Amp for electrostatic speakers

Old school recommendations for electrostatic speakers was tube amplification. In the early days many solid state amplifiers had difficulty with the capacitive load of an electrostatic. I think most modern power amps would do a good job with your MLs. I know many receivers today have good amplification but I still think an amp that doesn't share power supplies with 7 channels and digital circuitry has the potential to sound better. I would not be surprised if a Parasound, or B&K, or even an Adcom sounded better than your Yamaha. These amps are more likely to operate cleanly with the unusual load presented by the ML.

I would get the Emotiva, it's a fraction of the cost of a Parasound.

I personally do not think either the Parasound or the Emotiva amplifiers will be a big sonic improvement over the receiver.

To bring out the sound quality that the speakers are capable of, you need an amplifier with very high sound quality and a very large peak current capability, and those amplifiers are not quite in that league IMO.

I suggest that you consider the Vincent SP-331 amplifier, which has the sound quality and peak power capability you need for best results.

It was designated "Amplifier of the Year" by The Absolute Sound" a couple of years ago, and for its $1400 price it is a steal.

I "personally" have not heard of any problems with any speakers when using the Emotiva. Emotiva is even popular with Magnepan owners. Outlaw also sells MC amps and they have beefier power supplies than Emo but are twice the price. I would need to know more about the Vincent to judge whether it may be better. TAS is not exactly objectivist.
Edited by Theresa - 1/26/13 at 9:35am

Electrostatic loudspeakers present a very low (often below 2 ohm) load at high frequency. And it is essentially capacitive, so is a challenge for many amplifiers. ML says this: "Impedance 4 Ohms, 0.52 at 20kHz. Compatible with 4, 6, or 8 Ohm rated amplifiers." Magnepans are a much easier load, almost purely resistive albeit around 4 ohms.

While I have used tube amps on 'stats, count me as one of the few who does not care for tubes and ESLs. The HF's tend to get harsh, and the amp has difficulty controlling the panel at LF. The midrange is gorgeous, natch.

Sanders makes some nice amps targeting ESLs: http://www.sanderssoundsystems.com/

Look for amps rated to 2 ohms. However, most decent amps will drive ESL's OK, including Emotiva.

What is your budget?

p.s. I would start by trying your AVR. It's "free" and may work just fine in your room. Also note ESLs are dipoles and usually require care in setup (location).
Edited by DonH50 - 1/30/13 at 6:58pm

second Don's suggestion.

these amps are built by Coda to Roger Sanders' specs, designed for the demands of estat speakers. He utilizes a large number of output transistors, far more than a budget amp like Emotiva, very large power supply & capacitance, have very high current capability & stable to very low impedance. these amps have much more peak current ability than even the Parasound Halo's. IIRC the Halo is about 60 amps per ch & the Sanders amps are over twice that.

Both his present company & former one, Innersound, are highly noted for both e-stat speakers & the amps. the Innersound amps used to get highly reviewed; google Innersound & Sanders ESL amps. his new amps are even more robust, with even larger power supplies. due to a lot of Magnepan owners using his amps, he designed a variant, the Magtech amp.

I've had an Innersound ESL300 amp for ~8 yrs. & it barely gets warm no matter what the load, driving my Magnepan 3.6's to reference level volume. I use the class a/b ESL for the fronts and let my Pioneer class D receivers drive the rest in a HT setup.

for power nuts, Sanders has an ESL monoblock that's 1KW @ 8 ohm & 1.6KW ! @ 4 ohms
Edited by ss9001 - 1/31/13 at 4:22am

I think I'd get a large ribbon before I'd get ESLs. They are just too fussy for me.

one reason to look at Magnepans

but the absolute best speakers I ever listened to for realism, depth, soundstage were Soundlab e-stats. I only wish I had bought them when they were $3500. now the lowest cost ones they have are $18K!

but Maggies can get you 80-90% of the way there for a whole lot less $$

There's 50" BG ribbons for DIY from Part's Express too. I've always liked Maggies too.

The three technical keys to your question are the speaker's sensitivity, impedance curve, and how loud you want to play them.

No need to speculate becuase we have one of John Atkinson's fairly complete reviews to work with:

http://www.stereophile.com/content/martinlogan-montis-loudspeaker-measurements

He says: "I estimated the Montis's B-weighted voltage sensitivity as 88.7dB(B)/2.83V/m, which is higher than average but lower than the specified 91dB. "

The impedance curve is as follows:



Which is one wild piece of work! Atkinson says:

"As with other MartinLogan electrostatic loudspeakers, the Montis's impedance drops to a very low value at the top of the audioband. The Montis is specified as having an impedance of 0.52 ohm at 20kHz. My measurement (fig.1), taken with an Audio Precision System One, gave a figure of 0.7 ohm at 20kHz, but this included 15' of speaker cable. Repeating the measurement using the DRA Labs MLSSA system, which compensates for the speaker-cable impedance, gave a figure of 0.55 ohm at 20kHz. Either way, that this speaker is a difficult load for the partnering amplifier to drive is compounded by the high electrical phase angle, and ameliorated only by the fact that music rarely has high levels of energy in the top octaves."

My first reaction is "What was the designer thinking when he let this out of his lab?" I would not have these speakers except under duress on the grounds that they are incompetently designed, and there are so many alternatives that are equal or better. You have to love 'em to keep 'em.

I guess your post is evidence that Yamaha's rating of their current line of AVRs for low impedance loads is not all talk because your AVR has apparently not grossly insulted your ears or cremated itself with this extreme load. Either that or you just don't like to play your system very loud.



So the question is: What don't you like about your system's current SQ.

If I were going to match this speaker up with a power amp that would be likely to be able to take the abuse, I'm thinking about either a fairly modern pro audio amp (e.g. Crown) or a classic like a Threshold SA 6e.

Last time I saw output devices stacked up like that was that Threshold SA6e that I sold a couple of years back.

Those old Thresholds were a piece of work!

I keep planning to get up and visit Roger (Sanders) but something always seems to come up... I am mainly afraid I'll end up driving home with new speakers in the back!

actually, the people who started Continuum Coda, the OEM amp builder, came from Threshold so the resemblance is not surprising Nelson Pass went his way of course. I only wish I had the $$ for a Pass Labs amp but I'd need more air conditioning too

those Sanders amps are supposed to be stable to 1/3 ohm.
Edited by ss9001 - 1/31/13 at 7:04am

yeah you being in the same state! you know he used to have his company right here in GA, a small rural town north of Atlanta. before my time and interest so had no idea until I started reading about his speakers & his history. he moved to CO.

I've read where he's really a nice gentleman to deal with and great speakers, just doesn't have the dealership exposure of a Martin Logan. one of our vacations out west, I might have to take a side trip for a demo

He absolutely, positively is. A real gentleman who knows ESL theory and practice inside out.

Yes, I own one of his amps.

I met him a few times decades ago, pretty sure he was still in GA then (and I was in MO). He's now about an hour or two north and I really need to go see him. I get prods now and then...

He wrote the book, actually several books, on ESLs over the years. Soundlabs I have heard but rarely. They were really impressive but IIRC I thought the bass a little boomy and highs rolled off some. One reason I ended up with Magnepans instead of some flavor of ESLs so long ago, but I still have a yearnin' to have a pair of 'stats sitting in front of me.

For $1k it would have to be used. You might ping Emotiva and ask them if an XPA-2 would work. I have one on my Maggies but am not sure about it on big ESLs. Outlaw and Parasound make good relatively inexpensive amps but again I would check with the manufacturer.

I would advise against class D amps; they have come a long way, but I doubt most inexpensive versions would be happy driving that load.

Be prepared to spend much time dialing in their position. It will take time but be worth it.

OK, I just have to vent here.

With all due respect, that is just more of Sander's nonsense.

The load seen by a traditional amp in an electrostatic speaker is the primary winding of the step up transformer, an inductive load, and the current drawn is based soley on the impedence of that primary winding. a secondary having no load beyond the capacitance of the speaker has a negligable effect on the primary impedence - Sanders "high current" design is just stupid.

I have also met with Sanders and followed his "career" from the moment he got the idea to build his first speaker from Dave Heremeyer's project.

In fact, Heremeyer was working on his second system and offered to help Sanders - who clearly knew nothing about ESL's.

How did Sanders repay this generosity?

By stealing Dave's designs, and then trashing Dave's first design as clearly inferior to "his own".

I'll give you just one example of just how clueless Sanders really is - and there are many.

The conductive diaphram coating of an ESL needs to be of high resistance to assure the driver operates in charge mode. This is just a simple fact, and has been since Walker first started work on the quads.

This charge mode increases the slew rate dramatically because it lowers the current nesessary to charge the diaphram fully in order to respond to the plates.

Yet as Sanders published "I don't even measure it anymore, I just rub it on till it sticks"

So how then did he then determine what current range he needed - not knowing what his load would be?

By playing music and watching a scope for clipping!

That, my friends, is absolutely ridiculous - you can't "see" this current clipping on a scope, the output signal is still there, the diaphram just fails to respond accurately because it's unable to fully charge fast enough to react.

Sure,.... using other people's work.

Sander's "original" amplifier design is shown here esl amp

Notice how Sanders claims HE chose the 8086 tube, and made modest design changes to Hermeyer's amp?

It's Dave's second amp design, published in a following issue - Dave gave him the schematic. (yes, I asked Dave - though I didn't need to, as it was obvious to me that Sanders couldn't design an amp if his life depended on it.)

Dave also advised Sanders on lowering the plate spacing to decrease the voltage requirements, advised him on plate material, diaphram material, coating material, panel size vs resonance, - Dave led him by the hand through his own project, with Sanders offering little more to the article than construction tips.

No offense, but looking at his current white papers - i.e. his "high current" amp - it's obvious to me that Sanders still has no more understanding of basic ESL theory than my dog,... he hasn't learned a thing.

Sorry if this sounds harsh, but I've watched folks heap praise on this guy for decades and enough is enough.

Keep in mind, Dave's final ESL system was published over thirty years ago, and yet it remains FAR superior in every respect to anything Sanders has ever, or will ever build.

Dave is the REAL gentleman here, and were he not, you wouldn't even know who Sanders is.

But back to the topic, direct drive HV amplifier technology has come a long way from the incredibly unreliable acoustat OTL tube amps, and being designed specifically to drive ESL panels, no other amplifier even compares.

This is an exceptional design from someone who actually does understand how ESL's work, and whilst it was admittedly a challenge to build, nothing I have ever heard in 30 years even comes close to the speed and detail.

ESL amp

I've run these amps on Acoustat 2+2's, Hermeyers, and KHL 9's,.... absolutely astonishing.

I use an Anthem P5 for my 5 channel ML/ESL and am very happy with it. This amp does not have a fuse and is stable down to 0.05 Ohms and sounds wonderful for music/movies.

But it is not cheap.
Edited by sharok - 4/10/13 at 3:09pm

dirty larry joined just to dump on Sanders? Whatever...

The impedance drop is nonsense? Read any number of reviews and/or measure any number of ESL's and you will find out the impedance does dip very low at HF. Transformers transform impedances and add their own parasitics as well. They can turn a HV, low-current load into a low-voltage, high-current input. And, I learned all that long before I knew about Roger or his papers.

The direct drive amp looks interesting. When I piddled with HV amps a long time ago the then-current crop of HV transistors were not very reliable, required serious drive circuitry, and tended to be noisy. I have heard Acoustats, Beveridge, Dayton-Wright, KLH, Quads, Soundlabs, etc. ESLs but never Hermeyers so have no opinion about them. Was Hermeyers the guy who wrote in AA back in the 70's?

Please explain the following impedance curves, all slowing very low impedances above 10 KHz:

Actually I joined two months ago - long before I ever saw this thread.

However, this is a man who doesn't understand even the most fundemental concepts of electricity, and yet he feels compelled to "educate" others.

I'm sorry if pointing this out offends anyone, but letting his nonsense go unchallenged isn't fair to those who really do want to understand how ESL's work, and better their systems.

No, the claim that ESL's are essentially a capacitive load is nonsense.

Just as his claim that conventional loudspeakers are essentially resistive loads:

A magnetic speaker presents a mostly resistive load to an amplifier, while an ESL appears mostly as a capacitor.

And it gets even better:

When driving a resistor, the voltage and current flow together. In a capacitor, the current leads the voltage by 90 degrees. This out-of-phase power delivery drives transistors out of their safe operating area. In conventional amplifiers, it can cause output transistor failure and/or the premature triggering of protective circuitry. This can cause the amplifier to deliver only a small fraction of its rated power and cause harsh sound quality. The ESL amp's output stage is so robust that it can drive out-of-phase loads with ease, and since it needs no protective circuitry, there is nothing to ruin the sound quality.

Do I really need to point out how incredibly stupid that is?

And his white paper is riddled with that sort of clueless nonsense.

Look, audio transformers were never meant to drive a purely capacitive load - if you ever see a 10k square wave at their secondary, you would probably cringe.

Now, you can smooth out the capacitive reactance of an ESL by simply using a load resistor across it - that's how it's done in direct drive systems, and how it's done in step up systems which are properly designed.

This is also how slew rate is determined - the lower the value of the resistor, the more current available to the panel, and the faster it's able to charge and discharge.

But you cannot do this at the primary of a transformer because it's not a capacitive load, and it's impedence doesn't react like one - it's just as simple as that.

I'm confused... Neil S. McKean's ESL amplifier design you recommended highly in an earlier post:


Contains the following statement in the introduction:

"An ESL is essentially a capacitive load."

The phraseology is virtually identical between Mr. Mckean's ("someone who actually does understand how ESL's work") and your own, contrary statement.
Edited by RUR - 4/11/13 at 5:39pm

Yes, and so is Sanders.... that's my whole point.

McKean's amp is direct coupled - it does drive a capacitive load, and so his design employs several appropriate methods to control a capacitive load.

What Sanders can't seem to grasp is that those same methods do not apply to a conventional amp like "his" because they are transformer coupled to the speaker, and the load the amp sees is inductive. (in addition to many other major technical differences)

Sanders confusion stems from the fact that when Hermeyer was trying to teach him about ESL's, he outlined these appropriate methods of controlling a capacitive load because his amp was also HVDD, and more importantly, he understood how ESL's worked, and was very knowledgable about electronics in general.

It should be painfully obvious even to a layman who reads those articles side by side, that Sanders had absolutely no idea what Hermeyer was talking about, and from his "white papers" it's also painfully obvious that he still doesn't.

All I'm trying to say is that there are several reliable, knowledgable sources of information available today regarding ESL technology.

Roger Sanders just isn't one of them.

A proper analogy might be that If your wife's obstetrician starts outlining the migration habits of storks.... you would do well to look for a new doctor.

The confusion can be alleviated by looking at the facts, such as the impedance curves shown in this post http://www.avsforum.com/t/1454033/amp-for-electrostatic-speakers#post_23189233

Unfortunately, it might take a lttle EE training to interpret what they show.



What I'm trying to show is yet another case where an overly general statement fails.

Following the dashed impedance phase curve we see that below about 380 Hz, this electrostatic speaker is a capacitive load. From 380 to about 680 Hz it is inductive, From 680 to about 19 KHz it is capacitive, and above about 19 Khz it is inductive.

An engineer would weight his evaluation of this curve by the areas where its impedance is the lowest. Therefore, it would be probably called capacitive, even though it is inductive over much of the frequency range such as below 380 Hz and above 19 KHz. The frequencies above 20 KHz could receive less weight in the analysis because it is above the normal audio range.

Hey guys, I'm kind of in the same boat as the OP. I just received Martin Logan Theos, love them, sound great on my Denon 3313. I don't listen super loud and the receiver isn't shutting off or anything. Should I upgrade to an Emotiva XPR-5 or XPR-1 monoblocks? Would I hear any difference since I don't listen loud? Am I going to be ruining the ESL's or the receiver by not adding an amp with more power? I've heard that ESL's can kill the lifespan of normal AVR's, is this true? I believe the Denon is only supposed to do 6 ohm speakers, and these Theos can go down to 0.8 ohms so I am worried if I'm killing the receiver or at the very least missing out on higher frequency sounds.

And as you know, what they show overall is a poor design from a bad choice of transformers, and a failure to properly load them.

Like you said,... what was the designer thinking when he let this out of his lab.

It's absolutely insane to drive a panel from the secondary of a step up without at least some token level of resistive loading - to dampen ringing if nothing else.

Exactly.

If the load impedence dips to a point that your input impedence enters an undesirable range - you have obviously chosen the wrong transformer.

Again, as you said from a design perspective, your base is matching the lowest impedence the load will ever be, and work from there to bring the peaks down.

This is because you cannot increase the capacitive reactance of a given panel without affecting it's physical operation - however you can decrease it by adding resistance across it, and then upping the signal voltage which also increases available current.

Yes, if you add resistance in series and then up the signal voltage, you do up the impedence - but you also slow charge time by limiting available current.

In the load the amp sees, even below 380hz, current is lagging voltage by 90 degrees, just the opposite of the panel itself.

This is pretty important when attempting to best supply the current needed at the panel.

Again, were it a capacitive load, the impedence issues are relatively simple to address through basic R/C formulas.

You can call it capacitive if you like, but if you apply R/C methods at the primary of the transformer to smooth the capacitive reactance it won't work, because it is not a capacitive load.

And that is the point here.

When choosing an amplifier to drive ESL's, you have to understand what their limitations are, why those limitations exist, if they can be addressed, and how best to do so.