Sound-Lab A-3 loudspeaker J. Gordon Holt, June 1988
Although the A-3 has been my reference loudspeaker ever since I reviewed it (Vol.9 No.6), I have never been very happy with its hunger for amplifier power. During my initial tests, I measured 76dB of output at 1 meter, which is about 12dB lower than that of most other audiophile systems.
SL president and chief designer Roger West contends that a conventional sensitivity rating makes no sense for a speaker with a very large radiating area, because its SPL doesn't diminish with distance nearly as rapidly as it does from a system in which the radiating sources are small. He's right. But the fact remains that my early-model A-3 is the only speaker I have used that occasionally runs my Threshold SA-1 power amps into audible overload. (In fact, it turned out that my original observation, that the A-3's diaphragms seemed to bottom-out at 94dB SPL, was incorrect. It was the amplifier that pooped out.)
It was Roger's guess that the power supplies in my A-3s were not putting out a high enough diaphragm-polarizing voltage, and there was evidence that this might be the case. The speaker has a potentiometer on it that varies the polarizing voltage, and this is supposed to be set just below the point where the speaker produces random rustling noises. Full-up, my speakers never produced that rustling, even though my line voltage is around 115V. It seems I wasn't the only A-3 owner unable to get full sensitivity from the system, because Sound-Lab subsequently came up with a revised power supply, having a slightly higher step-up ratio in its polarizing-supply power transformer. Now used in all production A-3s (which now cost $6350/pair), the new power supply is available as a trade-up option for owners of older A-3s for $250 per speaker. A pair of them was shipped to me.
No installation instructions were supplied, but it's my feeling that they should be. Although much of the procedure is pretty much intuitive, some steps weren't. For example, each power-supply module weighs about 40 lbs, but because of the shape of the system, you cannot "dump" the module out of its well after you've removed its fastening bolts. Neither is there anything on top of it that allows you to get a good enough grip to lift it straight out. I found I had to use a small screwdriver to pry up one corner, use a larger one to pry it up more, then prop it up with a screwdriver handle while I lifted the opposite side. It was then possible to get my fingertips under the top plate. But it wouldn't come out.
Lifting one corner, I saw the problem. Three banana plugs protrude from one side of the power-supply board, and these were getting hung up on the underside of the well cutout. I could not reach them to remove them. But if I pushed the module toward the speaker screen and lifted its opposite edge (the edge toward the rear of the system), I was finally able to extricate it enough to reach in and pull out the banana plugs. The clearance was awfully tight, though, I was tempted to just tear the thing out and worry about replacing the plugs later.
Once I found out how the things went in, removing the second module and replacing both with the new ones was easy. The banana plugs and their sockets are color-coded, and—except for a few fastening-screw holes that didn't line up (and could not be persuaded to accept a screw)—the rest of the operation was straightforward.
This time, the polarizing-voltage pots produced the requisite rustling sounds at about 1 o'clock, and were in the quiet zone at around 12. Plenty of reserve!
I had been told to expect a 6dB increase in sensitivity, to 82 SPL; actually, I measured 7dB. This may not look like much, but it is equivalent to a 5-fold increase in amplifier power—having the effect of boosting the SA-1's output to 800Wpc! In theory, this should now allow the system to deliver upward of 105dB of clean output from the SA-1s, and that's the way it worked out, almost.
While the SPL meter was kicking up to 105dB on middle-range program peaks without my hearing anything amiss, the limitation on output level was (as before) at the low end, where occasional clicking was audible at around 100dB, indicating diaphragm bottoming. There was no question now about amplifier overload; their output meters were only registering about -10dB (16W) when the clicking set in. That's 6dB higher LF output than I was able to get from the A-3s before, and 100dB at 40Hz sounds very loud. And a midrange 105dB is much louder than most people will choose to listen.
When I first fired these up, I was not at all sure about their sound quality. It had been two weeks since I had last listened to them, as they had been removed from the room while I was working with the VMPSes and PSBs, but it seemed to me they used to sound richer and less aggressive than they now sounded with the new power supplies. So, I put some pink noise through them at around 70dB level, and let them "break in" for 48 hours. That apparently did the trick. I don't know what might have had to break in, but the sound was now much the way I had remembered it, while being cleaner and more effortless on fortissimos than it used to be. The A-3s still won't blow you out of the room, but with the new power supplies, they're no longer the "wimps" I described in a recent issue.
The upgrade effects a significant improvement to the A-3s' dynamic range, and it's more than worth the cost.—J. Gordon Holt