Anyone stupid enough to use an LTP or concertina or see saw or just about any phase splitter other than a transformer to directly feed output tubes drawing grid current will find quickly how bad an idea that is. I haven't met a person that stupid yet, though perhaps one of you has. So the latest suggestion of unbalancing the loads and trying to pull an "Aha!" by violating the equal-loads constraint strikes me as downright silly. My model still gives accurate results for equal loads, and if Dave Slagle had used it, he wouldn't be living in the back of a station wagon trying to feed his family from leftover Alpo cans.
Rich- Morgan was wrong about class AB1/B1 and the concertina in the third edition of Valve Amplifiers. He corrected this in the 4th edition.
Rich- Morgan was wrong about class AB1/B1 and the concertina in the third edition of Valve Amplifiers. He corrected this in the 4th edition.
SY, who suggested "unbalancing the loads"?
So no one who uses a Concertina or LTP will ever jack up the volume to the point where grid current is drawn?
Or perhaps everyone will starve the design so it's impossible to get the most possible drive out of the output stage, preferring instead to saturate the Concertina?
As for poor Dave, the whole point was that he got into the trouble he did because he used your model. He actually expected the Concertina to drive each of his LCR modules with a 6K source impedance - imagine that! In fact, he demanded it! And when his consultant gave him what he demanded, well, look what happened.
So no one who uses a Concertina or LTP will ever jack up the volume to the point where grid current is drawn?
Or perhaps everyone will starve the design so it's impossible to get the most possible drive out of the output stage, preferring instead to saturate the Concertina?
As for poor Dave, the whole point was that he got into the trouble he did because he used your model. He actually expected the Concertina to drive each of his LCR modules with a 6K source impedance - imagine that! In fact, he demanded it! And when his consultant gave him what he demanded, well, look what happened.
You did. Several times.
If you're clipping the output stage, the unbalance in the phase splitter (whether cathodyne, LTP, paraphase, cross-coupled, whatever) will be the least of your worries. And Dave is still trying to convince the kids that Alpo is just a tangy Salisbury Steak. If he had only listened to me, poor guy.
If you're clipping the output stage, the unbalance in the phase splitter (whether cathodyne, LTP, paraphase, cross-coupled, whatever) will be the least of your worries. And Dave is still trying to convince the kids that Alpo is just a tangy Salisbury Steak. If he had only listened to me, poor guy.
To recognize that in the normal way the amplifier is used, that “unbalance happens” hardly constitutes an intent of “unbalancing the loads”. Ignoring the phenomenon won’t make it go away. Better to accept the undesirable and understand it.
And oh, but Dave did listen to you! That's the whole point.
Does your model or does it not include two sources?
Did Dave’s design or did it not involve two matched loads?
Of course, yes to both! He then took the next logic step from the faulty model (there was no turning back) and demanded a 6K drive for each LCR module.
Poor Dave. He was just following your model.
And oh, but Dave did listen to you! That's the whole point.
Does your model or does it not include two sources?
Did Dave’s design or did it not involve two matched loads?
Of course, yes to both! He then took the next logic step from the faulty model (there was no turning back) and demanded a 6K drive for each LCR module.
Poor Dave. He was just following your model.
1/Mu attenuated cathode feedback easily provides for equal impedance.
http://www.diyaudio.com/forums/tubes-valves/189153-phase-splitter-issue-82.html#post3153990
Yet matched Cathodyne has exactly same A2 problem as ordinary Cathodyne.
These A2 problems have nothing whatsoever to do with impedance mismatch.
Caps coupled into diodes abscond with our Cathodyne's proper DC bias...
Argument against impedance match based upon example of grid conduction,
abuses a topology that is flawed, even if impedance were assumed equal.
The impedance are indeed unequal, but the A2 proof is broken and irrelevant.
In any circuit that works, the mismatch (and the fix for it) are irrelevant.
http://www.diyaudio.com/forums/tubes-valves/189153-phase-splitter-issue-82.html#post3153990
Yet matched Cathodyne has exactly same A2 problem as ordinary Cathodyne.
These A2 problems have nothing whatsoever to do with impedance mismatch.
Caps coupled into diodes abscond with our Cathodyne's proper DC bias...
Argument against impedance match based upon example of grid conduction,
abuses a topology that is flawed, even if impedance were assumed equal.
The impedance are indeed unequal, but the A2 proof is broken and irrelevant.
In any circuit that works, the mismatch (and the fix for it) are irrelevant.
Last edited:
No, he didn't. If he had used identical 5k5 (give or take a couple hundred ohms) series resistors from each node as my experimentally verified model suggests, he'd still be feeding his children people food.
Agreed, neither the LTP nor the Cdyne nor the balanced Cdyne solves the DC currrent pumping problem. None is better than another against this.
Flawed topology, yes. But it is employed. And when it is, balanced impedance drivers (LTP) and unbalanced (Cdyne) behave differently. Just... differently. Because of the impedance balance or unbalance.
Flawed topology, yes. But it is employed. And when it is, balanced impedance drivers (LTP) and unbalanced (Cdyne) behave differently. Just... differently. Because of the impedance balance or unbalance.
Well, SY, I'm not going to "Yes he did" you.
Your model does not "say" two equal 5K5 series impedances.
Your model says there are two separate sources connected only by ground. Once that is accepted, the only way to make things work is to demand that each provide a 6K drive impedance. The only way to verify that that is done is to test it. When tested, equal series resistors did not satisfy the requirement of 6K source impedances for each LCR module.
Even though we can both agree that that was what he needed, the logical step-by-step application - and test -of your model told him otherwise.
Your model does not "say" two equal 5K5 series impedances.
Your model says there are two separate sources connected only by ground. Once that is accepted, the only way to make things work is to demand that each provide a 6K drive impedance. The only way to verify that that is done is to test it. When tested, equal series resistors did not satisfy the requirement of 6K source impedances for each LCR module.
Even though we can both agree that that was what he needed, the logical step-by-step application - and test -of your model told him otherwise.
Last edited:
Balanced cDyn and LTP of otherwise equal impedance behave differently
because offsets of DC pumping integrate asymmetrically upon cathodyne.
Because of topology in the matched case, not because of impedance...
Primarily because of topolgy, even in the obviously unmatched case...
Last edited:
Well, yes, it does. That's exactly what it predicts, assuming 1/gm is 500R.
Tony, transformer. Otherwise, LTP with HV tubes and a high rail.
edit: You can also bootstrap the driver, a la McIntosh.
Every test you have proposed to prove unequal plate and cathode impedances has unbalanced the loads. I do not accept braking the model to prove it wrong acceptable.
In the case of the dual LCR's the only correct answer is ~5800 ohms in both the plate and cathode...
I agree in this case there will be no current flowing through the ground node. I will also accept that we need to call this the differential output impedance.
I think the key here is to draw the current loops and treat each loop impedance as it should. Currents common to the plate and cathode must have the differential output impedance. Currents common to just the plate or cathode loop (through ground) will have the respective plate or cathode impedance.
The whole conflict here is can ground still be a reference even if no current flows through it? I say yes since it was the only way to solve the problem that would have kept my job. (on the plus side cat food over rice tastes much better than kibble)
dave
......... i am doing screen grid drive in 2013...
I maintain what I said in post 850. I wish you address it.
Your circuit which purportedly tests P & K to ground impedances by simultaneously shorting plate and cathode to ground draws no ground current. So there can be no current through ground referenced impedances. So they can't possibly be tested by your circuit.
When you test them properly, you find that the ground-referenced P and K impedances are different.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.