This point is the important one regarding tone quality

"One technical complication is distortion cancellation. A second triode stage will subtract even harmonics while adding odd ones. More stages makes the accounting worse, but you can see that you might get a better balance of harmonic structure - it's just not reliable or predictable unless you laboriously account for the harmonic structure of every stage."

I used to make a big point of this "error" that's often in commercial designed tube preamps and amps that like to use the same part types over and over just for quantity discount. The simple problem is not enough understand the tonal richness aspect control you can do that SET guys are all into when you do it right.

Paul is absolutely correct about a cascade of two triodes. One inverts and feeds the next stage that also inverts. This is even harmonic canceling and will "dry out" the sound as PP stages do, for those that highly value that SET tone. One polarity is plus, the next minus, which when multiplied is a problem with close matched-distortion triodes.

I posted several times in the past to never use the same triode type twice in the full chain of the system, if you can help it. If you have two 6SN7's in successive common cathode stages in a 300B amp, harmonic richness will be lost from each of these 6SN7 tube stages. But if the preamp has a 6SN7 in a common cathode stage, and then the amp has a 6N1P input and then a 6SN7 driver for a 300B amp, the polarity aligns two 6SN7's in the same direction now: +, -, +, with the 6SN7's occupying the plus sides and the 6N1P the minus side. Now similar 6SN7's will just multiply their sonic signatures together, adding euphony instead of subtracting. Which is best?

In my experience, I don't want to multiply nor cancel the sound of any single triode type. It's then too much or too little. True detailed neutrality in a proper rich tonal midrange and highs will not be so radical if every triode is different and system complementary, and you can mix the sounds with good low distortion triodes all along the chain.

In my best system, changing 6ER5 -> 6ER5 to 6ER5 -> 6GK5 was a revelation in harmonic preservation. Then it went to a passive AVC. But where the AVC served as the phono output triode's variable parafeed WOT-style preamp stage; it was designed to be okay for this even though changing volumes (loading) is not optimal here. It was not just for strong low impedance drive needed, but for a lighter drive signal, too. This is how to make the AVC/TVC the superior choice for a passive IMO - make it do essentially two things at once.

Then the amp had triodes 01A -> 45. Just 4 triodes in the path from LOMC to speaker terminals (SUT, 6ER5, 6GK5, 01A, 45, and a high efficiency horn for less gain needed overall). Not a single duplication of triode sound to multiply that sound or cancel it, and tubes were selected for overall system quality taken together as a whole (and why 01A was selected over an attempt to use 26 as driver).

I would not go 01A -> 45 -> 45 or even 01A -> 46 -> 45. Those have issues with the harmonic balance of things and overall transparency. The 01A is more than capable for driving a 45. It is a mistake to think 1 mA of current drive is not enough into a 45's Miller capacitance. First verify that premise.

The slew rate into a 80 pF load from 1 mA is:
dV/dT = I/C = 1E-3 Amps/80E-12 Farads = 12.5 V/usec. For the large signal bandwidth, it was about 40 KHz, -3 dB, full power output best case. Max slew rate required for this to swing 45 Vpk into the 45 grid = 12.5E6 V/sec by that current slew rate, and by max bandwidth slew rate = 2*PI*(45V)*(40KHz) = 11.3E6 V/sec. So STILL, at only 1 mA here, you cannot go into slewing in this case. You are power-bandwidth limited instead.

-Kurt

Post a Message!