Home Tube DIY Asylum

Do It Yourself (DIY) paradise for tube and SET project builders.

Re: yep.... it's blue

209.6.201.196

Mikey, I'm not just disagreeing for the thrill of it. You asked me to have a look at this circuit and my comments reflect my sincere and carefully considered opinion.

(1) With respect to simplicity, if you compare this circuit to a single-ended amplifier, the "Compact" is definitely more complex. I don't necessarily mean in parts count, but surely in terms of function. In the single-ended amplifier, the output tube grid voltage is converted in a straightforward manner to a current flowing in the transformer primary. In the "Compact", the current in the first output tube is split between a resistor and the cathode circuit of a second tube, connected as a grounded-grid stage. There is some cathode negative feedback applied to the first tube. The two dissimilar currents produced these tubes operating under very different AC drive conditions are summed by the output transformer to produce the final output.

Parallel single-ended isn't really comparable. In PSE, both tubes run under (theoretically) identical AC and DC conditions. From a conceptual point of view, they are a single tube, just split into two envelopes.

As I think everyone agrees, the output stage in the "Compact" serves both phase-splitting and amplifying duty. I believe it's fair to call this a bona-fide complication.

(2) The definition of balance is well-understood by circuit designers. In a perfectly balanced circuit, the sum of the two signals is zero. Any departure from perfect balance results in a non-zero sum when the phases are combined. We can represent the imbalance as a common-mode residual superimposed on the perfectly balanced phases. This is first-year electrical engineering curriculum.

If you look at the EL-84 data sheet, two volts change on the grid gets you about 20mA change in plate current. This means the cathode impedance is going to be no lower than about 100 Ohms. With a 135 Ohm common cathode resistor, about half the signal current flowing in the first output stage valve is going to be shunted to ground instead of flowing into the cathode of the second tube (and therefore into the other leg of the output transformer). This means the circuit is fundamentally unbalanced and it can't be improved without increasing the impedance of the "tail" circuit (or resorting to other, more extreme fixes).

An important measure of a phase splitter is its common-mode rejection ratio (CMRR), the extent to which the output remains balanced even when the drive signal is unbalanced. If we lift the grid of the second tube from ground and connect it in parallel with the first tube grid, we will observe considerable AC plate current in both tubes when a drive signal is applied. This is because the stage has a poor common-mode rejection ratio. Because of the asymmetric driver, the output stage of the "Compact" sees a large common-mode input signal, which is propagated to the plate circuit for the output transformer to deal with.

In the MC-240, there are three well-balanced differential stages following the phase splitter. Each stage (assuming matched tubes) has a high CMRR, so that by the time the signal reaches the outut transformer any imbalance due to the phase splitter has been thoroughly canceled out. As well, the 10K tail resistor more closely approximates a current source than the 135 Ohm resistor in the "Compact" (even taking into account the difference in cathode circuit impedances between the two circuits). And finally, the plate resistors in the phase splitter are deliberately mismatched to correct the imbalance. So, in my opinion, this part of your argument doesn't really apply. McIntosh knew what they were doing, and it was very, very, VERY different from the design of the "Compact". In fact, the McIntosh amplifier is arguably the very antithesis of the "Compact" -- just about as complex a push-pull topology as you can imagine.

With respect to John Tucker's implementation, I can't comment without seeing it. The version of the amplifier on diyparadise doesn't really make any changes to deal with this balance issue.

(4) There's no question that in spite of its other limitations, a circuit using a quality interstage transformer will yield orders of magnitude better balance at the output transformer primary than the "Compact" will.

(5) Philosophically, the "Compact" and Allen's design are quite different. The essence of Allen's philosophy is balanced signals at all points. The output stage does not serve phase-splitting duty in his circuit; this is handled by the cascode differential driver. The current source in the output stage is there for a number of reasons that have been discussed elsewhere. While it's true the "Compact" strives to have a differential output stage, the implementation (of a diff amp) is so crude, and the driver stage is so different, that it's hard to claim this circuit is similar to Allen's.

There's really nothing "funny" here. Just some straightforward and unemotional engineering observations.

There are about a thousand ways to design an amplifier that sounds good and communicates music well. Sound quality for the sake of audiophile bragging rights isn't my priority. On the other hand, I enjoy the theory and practice of circuit design. I view a circuit as a statement of a design concept. The concept of push-pull is balance. The "Compact" doesn't express that concept very well. Nor does it particularly express the single-ended concept. It's some kind of hybrid. I think the main point of this circuit is "cost savings". Otherwise, it doesn't say much to me conceptually. That's not to say that it's "bad" or "wrong", or that it doesn't play music well. I just don't find the concept as noteworthy as you do.

-Henry



This post is made possible by the generous support of people like you and our sponsors:
  VH Audio  


Follow Ups Full Thread
Follow Ups


You can not post to an archived thread.