Compound amplifier." This is a basic cathode biased triode stage followed by a cathode follower, and the cathode follower is biased off the DC plate voltage of the first stage. I have seen this used in the design of many 60's era tube amps. It has fallen out of grace somewhat because it has been shown to have nonlinearity. But the current through either stage (and it is technically 2 stages) is not constant; the current through the first stage fluctuates with the signal placed at the grid of the first stage and the current of the second stage fluctuates with the signal at the grid of the second stage (which happens to be coming from the voltage changes across Ra). The current from the second stage does not drive the first stage, as it does in the SRPP circuit. So how is this considered a Constant Current amp. I have several other questions regarding your article, but his seems like a good place to start.

   Lets start with the claim that the bottom tube is loaded by a constant current source. Close, but no cigar. The load seen by the bottom triode's plate is (in the absence of an external load impedance) equal to
   rp + (mu + 1)Rak,
which in the case of the 6DJ8 with a Rak resistor equal to 200 ohms equals 3,300 + 6,800 = 10,100 ohms. While 10k is greater than the rp of the 6DJ8, it is not greater than the equivalent plate resistor that would be used in a straight grounded cathode amplifier with the same tube and power supply voltage. True, as the effective impedance of the active load approaches infinity, it approximates a constant current source. In the presence of a load impedance, however, the effective impedance of the active load decreases. In other words, the bottom tube is loaded by both the external load and the top tube. Given a load of 1 ohm, the effective impedance of the active load becomes roughly 201 ohms, hardly an ideal constant current source.

This month's E-Mail is an exchange from reader Pickacr who argues that the SRPP is not a push-pull amplifier. Here is the first of two e-mails.
Re: SRPP and constant current sources…

   I have a few questions regarding the author's analysis of the SRPP amplification stage. In particular, every time I see this circuit, it simply appears to me to be an attempt to create a constant current source driving a cathode biased gain stage. The current source tube (let's reference it as V2) appears to be set up in a cathode follower configuration who's input is determined by the current flowing from V2's cathode across what you have listed as Rak. As the input voltage on the bottom triode increases (V1 grid), the current drawn across Rak increases, so the voltage across Rak drops and so to the voltage at V2 grid must drop. As the voltage at V2's grid drops, the current through V2 decreases, attempting to maintain the current through Rak constant. To me, it appears then that V2 coupled with Rak (or better yet, R1 in the figure on page 2) forms a constant current source.
   I don't know if this is an attempt at a constant current source or not, but it sure looks a lot like high speed computer and op-amp circuits. The only difference is that in those circuits the transistors and resistors are doped and laser trimmed to create the constant current source, so the cathodes (if JFETs or MOSFETs are used) can simply be tied to the gate.
Also, I seem to remember seeing this circuit in an Audio Amateur article in 1992 or 1993. If I remember right the author stated something to the effect that the circuit was originally designed for high-speed digital computer circuits, back when triodes were used for such circuits. If this is accurate, then it would seem that calling this a push-pull circuit would be a misnomer.
   Also, looking at the diagram listed on page 2, you have a circuit titled "Constant Current Draw



www.tubecad.com   Copyright © 2000 GlassWare. All Rights Reserved

pg. 16