The Lost Speaker: Don’t Know What You’ve Got ‘Til Its Gone

Posted by: Tom Martin at 11:11 am, April 25th, 2009

(You may want to read my prior blog on this topic for Part 1 of the quest to create a speaker with electrostatic transparency and superb dynamics.)
 
Part 2 -- Now that we know John Wolff at Classic Audio Loudspeakers has chosen horn drivers among other technologies, here is a brief and approximate history of why one might take this path:
 
In 1947, Bell Labs scientists succeeded in making the first transistor. The goal was to make more reliable amplifiers for telecommunications. That invention had profound implications for the audio business.

By the early1960s, consumer-priced high fidelity amplifiers were replacing tube amplifiers. As with many new technologies, the insurgent transistor products won on lower cost and (sometimes) higher reliability, at the cost of reduced quality.  Transistors also made it easier to create high-powered consumer amplifiers. The watershed events on this trajectory were the introduction of the Crown DC 300 with 150 watts per channel (1967) and the Phase Linear 700 with 350 watts per channel (1970). I believe the Phase Linear 700 was reviewed in Issue 1 of the abso!ute sound.

To cries of “power is essentially free”, a whole group of companies (Acoustic Research, KLH, Advent and others) began making small “acoustic suspension” speakers that had more or less full frequency coverage. They also had low efficiency, but with high-powered amps this didn’t seem to matter.

Naturally, these small acoustic suspension speakers were cheaper to make than the large horn speakers they replaced in the market. Just as the transistor amplifiers were more cost efficient than tube amplifiers. In a typical market, lower cost products with reasonable quality often dominate, and so it was with speakers and amplifiers after the mid-60s.

Now, as we all know, tube amplifiers didn’t go away. Thanks in part to the fact that there was a large installed base of tube amplifiers needing new tubes, and in part to Soviet technological backwardness, tubes continued to be manufactured. Since tubes and transformers are the primary unusual components in a tube amp, anyone who wanted to make a tube amp could do so (transformers are needed for other purposes, so are relatively easy to have built). And as high-end audio developed, the moderate cost premium for tube amps could be supported at the middle and upper end of the market.

The horn speaker, one of the technologies replaced by small loudspeakers during this upheaval, did not enjoy such an easy fate. As component speaker manufacturers shifted production to dynamic drivers, loudspeaker designers (often primarily integrators of components into a system) who might have considered using horns were faced with little choice on that front.  Technological development in horns slowed relative to dynamic speakers. If that weren’t enough, horn speakers are big and domestically challenged. It made sense for speaker manufacturers mainly to build speakers with dynamic drivers.

Horn speakers, lacking development effort, and with roots in PA applications, developed a reputation for low quality. This was amplified, no pun intended, by the mismatch between early transistor sound and horns. So, it wasn’t cool to make or own horns. Interestingly, note that none of that says that horns are bad; it simply says that horns are big, hard to make and out of favor.

Now, roughly 40 years later, transistor amps are extremely good, and we have a plethora of excellent dynamic speakers. So why bother with horns?

Well there are several reasons. First, we’re still faced with the Mark Levinson problem. Among others, Mark Levinson (the man) focused some light on a key audio problem. If I have a speaker that is 85 db efficient with one watt input at one meter listening distance, then to reproduce a 109 db peak at 4 meters, I need about 1024 watts of power. That isn’t an impossible amount of power, but most of us don’t have kilowatt amps. Adding to the practical side of this is that one would likely need a 20 or 25 amp circuit dedicated to this amp, if one had it. Many conventional dynamic speakers are in the 85-88 db efficiency range, so this is not purely hypothetical.

If I substitute a 97db efficient speaker for the 85 db speaker in the paragraph above, the same 109 db peak requires roughly a 64 watt amplifier. Horn speakers are commonly in this efficiency range.  You can see that a horn speaker might be better at dynamics than other speakers.

But wait, there’s more. It turns out that every technology travels with a set of assumptions. None of these are set in stone, but to see what John Wolff is trying to do, it helps to examine these.