This is explained in some detail in the Red Light District article and in even greater detail in the Bevois Valley section of Morgan Jones's "Valve Amplifiers."
Walks like a Hippie, smells like a Hippie, looks like a Hippie...I'm pretty OK with calling it a Hippie.
cheers,
Douglas
Correct. The cathode bypass cap mostly doesn't change plate impedance
excepting consequence of 1/(mu-1) portion that leaks through from the
plate... Shunting away or adding other arbitrary cathode currents, that
may or may not be correlated, is irrelevant to plate impedance. Unless
you are going to close a global loop and make them relevant.
Can make a big change in the illusion we are miscalculating as impedance.
And I'm not saying we shouldn't make use of that calcultaion. Recognize it
for what it is, a disconnected cause and effect that looks like impedance.
Last edited:
Ken, it's a hippie. When you can take ANY arbitrary pair of loads and get results consistent with equal source impedances, what you have there is equal source impedances.
you mention plate impedance. Isn't that quite different from output impedance?
If you load a 12AX7 @ 2ma with a CCS and then with a 100K resistor
does the plate impedance stay the same?
does the output impedance stay the same?
dave
dave
I've attached 2 times in this thread a picture bout that: pentode stage, Concertina, LTP, all directly coupled.
I may be missing some context here, but the impedance "looking into" the plate, with respect to ground, in a circuit like the Cathodyne (and absent any external loads) is given as Zpg = rp + (1+u)*Rk. (See p. 404 of Morgan Jones' book, or the oft cited Preisman or George Ellis Jones papers, or the Radiotron Designer's Handbook, 4th edition, p. 330, equations (30) and (31).)
If you capacitively bypass Rk, at high enough frequencies, that term in Zpg disappears, and the impedance becomes Zpg = rp.
If you want to discuss the plate to ground impedance of the circuit as a whole, you parallel the impedance connected to the plate ("looking out" of the plate) with Zpg.
Regarding the question about output impedance vs. impedance, between any pair of nodes in a linear circuit, the impedance = the output impedance = input impedance. Thevenin's theorem proves this.
Thevenin says that for any two nodes in a linear circuit consisting of resistors and independent and controled sources, you can construct an equivalent circuit consisting of a single voltage source in series with a single resistor (Wikipedia actually gives a good treatment of this.)
If you test the "output" impedance of that circuit with the traditional open circuit voltage, short circuit current technique, you will see that it equals that single resistor.
If you test the "input" impedance by setting the single voltage source to zero, it is clear that it is equal to that single resistor.
(By the way, if you are unfamiliar with the terms, a controled source is a device in which a signal applied to its terminals affects its characteristics in a circuit, a triode being a good example. An example of an independent source would be a battery, whose output could be varied only by the combination of its internal resistance and a load placed across it.)
If you capacitively bypass Rk, at high enough frequencies, that term in Zpg disappears, and the impedance becomes Zpg = rp.
If you want to discuss the plate to ground impedance of the circuit as a whole, you parallel the impedance connected to the plate ("looking out" of the plate) with Zpg.
Regarding the question about output impedance vs. impedance, between any pair of nodes in a linear circuit, the impedance = the output impedance = input impedance. Thevenin's theorem proves this.
Thevenin says that for any two nodes in a linear circuit consisting of resistors and independent and controled sources, you can construct an equivalent circuit consisting of a single voltage source in series with a single resistor (Wikipedia actually gives a good treatment of this.)
If you test the "output" impedance of that circuit with the traditional open circuit voltage, short circuit current technique, you will see that it equals that single resistor.
If you test the "input" impedance by setting the single voltage source to zero, it is clear that it is equal to that single resistor.
(By the way, if you are unfamiliar with the terms, a controled source is a device in which a signal applied to its terminals affects its characteristics in a circuit, a triode being a good example. An example of an independent source would be a battery, whose output could be varied only by the combination of its internal resistance and a load placed across it.)
For the real life this whole impedance - related talk is useless. If resistors in anodes and cathodes are equal, if load impedances are equal and linear, phase splitter works as intended and supply loads by signals of an opposite polarity. The problem is in load on non-linear resistance. Since loads of outputs are asymmetrically reflected on each other it creates a problem of non-linearity of the resulting drive.
It is an axiom that the system made of optimal subsystems can not be optimal. That means, for practical engineering that is search for optimal solutions we must optimize the whole system that contains more than Concertina with it's triode and couple of resistors.
It is an axiom that the system made of optimal subsystems can not be optimal. That means, for practical engineering that is search for optimal solutions we must optimize the whole system that contains more than Concertina with it's triode and couple of resistors.
He did, actually. Once, when suggested to load on equal capacitances that are much greater in values than variations of input capacitances of output stage tubes, then when mentioned grid stoppers of relatively high values. Third time when wrote about Power Drive that means buffers between outputs of Concertina and grids of output tubes.
If by this you mean the anode to ground resistance equals the cathode to ground resistance, I disagree, and so does Mr. Vogel in section 2.2 of the MathCad portion of his recent letter to Linear Audio, along those who wrote the references I cited.
However, this has no bearing on the fact that the plate-cathode impedance is a little less than 2/gm, as Mr. Vogel also states.
I agree with this shortened version of your (partial) quote.
I agree with your non-linearity comments.
The increased grid stoppers, within reason, are a good idea. They maintain balance as long as they are equal, and they increase the total resistance, especially in the otherwise low impedance cathode-to-conducting power tube grid path, reducing the unwanted charge left on the grid coupling capacitor when grid current flows.
This is of much greater benefit in the cathode path than in the plate path because of the lower cathode-ground than plate-ground impedance to start with.
Of course, global negative feedback works against these ameliorative effects.
This is of much greater benefit in the cathode path than in the plate path because of the lower cathode-ground than plate-ground impedance to start with.
Of course, global negative feedback works against these ameliorative effects.
...it is not due to impedance, but due to the fact that cathode resistor is in feedback loop, i.e. voltage drop on it is applied to the input signal providing negative feedback by voltage in series with input signal. Anode resistor is outside of this loop. But if you apply input between anode and grid instead of between ground and grid, things change drammatically.
...but it is a different story...
The feedback in the triode reduces cathode-ground impedance and increases plate-ground impedance. Radiotron Designers Handbook has a good treatment of this.
You might consider the difference in performance between a standard Cathodyne and one modified so that the grid signal was referenced to the cathode rather than to ground. In this case, there is no feedback in the Cathodyne - Ec = eg - ek is determined completely by the source signal. In such a case, the plate to ground and cathode to ground impedances truly are identical - there is no feedback to modify the impedances.
You might consider the difference in performance between a standard Cathodyne and one modified so that the grid signal was referenced to the cathode rather than to ground. In this case, there is no feedback in the Cathodyne - Ec = eg - ek is determined completely by the source signal. In such a case, the plate to ground and cathode to ground impedances truly are identical - there is no feedback to modify the impedances.
Yes, if signal referenced to cathode it turns into a common cathode stage with split in half load. If signal is referenced to anode it turns into a cathode follower with split in half load.
Correct.
But as used, it is a cathode follower with one more equal resistor in series. And it is easy to demonstrate that when it is equally loaded it works well as desired, but attempts to "compensate unequal output resistances" leads to wrong results, what SY tried to show in his article. Load it equally, do not worry about different output resistances, and you will be happy.
Correct.
But as used, it is a cathode follower with one more equal resistor in series. And it is easy to demonstrate that when it is equally loaded it works well as desired, but attempts to "compensate unequal output resistances" leads to wrong results, what SY tried to show in his article. Load it equally, do not worry about different output resistances, and you will be happy.
Last edited:
Actually, I think it is well viewed as a two terminal (plate to cathode) signal source, which is why the plate-ground and cathode-ground impedances are equal in this case.
But not with a standard Cathodyne.
When equally loaded, works well as desired. Agreed, we should never try to compensate unequal output resistances. But when grid current unbalances load, the different cathode-ground and plate-ground resistances become important.
But not with a standard Cathodyne.
When equally loaded, works well as desired. Agreed, we should never try to compensate unequal output resistances. But when grid current unbalances load, the different cathode-ground and plate-ground resistances become important.
Last edited:
What is primary, is feedback. How it is applied, is important. Resistances are consequences. And they depend on load resistances, so without loads we can't say about them.