Tube Power Amp Reviews

What's the buzz?
The Avantgarde Uno's high sensitivity means that any residual noise in the signal is going to be more noticeable; this just comes with the horn territory. Even so, the noise emanating from the speakers when I initially fired up the Geminis was considerably higher then I've been used to, and tweaking the hum pots didn't help. The noise, present in both channels, was the kind of buzz normally associated with ground loops, so I tried the normal variations on grounding (connecting/lifting grounds at various points in the system). Nothing seemed to help. A call to Rankin followed, but he was at a loss to provide an explanation or a solution. Apparently, he had tried these very amplifiers with his own pair of Unos without any noise problem.

The solution came courtesy of Neil Muncy, who wrote the classic Audio Engineering Society paper (footnote 2) on the "pin-one problem," dealing with problems in grounding. Muncy is a noted consultant to recording studios, movie theaters, concert halls, and auditoriums, troubleshooting problems involving electrical noise. If anyone could solve my problem, he'd be the one. Conveniently, he lives about a 30-minute drive from me—and he makes house calls!

Muncy arrived with an arsenal of electrical measuring instruments, including a wide-bandwidth oscilloscope, and quickly determined that the problem had to do with the Gemini's grounding arrangements. He thought he could solve the problem, but it would require changing the ground hookup inside the amplifier. He got on the phone to Gordon Rankin, explained the problem and the proposed solution, and, with Rankin's blessing, made changes to the signal and electrical grounds (footnote 3). The effect of these modifications was to reduce the AC buzz to a level that was measurably and audibly much lower. Previously, it had been audible as soon as I walked into the listening room; now, I could just hear it from the listening seat with no music playing; it was masked by music playing even at a fairly low level. Although apparently the noise problem I encountered has not been reported by others, Rankin said that he felt that Muncy's grounding changes were very worthwhile, and will be incorporated into future production of the Gemini.

Are we there yet?
With the ground problem taken care of, I was ready to listen in earnest. The output tubes were Majestic 45s, and I was just starting to appreciate the Geminis' special sound when, one morning, turning on the amplifiers resulted in loud thumping sounds emanating from one of the speakers. The thumping subsided in about 10 seconds, but it was disconcerting. The same thing happened the next day, and this time I noticed that the 45 tube in that channel was flashing blue in time with the thumps. I e-mailed Gordon Rankin about it, and he diagnosed the symptoms as indicative that the tube was "going bad." He promptly shipped me another pair of 45s, these branded Silvertone. The amps were fine for a few days, and then I noticed a slight "ticking" noise coming from one of the speakers. Gently tapping the 45 tube in that channel stopped the ticking, but it later started up again, and this time it persisted.

At the risk of giving Rankin a heart attack or at least the beginning of an ulcer, I e-mailed him again, advising him of the latest tube misbehavior. The third set of 45s he sent me were made by Sylvania in the early '50s (the Majestics and the Silvertones were both from the '30s), and, whether because of their comparative youth or for some other reason, they worked flawlessly for the rest of the review period. There was no problem with any of the 2A3s.

The Geminis were plugged into a PS Audio P300 AC-synthesizer set to output pure 60Hz AC. I tried higher frequencies and various MultiWave settings, and there was some enhancement of the Geminis' dynamics, but any setting other than 60Hz resulted in a significant increase in transformer noise, so I ended up leaving the P300 in the 60Hz position. The indicated power draw was only about 50W.

The final tweak, effected just a couple of weeks before the amplifiers had to be shipped off to John Atkinson for measurement, was placing each amplifier on three Vistek Aurios MIB component-support devices (see Stereophile, May 2001, p.134). These made a greater difference than I'd ever heard from a change in component support: the sound became much more airy, with a wider, deeper soundstage and tighter, better-controlled bass. Except as noted, my comments on the sound of the Gemini refer to the amplifiers as supported by the Aurios. Although three of these devices add about $600 to the total cost, the sonic improvement was such that I consider them a must-have accessory.

Which Gemini?
Although it's recognized that the performance of an amplifier depends on the associated components (eg, whether the speakers present an easy or a difficult load), we normally assume that, in a theoretical sense, the amplifier itself is a constant. However, this assumption does not necessarily hold for tube amplifiers. Tubes given a generic designation (eg, the 2A3) can take a variety of forms, with little in the way of standardization beyond the pin complement and certain basic electrical parameters. Gordon Rankin estimates that there are some nine versions of the driver tubes, maybe 50 different rectifier tubes, more than 100 different New Old Stock (NOS) 2A3s, and 60-odd 45s. Many tube fans like to try different tubes to tweak the sound to their taste; with each tube change, the amplifier becomes in effect a different product.

I was not about to spend the rest of my days trying all possible combinations of tubes in the Gemini. To simplify things for review purposes, Gordon provided me with a pair of 45s—the original Majestics, which, as noted above, were subsequently replaced by the Silvertones and finally by the Sylvanias. (The price of 45s varies from $40 to $60 each, and up to $150 for a matched pair.) He also sent three pairs of 2A3s: NOS Raytheons ($65 each); new Sovteks, the stock tube normally provided with the Gemini (Platinum Matched Sovteks are $46 each); and new AVVT mesh-plate tubes ($280 each). The sound did change with the output tube, to the point that we might consider the Gemini as having multiple personalities.

---

Footnote 2: "Noise Susceptibility in Analog and Digital Signal Processing Systems," the best-selling publication in AES history, is available from the AES website as Preprint No. 3930. Neil Muncy can be reached at muncy@attcanada.ca.—Robert Deutsch

Footnote 3: I hasten to note that such design consultancy is not provided to manufacturers as part of Stereophile's regular review services. In this case, the review would have been stillborn without Neil Muncy's rearrangement of the Gemini's grounding requirements.—John Atkinson