Doesn’t cathode feedback result in a higher output resistance? Even higher than the pentode itself normally has? I get that the distortion could be quite low, because it attacks multiple distortion mechanisms at once. But the output characteristic would be very un-triode-like, unless schade were added on top of it.
Nope, the simple fact that part of the output is from the cathode should hint to lower Zout. Typically 10-12% cathode fbk with most pentodes results in triode Zout. Damping factor around 1.5-2 as a typical figure. The effect on Zout is in the same direction as Schade and in fact the percentage above has the same meaning (i.e. percentage of output voltage coming out of the cathode). There is little point in using both. Crowhurst use call it also with another generic name: partial cathode follower. I use cathode fbk because it's better overall in my opinion as the front end only needs to deliver the extra driving voltage but input impedance can be very high. As always, there is a trade-off and that is a more complicated output transformer. However it is not a problem for me. I have found a winding geometry that works great, with bifilar primary and very large bandwidth. I guess in the current market would be VERY expensive as it takes a lot of work to do it right....
Here is an example applied to a SE amplifier:
http://www6.plala.or.jp/Michi/Renewal838SE/Renewal838SE.html
At point 4 it says that with 7.2 dB cathode fbk damping factor goes from 0.35 to 1.8.
From what I have seen, higher gm tubes get more benefit from the same amount of cathode fbk.
Here is an example applied to a SE amplifier:
http://www6.plala.or.jp/Michi/Renewal838SE/Renewal838SE.html
At point 4 it says that with 7.2 dB cathode fbk damping factor goes from 0.35 to 1.8.
From what I have seen, higher gm tubes get more benefit from the same amount of cathode fbk.
Last edited:
It depends on what is fed back, as always: If the FB signal is proportional to output current, output impedance is increased. Proportionality to output voltage decreases output impedance.
Best regards!
Best regards!
That is not case. The cathode winding is an independentnd widing and part of the primary. It is coupled to the secondary exactly like the plate winding. There is no loop. It's like a conertina splitter with 1 input and 2 outputs, except that the cathode winding has inverted phase so that cathode signal and anode signal add up.
The partial cathode follower was well explained by Chrowhurst in one of his famous articles.
The partial cathode follower was well explained by Chrowhurst in one of his famous articles.
Last edited:
Every winding is coupled in a transformer , for the cathode winding it is negative feedback from plate primary .
Extreme example in McIntosh where ratio is 50-50 and the gain is 1 for the output stage , a lot of negative feedback
Extreme example in McIntosh where ratio is 50-50 and the gain is 1 for the output stage , a lot of negative feedback
Last edited:
Of course it is negative feedback but that's not a good description. The feedback is internal, it's not coming out of windings. It is the tube the active element that adjusts itself to different loading conditions.
The signal is applied to the grid and develops at both plate and cathode simultaneously. Both cathode and plate are loaded simultaneously through the dedicated windings and coupled to the secondary.
See Crowhurst's article attached. The circuit I am talking about is called "modified Ultralinear" in Fig 16. In the article it is shown with a partition of 57% at the plate and 43% at the cathode, in my case it is 90% and 10%, respectively.
The signal is applied to the grid and develops at both plate and cathode simultaneously. Both cathode and plate are loaded simultaneously through the dedicated windings and coupled to the secondary.
See Crowhurst's article attached. The circuit I am talking about is called "modified Ultralinear" in Fig 16. In the article it is shown with a partition of 57% at the plate and 43% at the cathode, in my case it is 90% and 10%, respectively.
Last edited:
Nobody said it is not simultaneous , it is obvious that both windings are working together in series with the same tube current like a single primary , but if the mechanism is so simple without coupling between them why not use a separate smaller core for the cathode winding coils ? Then it can work with any normal transformer ( a big plus ) and you would have just an AC voltage drop on the coil , like on a cathode resistor ( negative feedback ) but without DC voltage drop .
Last edited:
Partial cathode loading is feedback in the sense that signal voltage is applied between grid and cathode (the valve's input) and series feedback is applied between cathode and signal ground, and therefore between grid and cathode. The valve itself operates exactly the same, on the same loadline, as one with that same total load impedance all in the plate load). It has the same gain, measured across its elements, and the same distortion (before feedback) as it would with all of the load in the plate portion. No magic allowed. Of course this is also true for fully cathode loaded valves, or any other combination.
All good fortune,
Chris
All good fortune,
Chris
A simple coil in the cathode circuit would not represent the same thing. There would be no output signal from the cathode. It's the same as with resistors: if there is an anode load and a cathode resistor and one only takes the output from the anode, the output impedance is actually higher if no bybass of the cathode resistor. If one takes the output from the cathode (with the right resistors and supply) then gets a cathode follower. If one takes both outputs with equal resistors that is concertina splitter, which has low Zout (equal Zout's from plate and cathode in this specific case) when BOTH anode and cathode are loaded (with the same load).
Last edited: