Tubes and Summits
Hi Tommy, for some time now I've been preaching tube amps with Summits, but let's go back a bit . . . . . . . .
I've always felt electrostats generally sound best with tube amps. Going all the way back to Peter Walker's Quad ESL 57's and his 15 watt Quad amp. There are always pros and cons of course, but in this situation, the pros are all sonic, and the cons are all electrical (amp/speaker matching).
First the pros. The inherent sonic qualities of tubes amps are well known. It is a characteristic of a tube gain stage that it is able to preserve the natural timber, harmonics, and decay, that our brains associate with live sound, while amplifying the signal strength. Solid state gain stages if not very carefully designed can easily lose these delicate signal components while multiplying the signal strength. Some people characterize their sound as "hard", but I think it's because only the best (and most expensive) solid state amps are able to preserve most of these subtle nuances; something tubes seem to do with ease.
The cons. Tube amps need an intermediary (the output transformer) to match the high impedance of a tube to the low impedance of the speaker (the load.) This OT is like a gearbox on a vehicle, where you select the right gear to match the engine (the tube) to the steepness of the hill (the speaker.) As long as the slope of the hill (the speaker impedance) stays relatively constant, no need to change gears. But if it goes up or down, you find you are in the wrong gear for the load. In a car you can change it. In a tube amp you can't. A solid state amp doesn't care. The power output transistors plug right into the speakers. No OT needed. The metaphor here would be a fluid coupling transmission, which is constantly self adjusting.
Electostatic membranes are so accurate and responsive to every little variation in a signal, that they are the most "revealing" transducer we have come up with. And they're so light weight, it doesn't take gobs of power to move them back and forth accurately. So a modest powered tube amp (which in high powered form, is terribly inefficient compared to a similarly powered solid state amp,) would seem a perfect fit for a stat. Accurate signal amplification, low power requirement, low damping factor requirement. But there's a problem. Electrostats have characteristically wide-ranging impedance curves -- the steepness of the hill varies a lot from the low frequencies to the high frequencies. So if you select a tap (2,4,8,16 ohms) on the output transformer to match an electrostat's impedance at one place in the frequency spectrum, it probably won't match very well elsewhere.
There are several strategies to minimize this problem, which I will skip in this discussion, but the bottom line is that many tube amps can't be successfully paired with electrostats, and many that can, do so only if certain of the aforementioned strategies are employed.
For years, MartinLogan has manufactured hybrid speakers. Stats + cone woofers to "fill in" where (reasonably sized) stats would not go. Because of the woofer, you couldn't afford to drive the whole shebang with a tube amp, unless you were an oil sheik OR you were willing and able to buy a tube amp for the stats and a solid state amp for the woofers, double the speaker and interconnect cables, and bi-amplify the speaker.
With the Summit, MartinLogan has finally addressed the problem. They have basically provided an already bi-amplified speaker, complete with solid state amp for the woofers, and all the user has to do is provide a modest tube amp for the panels, but which tube amp to use!?
My personal favorite at this moment in time is the newly re-designed McIntosh MC275 Mk IV or Mk V. (They're the same except the Mk V has standard speaker binding posts.) Before I explain why, I should say I don't own a pair of Summits. Rather I own "component Summits": a pair of MartinLogan CLS IIz's and a MartinLogan Depth.
I have driven my stats with some very expensive tube monoblocks, and a very expensive solid state stero amplifier. The trioded monoblocks (Audio Research) were humongous, in money, size, and heat (eight power tubes per side!) and put out a gigantic (not!) 140W per channel. But sounded lovely. The SS amp (Mark Levinson 23.5) was smaller, more efficient and delivered 200W/8ohms or 400W/4ohms. It sounded lovely too -- until one night I brought over a friend's (smaller) tube amp, which did many things wrong, but one thing very very right: it made the stats make music again. And let me say, I'd grown quite used to, and I guess happy with, the beautiful Levinson sound. That is until I heard some (relatively crappy) tubes again.
The new MC275 tube amplifier is unique in its ability to drive electostats. With only four power output tubes (two/channel) it can put out an honest 95W/ch and drive the varying load stats present. How? Something that McIntosh pioneered and patented ages ago: The 'Unity Gain' circuit. Power is drawn from both sides of each power tube and fed to two (instead of the usual one) primary transformer windings. This creates better efficiency (they run comparitively cool for a tube amp) and the output transformer can drive the speaker over a wider impedance range. In addition, if one needs more power, the amp can be converted to a 150W+ monoblock at the flip of a switch. In fact many McIntosh dealers are using Summits to demonstrate this amp because they sound so amazing with it. And at $3900 apiece, they represent one of the best values in audio today.
Questions?
