Features

My survey of 300B power tubes last November (Vol.22 No.11, p.72.) was bad news for Svetlana, as four out of six samples of their 300B failed after 300-400 hours. Svetlana corrected a problem they had found and sent me four samples from the new production run. I installed these in two amplifiers and, to gain time, left them running day and night for 900 hours—nearly six weeks! To mimic real-use circumstances, each morning I switched the amps off to let the tubes cool down fully, then switched them on again, in order to expose the tubes to a substantial number of full thermal cycles. None of the four new-production tubes showed any kind of trouble; we can conclude that the Svetlana 300B is now reliable.

That's good news, but even better news is that the new Svetlana sounds considerably better than the earlier version. That improves its already respectable position in Group B (see p.79) by so much that the Svetlana tube clearly occupies a leader's place in this group, ahead of the Golden Dragon, KR, and JJ Electronics (formerly Tesla) 300Bs.

We listened to it after about 500 hours of burn-in. Its dynamics and control remained unchanged, but it sounded more open, even more lush, sibilants were cleaner, and the trace of midbass hollowness when used in the Kit One test amplifier was gone. Though still not as good as the Western Electronic 300B, it came closer to that tube than any other. I'm tempted to move the Svetlana into Group A, which has so far been the exclusive province of the WE.

In the review, I issued a small warning concerning the JJ Electronics heater structure, saying that it was "the only 300B tested here that is not equipped with a center-tapped filament, meaning that it might be somewhat more sensitive to (residual) hum on the filament supply." The opposite appears to be true, according to a reader who experimented with AC heater supplies. He found that the JJ Electronics is the least sensitive to hum from this source.

In hindsight, this makes sense. Think about what happens in a tube: The grid-to-cathode voltage controls the anode output current, but the distribution of that control along the cathode is not uniform: what happens at the cathode's extremities differs from what happens near the cathode's center. In a 300B, the cathode is the heater, as it is a direct-heated tube. All 300Bs, including the WE, feature a center-tapped heater: during a 50Hz or 60Hz AC cycle, both ends of the heater are "up" while the center is "down" (and vice versa half a cycle later). As the AC cycle is symmetrical whereas the cathode's mid- vs end-control over the anode current isn't, hum cancellation will be poor.

In the JJ, however, one end of the heater is "up" when the other is "down" (and vice versa). This is a perfectly symmetrical situation, with the natural benefit of a decent amount of canceling!

I happened to meet Western Electric's owner/director, Charles Whitener, at the London HiFi Show last September. I told him about our amazement when we found out that the WE tubes clearly needed more hours of burn-in than the others; and about our confusion over the first pair of WEs, which were supposed to have been burned-in at the factory but didn't sound right until after we'd given them another 200 hours. Whitener confirmed our guess: Other than a standard 24-hour test, the first pair of WEs hadn't been burned-in at all. WE never does that. (The guy who told me that they do no longer works at WE.)

Whitener also explained to me why WE tubes take a longer time to break in than other brands. As in all other makes, the coating (footnote 1) on a WE300B's heater contains barium oxide and strontium oxide, but WE does not add the usual small amount of calcium oxide. The calcium acts as a kind of pep pill in activating the barium, meaning that calcium-doped 300Bs reach their peak rather quickly, within 100-200 hours, and thereafter (according to Whitener) degrade slowly. Though the WE takes considerably longer to reach its top level of performance, its absence of calcium is said to maintain that high level over the tube's entire lifespan.

As of now, Western Electric suggests a minimum of 500 hours' burn-in, that recommendation to be included in their future literature.—Peter van Willenswaard