[markmaloof]

Anybody have any idea how the 100 watt AKSA will deal with difficult loads, in paticular electrostatic speakers, which don't have the impedence dips in the lows that many "difficult to drive" conventional dynamic speakers have, but rather impedence spikes in the higher frequencies.  Besides some DIY dynamics (Dennis Murphy's MBOW1s, as two-ways, also toying with the idea of making them into his 3-way design, hence the desire for 100 watts), I have Quad 63 USA Monitors.  These are not the most difficult electrostatics to drive (no problems with a 25 watt Monarchy SM70 Pro ss amp, also a 40 watt Harman Kardon Citation V tube amp, and a 100 watt Music Reference RM9 tube amp made to handle difficult loads.)  I do toy with the idea of trying other electrostatics some day, perhaps Just Real Music's custom electrostatic panels with DIY dipole subs, or Martin Logan CLS, etc.  These sorts of stats can present problems to some SS amps (not as much of an issue with tubes, which often work well with stats, esp. the Music Reference RM9 which was voiced with Beveridge electrostats.)  I'm wondering if the AKSA would be up to the task of the weird loads of stats.

[AKSA]

Hi Mark,

Thank you for your post.  This question has been asked before, and I have copied a response here from some months back.  However, that said, I do have a 100W AKSA someplace in Sweden which apparently pulls an ESL 63 without complaint, but I've not tried it myself.

***********************************************************

The electrostatic loudspeaker was artfully designed specifically to destroy push pull Class AB solid state amplifiers. There is no other way to put it.

The argument goes this way.

The PP SS amplifier, both Class AB and Class A, has one mortal enemy. Amp kryptonite is called cross conduction (CC). CC puts one rail in direct communication with the other - we can forget about the speaker here - via the output transistors. A horrific current flows, aided and abetted by the largish capacitance at the rails, and the output devices lay down their lives in the electronic equivalent of a massive auto burnout.

So, what causes CC?

Oscillation; uncontrolled switching of the output devices precipitated by phase shifts across the global feedback network. Negative feedback must remain NFB, and when it strays into the PFB region, the amp oscillates in short order.

And what causes PFB?

Capacitance in the output stage, amongst other causes, typically around 2uF. Capacitors change the relationship between current and voltage; phase shift is the result.

The phase shift induced by high capacitance at the load transmutes negative to positive feedback - voila, oscillation, CC, and solid state death!

An ES speaker has planar electrodes stretched quite close together. Of itself, this is perhaps not too much of a problem since the resulting capacitance is only of the order of picofarads. However there is a problem; ESs must be driven with very large AC voltages and for this we need a transformer with a high step up ratio.

Typical transformers used in the Quad have a step up ratio of 100 times. This means (by simple transformer physics) that impedance, and specifically capacitance, is magnified by 100 times squared, viz 10,000. This is the problem........

If we take the 200pF interelectrode capacitance of a Quad (these are guestimates; don't have the spec sheets in front of me) and multiply this by the requisite 10,000 to get the 'reflected' capacitance at the low impedance primary, we finish up with the 2uF specified earlier.

Now, the AKSA will cope with 2uF at amplitudes where the inevitable ringing does not precipitate oscillation. It will handle both the ESL57 and the ESL63 AT LOW VOLUME. However, BE WARNED: Turn up the wick, and destroy the amplifier!!!

OK, why the hell can't the AKSA handle full power in an ES? Good question.

The reasons have to do with careful dimensioning of the feedback system which promote sonics, but take the amp just inside the Bode-Nyquist feedback criteria. The AKSA is much more stable than most SS PP amplifiers, but even it has limitations, so be warned: the ESL is the proverbial wooden stake/silver bullet for the AKSA, just as it is for other SS PP amps.

If you want to drive ESLs there are only a couple of choices. One is tube amps, which, with low feedback circuits, are assuredly good with these loads. The second is zero feedback mosfet SE designs, a la Nelson Pass, which drive them beautifully. Some years ago I designed a 28W single ended tube/mosfet hybrid called the Glass Harmony which is eminently suited to driving ESLs, and in fact four years ago I installed two pairs of these monoblocks with ESL63s and a Scandinavian active subwoofer system. They have never given a moment's trouble, and the owner is as captivated with the sound now as he was when he first heard them. They go remarkably loud, too.

As far as PP SS technology is concerned, the only SS PP amp which can drive an ESL with complete safety must necessarily be agricultural in design, with enormous lengths taken to prevent oscillation. The 200W Levinson is such an animal, and not a bad amplifier. But these measures ALWAYS compromise the sonics, and in truth the best amps for ES speakers are zero global negative feedback amplifiers - a VERY tall order for the designer, I can tell you......

Moral - wait until I have a SE tube/mosfet hybrid in the commercial market. It won't be cheap, but expensive, exotic speakers demand unusual amplifiers.

I hope this explains the issues. It's complicated stuff.

Cheers,

Hugh
_________________
Hugh R. Dean
Director
Aspen Amplifiers P/L
Melbourne AUSTRALIA
www.aksaonline.com

[markmaloof]

Innersound, since they sell Electrostats, makes a PP SS amp (300 watts per channel) that is made for stats and any difficult load.  It sounds GREAT on stats (I'm more of a tube head, but paired with a tube preamp, it's super nice, have listened to it at a friend's house on his Innersound Eros many times.)  Not sure of the topology except that it is PP, and Roger Sanders does not recomend it with high capacitance (Goertz, CAT 5, etc) speaker cables.  It's beyond my means at this time (even used, at least $1,600).  Maybe someday.

So, guess I should hang on to my Monarchy SM70 Pro (25 watts, class A mosfet, single ended, no negative feedback), now I understand all the more why it works just fine with my Quad 63 USA Monitors (and, of course, I have two tube amps to use as well.)  I may get a second one someday to have 80 watts 8 ohm, 120 watts 4 ohm, or get a pair of the Monarchy 100 watt monoblocks.  HOWEVER...I have not ruled out the AKSA for my Dynamic speakers, and I'm curious to try out the home audition of the 100 watter here in the US (I live in NYC, I'm sure there are other folks here as well interested in listening to this amp.)  I'll search around here to see about the home demo.  So here is my question in regards to the 55 vs. 100 wattt amp AKSA.  I have a feeling that if I transform my current DIY dynamic two-way (Dennis Murphy MBOW1) into his three-way design, that the 4 to 5 ohm load, 86 DB efficiency of the MBOW1 3-way would like a bit of solid state control, and I've been intrigued about the raves around the AKSA amp.  I sit somewhat nearfield (7 feet from the speakers), and listen at moderate volumes (maybe, at the loudest, peaks in the mid 80s or maybe low 90s, but not often, more likely at night and in the upper 70s to low 80s.)  I'm wondering if the 55 watt AKSA will be enough, given that I sit close to the speakers, not into cranking it even when "rocking",etc.  Again, the impedence of the MBOW1 3-way design is pretty even, mostly at 4 to 5 ohms, never dipping much below 4 ohms, and the sens. is around 86 db.  I sure my 100 watt tube amp will be fine, possibly the 40 watt tube amp for moderate levels, but I do enjoy switching to SS with a tube pre for a different sound (hence, the Monarchy, which I'm glad I have, since I have to do a little work on both the tube amps.)  I have read that the 55 watt AKSA sounds a little better than the 100 watt, so I'm curious if I could get away with 55 watts of AKSA power with the dynamic speakers.  And, of course, I would avoid using it with any electrostats!

[bhobba]

The electrostatic loudspeaker was artfully designed specifically to destroy push pull Class AB solid state amplifiers. There is no other way to put it.

and

If you want to drive ESLs there are only a couple of choices. One is tube amps, which, with low feedback circuits, are assuredly good with these loads. The second is zero feedback mosfet SE designs, a la Nelson Pass, which drive them beautifully.

Really thought provoking stuff.  Because I like to think about technical stuff I went around looking for exactly what amplifiers would have no trouble.  I found circuits for the ZEN (able to deliver 10 watts) and variants (able to deliver power up to 40w) but was concerned about the class A operation.  As one site said if you are like a lot of audiophiles you leave the amps on all the time and over ten years this can amount to - wait for it, $12,000 worth of electricity.  YUK - I like hi-fi but not that much.  Indeed $12,000 will buy one hell of a good hi fi system. Of course you can switch it off - but even say using it on the average of 3-4 hours a day it will add up to $2000 over ten years.

There simply must be a better way.

Now, the AKSA will cope with 2uF at amplitudes where the inevitable ringing does not precipitate oscillation. It will handle both the ESL57 and the ESL63 AT LOW VOLUME. However, BE WARNED: Turn up the wick, and destroy the amplifier!!!

That being the case would a possible solution be to limit the maximum voltage than could be fed into the AKSA so it is not possible to turn up loud?  It is well known that electrostatics produce light bass and many have a subwoofer to handle the low frequencies.  If this was crossed over at say 300hz then only a small signal would need to be sent to the electrostatic - possibly enough to never cause trouble for the amp.  I wonder if this would be size deponent - ie I wonder if say the 25w Aksa would not experience this problem without any kind of limiting?  That being the case, and a crossover at 300hz, this would subjectively be equivalent to 100W?  If that was not possible then I would simply be inclined to say - I will not turn the volume up - simple as that.

What do others think?  Or have I missed something?

Thanks
Bill

[kyrill]

Clever, but you force to open Pandora's box?

[bhobba]

Clever, but you force to open Pandora's box?

Yep - it is just musings on my part - as I said I like thinking about technical issues.  

There are a number of considerations here.  First electrostatics are not everyone's cup of tea - I personally love them but know not everyone does.  And of those that like electrostatics some would consider it sacrilege to cater for the low frequencies with a conventional drive.  I am not one of those - I have heard Martin Logan's and they are a great speaker.  In fact I chose my current speakers (Axis LS88's) by auditioning them against Martin Logan's - they were the closest conventional driver speaker I could find.

But since I made my previous post I have come across a wonderful site http://www.tcaas.btinternet.co.uk/index.htm.  It contains details of a 15W class A amplifier specifically designed to drive electrostatics.  Now even at 15w a class A amp will give off a lot of heat (it draws a current of 4 amps or with a +-20v rail 160W) but I believe it is manageable at this level.  Combining that with a 25W AKSA to handle the region say 60 - 300hz from a closed box and a subwoofer such as the Siegfired Linkwitz Thor to go down to below 20hz I think it would be one hell of a good system.

Thanks
Bill