Using tube amp with electrostatic speakers.


Moons ago I started similar discussions and thought I had been given enough good advice not to approach the subject again. Here goes anyway. I've used Martin Logan electrostats for well over 30 years with quite a few different amps but have recently switched to a tube amp and dynamic speakers with which I am very satisfied.  It consists of the Cary Rocket 88R amp and Serie Reference 3 speakers. 

My brother was visiting last week and was so impressed with the sound that he decided that he might want to try a tube amp also (probably the same one as mine).  However, he is using a pair of SL3's that I gave him years ago and I'm concerned primarily about the current requirements of the Martin Logans as well as other concerns that I'm not thinking of.  I don't want him spending money on something that may not bring him improved sound so would appreciate more advice to pass on to him.  He currently uses a Rogue Audio SS amp with his SL3 speakers and, to me, it sounds very good. 
jimbreit
lewm,

Thanks for the response.  I have an old set of Soundlab speakers (A1 and B1S) that I have been working on and upgrading.  I am currently driving the system with SS amps (Innersound ESL 800) but would like to try a couple of tube amps I have laying around (both are 60W).  One of the tube amps does have a 2 ohm tap (it is a Margules U280C).  I would use one amp to drive the A1s and the other to drive the B1S (with a crossover).  I have the luxury of optimizing the frequency band for HF and LF by using the B1S as my bass speaker.  Thus my question about using the 2 ohm tap for the A1s if bandwidth limited.  Would there be any benefit driving the B1S with a tube amp at the high impedance or would this work equally well with a SS amp.  Once again, thanks I really appreciate your input.
Also, since you seem to know Soundlabs, would there be an optimal frequency band for using the A1s from an impedance standpoint (e.g. minimize the wild impedance swings that the amp sees).

Regards,
Gary
I am fairly sure that the A1s would look like my 845PX with respect to impedance, which is "very high" up to 100Hz (50 to 100 ohms), then around 4-5 ohms up to around 1kHz to 2kHz where there is a sharp notch dip to 2 ohms or so, then back up to more reasonable values, then back down at 5kHz and above, to 2 ohms or less at 20kHz.  This is from memory.  (I could REALLY hear an improvement in vocals when I got rid of that 2-ohm dip; see below.)

You can probably assume, if you don't want to or cannot measure, that the B1S will present a fairly high impedance to the amplifier, and in that respect it should be an easy load for any tube amp, but the question would remain how much amplifier power you need to make it sing.  You might ask the folks at Sound Lab.  

One option you didn't mention but which works quite well is to bi-amplify your A1s.  I've never done it, but you can experiment with that idea.  See if you can find the old "Sound Lab Owners Group" archives for some hints.

And finally, you might do what I did with my 845PXs: (1) Acquire a full-range ESL step-up transformer with a turns ratio of 1:75 up to 1:90.  Plitron makes a nice one, toroidal at 1:75, but there are several others. I used a EI type transformer made in Australia that is unfortunately no longer available.  (2) Remove the existing treble transformer in your A1s and also remove the RC network that effects a high pass filter feeding the treble transformer. That RC network is the problem; it creates the low impedance dip right around the crossover point, because of the value of the R that SL chose to use. Also, the R (typically 5 to 8 ohms in value depending on how old your speakers are) soaks up amplifier power, because it is in parallel with the panel at all frequencies. The intrinsic impedance of the panel is mostly way higher than 5 to 8 ohms, so most amplifier energy flows through that R.
(3) Install the full-range transformer in place of the existing treble transformer AND the RC network, so there is no longer an RC at all.  You can retain the inductor that feeds the bass transformer.  Make sure the two transformers are "in phase".  And then your A1s can be driven by your 60W tube amplifier all day long. The resulting impedance is going to be 20-30 ohms from about 100Hz to 5kHz, going down from there but fairly constant between 100 Hz and 2kHz. And the efficiency is ridiculously improved.  I doubt my speakers draw much more than 20W, steady state.

Or finally, talk to Roger West about upgrading your treble transformer to his latest device, which is more nearly full range. That allows for a lower crossover point and also gets rid of that midrange dip in impedance, according to them. With that option you keep the crossover (but change the values according to his instructions). Plus, this mod is factory sanctioned, unlike what I did.
Does the 4 ohm tap in this case make sense to you?
@garyosoba Yes.

Lew, I've heard the new Sound Lab backpanels a number of times and its pretty obvious that Dr. West got it right with this one. Its much easier to drive (we can do with an MA-1 what took the MA-2s before) and sounds better too.
Ralph,
Your information about the new Sound Lab treble transformer ("back panels") is consistent with what I have heard too, from other Sound Lab owners.  They all seem to like it, and they all sense the improvement in sound compared to the old set-up. Even the one or two guys who first tried "my" approach and who were not satisfied, are now happier with the new toroidal treble transformer.  No doubt that is the simplest thing to do, if one owns an older set of panels.  

Note to others that getting rid of the crossover entirely was not originally my idea; it was communicated to me by "Will" in Australia, via SLOG.  Will also helped me and others to obtain the Australian-made full-range 1:90 step-up that he used too to modify his 945PXs. I bought the last pair of the original run of those transformers, but the vendor had more of them made due to the demand from among SL owners. There may have been something odd or different about that second batch of production units, because I know of at least two guys who were not happy with them. We'll never know why, with certainty.  This is why I would recommend the Plitron 1:75 full range toroid, not the Australian transformer, to anyone who wants to try the mod I describe above.  I personally have never thought twice about adopting the new SL toroids to replace what I've got.  Anyone in the DC area is welcome to stop by my home for a listen.  

To be more clear, the Plitron toroid is a full-range step-up that one might choose to install if one wants to go with no treble crossover at all.  SL sells a toroid that requires a high pass filter in front of it; it's not designed to run full range.  But with the SL toroid, the hi-pass crossover point is down around 200Hz, which does dramatically improve the impedance curve compared to original and the efficiency of the speaker.