of each side of the output stage and lowered distortion. But unless the power supply noise signals on the cathodyne outputs were the same (which I don't believe they are) it's not clear to me how that would reduce the overall buzz.
It probably improved the gain balance of each side of the output stage and lowered distortion. But unless the power supply noise signals on the cathodyne outputs were the same (which I don't believe they are) it's not clear to me how that would reduce the overall buzz.
I'm guessing that transconductances came more into balance as idle currents came more into balance. Buzz cancellation in the output stage depends on setting the current close to equally in the output tubes. With four tubes sharing a common cathode resistance, this is not going to happen (at least, not in my experience) and results in a tube that starts red-plating as they age and spread out in characteristics.
With individual bias control of output tubes and tightly matched idle currents, the output stage behaves like it should and cancels the buzz that rides on the B+ to the transformer.
Now, that didn't do anything to help any buzz that could have come from previous stages to the grids of the output tubes, but overall it made buzz inaudible in my amp.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.