What amp should I buy to power Martin Logan Esl's?

08-05-13: Bifwynne
Bombaywalla, how do tube amps mate with ESL/capacitive loads? I kinda' recall that tube amps "generally" like resistive loads with positive/inductive phase angles. Come to think about it, I'm not so sure that SS amps particularly like super-low impedance loads with large negative phase angles.

In general, tube amps do not mate very well with ESL/high capacitative loads.
Tube amps, generally, have a much higher output impedance (as you now know) & together with the ESL looking like a big capacitor, the combination acts like a low-pass filter with a low(er) frequency -3dB frequency. This will have the tendency to roll-off the high frequency content of the music signal.
Now, tube amps can be used with ESL speakers but, if you notice, these tube amps are physically large units. Examples are Tube Research Labs products, Wolcott Presence amps, large VTL monos, CAT JL2/JL3, etc. These large tube amps have many parallel output tubes thereby reducing the overall output impedance & widening the bandwidth of that low-pass filter created by the amp o/p imp & the ESL capacitor.
So, this is from the impedance point-of-view.
Another point-of-view is the ability of the power amp to generate power into a low(er) impedance load. Tube amps have a higher output impedance & when you consider the voltage-divider effect, as the ESL impedance goes down, the tube amp cannot generate sufficient power into that low impedance thereby giving you anemic sonics. Per Ralph Karsten's constant-power theory, the tube amp will change voltage output swing & current output swing such that the product of voltage & current will equal the max power rating of the tube amp. So, if the ESL impedance goes down, the current delivery will go up & the voltage swing will go down such that the product of increased current & decreased voltage will equal the max power rating of the tube amp. Most tube amps do not double their output power as the load impedance halves (like s.s. amps) so power delivery over varying load impedance is mostly fixed. OTOH, s.s. power amps have a much lower output impedance & will deliver much more power into a lower impedance (within the limit of their power supply/transformer). Thus higher current delivered into a lower impedance mostly makes up for the lower voltage swing that would have resulted from a lower impedance ifff there wasn't an ability to deliver enough current. IOW, ability of a s.s. amp to deliver more current into a lower impedance gives a higher voltage swing. And, this means you will get more robust sonics because pk-pk voltage swing generated in the speaker directly xlates to driver pistonic action (which is what creates sound).
Power delivery into a low(er) impedance matters a lot when we have a highly capacitative load because the phase angle of a capacitor is negative & the effect of this ph angle is to reduce the real part of the music waveform (when a music signal goes thru a capacitor, it becomes a complex waveform meaning that the waveform now has a real & imaginary part. the real part is what's of meaning to us as it gets converted to sound energy. the imaginary part is the instantaneous power dissipated in a reactive/capacitative element. Average power thru a capacitor will be zero but it will have instantaneous power dissipation). So, when the real part of the music waveform is reduced, you need to pump more power from the power amp to compensate (or submit to listen at a lower volume/an anemic or less dynamic signal). This is where tube power amps eventually come up short/run into their output power limit. Big/huge tube power amps fare much better & of course, s.s. power amps also fare much better.
you are correct in stating that speaker loads with low(er) phase angle shifts are much easier to drive & the power amp (tube or s.s.) is much happier in that sort of condition.

Thanks Bombaywala -- very good explanation. After breaking big rocks with a sledge hammer, courtesy of Ralph and Al, I'm starting to better understand speaker/amp electrical matching. Of course, I've also come to better understand that speaker and amp designs are often the result of engineering/technical compromises. I believe Al and others made that point as well many times.

One such compromise that comes to mind is the benefit and cost associated with using negative feedback. Ralph and Al have spoken about many of the costs, such as TIM distortion and odd ordered harmonic distortion. They have also mentioned some important benefits, such as lower overall distortion, increased bandwidth and lower output impedance.

Case in point: I believe that ARC uses NF in most (if not all) of its tube amps. For example, the somewhat new ARC Ref 150 tube amp uses 14 db of NF and has approximately 1.1 ohm output impedance off the 8 ohm tap and about 60% of that value off the 4 ohm tap. See Stereophile/John Atkinson Ref 150 bench test last year.

Not sure if that turns a low impedance/high capacitance ESL speaker into vanilla ice cream for the Ref 150. Not even sure if the Ref 150 likes ice cream. But I wonder how well tube amps that use NF fare with ESL loads.

08-06-13: Bifwynne
Thanks Bombaywala -- very good explanation.

thanks for your kind words Bifwynne.

But I wonder how well tube amps that use NF fare with ESL loads.

I would think that tube amps using large amts of NF would fare quite well with highly capacitative loads from a purely electrical interface point-of-view. As mentioned by you in your post, NF would make the power amp measure really well on the testbench. I would think that you would get pretty decent sonics from a high-NF tube amp but what would be missing would be the soul of the music. I think that a high NF tube amp would never truly 'sing' to the music - it would have many of the audiophile attributes but would sound very clinical & non-engaging.

I forgot to write in my prev post - comments from other members who have experience with tube amps driving highly capactitative loads are most welcome. Hopefully I'm the person with the least amount of knowledge on this topic. Thanks.

Bombaywalla, not sure what constitutes "large amounts of NF," but the ARC Ref 150 uses 14 db of it. Perhaps somewhat responsive to the issue of tubes and ESLs, ARC reports (and Atkinson confirms) that output regulation is approximately +/- .5 db off the 4 ohm tap. So ... maybe the amp may add a touch of "flavor" to the acoustic presentation, but who said life is perfect.

Can't say that I recall reading comments that the Ref 150 doesn't sing to the music, but I'm tone deaf. What do I know??

I suppose we still don't know for sure what all the techno-babble means in real life. So I echo Bombaywalla's request that folks who have matched tubes and ESLs report their real-life experiences. Kinda' an interesting topic.