Shoot out

October

Page 12

As 32 ohms divided by 4 equals 8-ohms, rather than burden the SPICE engine with eight 6AS7s, we’ll use only two 6AS7s but increase the load resistance by fourfold. (The 1-ohm cathode resistors are a small insurance policy against the triodes running into excessive current conduction and could be removed without altering the circuit’s functioning in the least; but I would not recommended it, as these tube are notorious for burning up—in fact, a 4 ohm resistor would be safer still.) This amplifier has a gain of 0.33 and puts out 28.5 volts just before the grid goes positive relative the cathode, which equals 12.7 watts into the 32-ohm load and, when the full array of output tubes are used, 50 watts into an 8-ohm load. The output impedance is 30.5 ohms for one pair of 6AS7s and 7.6 ohms for a pair of 6AS7 quartets. The THD is 3.37% and the harmonic distribution is shown below. (Typical push-pull harmonic structure.)

And now for something completely different, let's see what the modern circlotron has to offer in opposition to this clunky, push-pull, class-AB amplifier.

The amplifier has a gain of 0.33 and puts out 28.5 volts just before the grid goes positive relative the cathode, which equals 12.7 watts into the 32-ohm load and 50 watts into an 8-ohm load, with 8 output tubes. The output impedance is 30.5 ohms for one pair of 6AS7s and 7.6 ohms for a pair of 6AS7 quartets. The THD is 3.37% and the harmonic distribution is shown below right. (The only difference I can spot is at the 2nd and 4th harmonics.)

Note the 100% classic push pull signature: even harmonics are notched out and odd order harmonics predominate (lots of odd order harmonics in this example, much like a transistor amplifier). There isn’t a single ended amplifier in the world that has a harmonic structure like the one shown below (or sounds like this amplifier). Well, so much for the widely imputed single-ended aspect of the modern circlotron.

Or, if more proof is needed, look at the current conduction through each triode in the graph on the next page. Once again, one-hundred-percent-classic-push-pull operation. Notice how the class-A overlap only extends to twice the idle current, about 100 mA or about half a watt into 8 ohms with 8 output triodes. Well, so much for pure class-A operation by necessity. But couldn’t the idle current be increased to further the amount of class-A operation? Not by much, as the triodes are already dissipating 8 watts each, so there isn’t much safety headroom left to go.

What happened? One dB at –140 dB (1/10,000,000) is no difference (probably a SPICE engine error do to 1G-ohm node shunts), at least no difference worth arguing over. Where did the magic go? All we have left to the amplifier that stood so tall is a crummy, miserable push-pull, class-AB amplifier that topologically isn’t that different from the cheap solid-state amplifiers from the 60s.

Wait another minute; how do we know that the SPICE engine isn’t rigged to show no difference between these two wildly differing amplifiers? That’s possible isn’t it?