• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Electrostatic speakers?

misterdog

Addicted to Fun and Learning
Joined
Dec 7, 2018
Messages
533
Likes
407
A properly designed enclosure has no box sound.

Though what of the distortion from magnetic drivers in speakers from ' good conventional speakers which don't cost a lot'



So this is an electrical motor or electro-magnetic motor system. Now, the magnetic circuit, which is these metal parts and the black disc here, which is the actual ceramic magnet, it inherently is not perfectly linear, particularly when you get to high currents or high power levels, or the system heats up.

So that's one area where there can be some distortions. The voice coil moving in the magnetic field itself, as it moves more and more, it actually creates its own current because it's moving, which then interferes with the magnetic field. So the two actually upset each other to a certain extent, and can cause some distortion.

Now, obviously, the radiating and the moving surfaces, the cone, the dust cap, the rubber surround, the spider suspension, all of those things can create and do create levels of distortion. Those distortions can be the fact that if I put a perfect sine wave in, the cone is not going to move perfectly in and out in a perfectly what's called pistonic fashion, so it's moving perfectly.

We all love distortion ?

Assuming your 'box' induces no colouration and your crossover is somehow flawless in integrating your 2/3 or 4 different drive units.
 

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,615
Likes
3,992
Location
Princeton, Texas
Electrostatics are pricey.

That's pretty much true of new electrostats. Used Quads are often affordable, but some refurbishing (often DIY-able) may be required. I'm not familiar with all of the electrostatic loudspeakers on the market right now, but you can get bookshelf hybrid electrostats for under four grand a pair.

They don't measure well,

Imo this is a generalization that doesn't apply to all electrostats. There are exceptions to most of the generalities that people make about electrostats, just as there are exceptions to most of the generalities that people make about other types of loudspeakers, and very often it is the exceptions that are the most interesting.

but maybe they run under a different set of rules than conventional loudspeakers.

Dipolar electrostats interact with rooms different from the way conventional speakers do, and imo this puts some of the "goal posts" in different places than for conventional speakers:

"The Harman curve relates to conventional forward-firing loudspeaker designs. Legitimate reasons for differences are different loudspeaker directivities - omni, dipoles, etc. - or rooms that are elaborately acoustically treated, or both." - @Floyd Toole
 

Ron Texas

Master Contributor
Forum Donor
Joined
Jun 10, 2018
Messages
6,353
Likes
9,516
@Duke used gear is always more affordable than new gear whether it's electrostatic speakers or anything else. Every generalization has exceptions. On the internet a lot of people will read one's statements so the exceptions will be found. Exceptions don't make the rule.
 

gnarly

Major Contributor
Joined
Jun 15, 2021
Messages
1,042
Likes
1,483
They don't measure well, but maybe they run under a different set of rules than conventional loudspeakers.

I think it's more a matter of that they are hard to measure.....
They are essentially line arrays, both vertically....and horizontally.
Plus, they are dipole.
A little tougher measurement to make, than sticking a mic in front of a small two/three way box.

I either drag them outside or close mic them in the middle of a room.

here's some transfers of Acoustat-X....
Their direct drive tube amps have a HF gain control. I wanted to see what it was doing.
Red trace is HF normal, green full cut.

These are without EQ. It would be brain dead easy to take this speaker to 'ASR approved flat' with minimal EQ. :)
The phase trace plain rocks with finesse, imso.

I've owned these for almost 50 years. They still have a bit of magic left in em.
Might be my "return to speakers", when I finally get so old i have to downsize from my DIY synergys. Lol



acoustat x pair.jpg



acoustat x hf gain compare.JPG
 

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,615
Likes
3,992
Location
Princeton, Texas
Every generalization has exceptions...
Electrostats are a relatively obscure loudspeaker type and many people are unaware of the existence of notable exceptions to the oft-mentioned generalities, especially on this site given that many here are primarily aware that some electrostats have measured poorly and scored poorly in controlled blind testing. Hence my mentioning of the exceptions, or at least of their existence.
 
Last edited:

Ron Texas

Master Contributor
Forum Donor
Joined
Jun 10, 2018
Messages
6,353
Likes
9,516
Electrostats are a relatively obscure loudspeaker type and many people are unaware of the existence of notable exceptions to the oft-mentioned generalities, especially on this site given that many here are primarily aware that some electrostats have measured poorly and scored poorly in controlled blind testing. Hence my mentioning of the exceptions, or at least of their existence.
If they score poorly in blind testing, isn't that the acid test. Maybe you are trying to say the exceptions show there is a better way to build an electrostatic. Still, I feel strongly that exceptions do not disprove a generalization. By the way, your subwoofer swarm looks like a great idea to me.
 
Last edited:

dlaloum

Major Contributor
Joined
Oct 4, 2021
Messages
3,246
Likes
2,507
When I worked at PolyGram (around 1975ish) they had a "special version" of the ESL57 (silver grille). I never really figured out what was special apart from the different grille, but it was quite ubiquitous especially for classical music (Polyhymnia et al), it was also used in quality control and mastering, along with JBL 4315's.

I'm running a pair of (ordinary :)) ESL57's complemented by two DIY 2x12" Ripole subwoofers and they manage to amaze me every time again.
They can do magic that few other speakers can...

But... narrow sweet spot, limited low end (high quality... but limited extension) - and the clincher is that they are delicate... give them a smidge too much voltage, and they arc.... and once they arc, it takes less voltage to arc them next time... a recipe for self destruction!

But I've never heard other speakers sound quite as good.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,946
Likes
38,062
To me the actual frame rigidity is less important than the attachment of the panels to the frame.

Here is the actual aluminium angle (not more rigid T section) that my 989's panels were attached to, 1200mm. long and 1.2mm thick section, you can see where each of the 6 panels were attached. The soft open cell black foam is non sense to me.View attachment 365123

Here is the deflection of this aluminium L section with a carboard box and 9 DVD's.

View attachment 365124

That L section is only held at the top and bottom frame plates.

With the weight of that box and DVD's the aluminium deflects by 4 mm in the centre (less than the weight of 3 of the panels).

The membrane in each panel only has an excursion of around 1mm when playing music, this is (was) the problem to me with the lack of dynamics.

My cat litter filled, steel box section frames, with the panels attached to a 30 x 10 x 2 mm aluminium angle, attached to the frames along their full length solves this problem to me.

When I first played them after the rebuild I feared I had 'broken' them, the reality was that there was so much more detail and dynamics that I had to reduce the cartridge loading on my phono stage by 20 Ohm.
I did use Arcici stands with mine. That helped a little bit with possible flexing. Don't know if it would make the level of difference you describe with your efforts (probably not). It was helpful.
 

Axo1989

Major Contributor
Joined
Jan 9, 2022
Messages
2,940
Likes
2,997
Location
Sydney
If they score poorly in blind testing, issn't that the acid test. Maybe you are trying to say the exceptions show there is a better way to build an electrostatic. Still, I feel strongly that exceptions do not disprove a generalization. By the way, your subwoofer swarm looks like a great idea to me.

In the often-cited comparison published early on by Toole, the detailed discussion highlighted certain caveats and characteristics, notably low preference for the Quads with hard-panned pop material used. The mono vs stereo difference is often misinterpreted. And (in stereo for spatial rating) they were preferred for jazz, and second preference for chamber, if you check all the published graphs. So the picture is more complex than "score poorly in blind testing".

Screenshot 2024-04-22 at 9.29.11 am.png
 
Last edited:

gnarly

Major Contributor
Joined
Jun 15, 2021
Messages
1,042
Likes
1,483
If they score poorly in blind testing, isn't that the acid test.

No sir, a million miles away from an acid test....
Those preference "scores" are simply means of statistical distributions from the population of folks involved.

Even if those scores, for that limited population, are dead on accurate for preferences of the whole world......

....does that mean I've got a bad speaker cause it scores a few standard devs away from the mean preference?
Or I have a bad brain cause it scores a few standards devs away?
Not saying I'm anything special, but I'd prefer to keep my own mellon, than trade it for the mean :)
 
Last edited:

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,615
Likes
3,992
Location
Princeton, Texas
If they score poorly in blind testing, issn't that the acid test.

Is it really as simple as something like "one or two electrostats scored poorly, and we can reliably generalize from there"?

Let's just look at radiation patterns: If one or two moving-coil loudspeakers with "medium width" radiation patterns measure and/or rank poorly, does that really inform us about how a moving-coil speaker with a radiation pattern width three times as wide would fare? How about a moving-coil speaker with a pattern 1/3 as wide? This (and more!) is the range of radiation pattern widths that exist among electrostats. Not to mention all the other variations in execution that exist among electrostats.

In other words, imo it would be a mistake to extrapolate too many generalizations from too few examples.
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,946
Likes
38,062
Is it really as simple as something like "one or two electrostats scored poorly, and we can reliably generalize from there"?

Let's just look at radiation patterns: If a moving-coil loudspeaker with a "medium width" radiation pattern ranks poorly, does that really inform you about how a moving-coil speaker with a radiation pattern width three times as wide would score? How about a moving-coil speaker with a pattern 1/3 as wide? This (and more!) is the range of radiation patterns that exist among electrostats. Not to mention all the other variations in execution that exist among electrostats.

In other words, imo it would be a mistake to extrapolate too many generalizations from too few examples.
I think the only ESLs used in Harman tests were the Quad ESL63 and one of the Martin-Logan hybrids with a woofer. Maybe I'm remembering wrong, but it seems one odd result with the Quad was it did poorly in mono test, but it was used in some of the stereo tests and was closer to other good box speakers. All speakers doing better in stereo blind tests, but the improvement in the Quad was larger.

I see Axo1889 had already posted about this.
 
Last edited:

goryu

Addicted to Fun and Learning
Joined
Oct 4, 2019
Messages
541
Likes
279
@misterdog there good conventional speakers which don't cost a lot. There are no good electrostatics which don't cost a lot. It proves nothing to show that some conventional speakers are expensive. You are trying to pass off an exception as the rule.

A properly designed enclosure has no box sound.

I just picked up a pair of Acoustat's top model, the Spectra 6600's, for $750. And while disassembling for transport, learned that the interfaces had been refurbished 2 years ago. There are incredible bargains available on old Acoustats which seem to just keep playing for decades. Oh, I was planning on spending $7500 on a new pair of Final Acoustics ESL speakers at Axpona but due to the games the distributor and company owner were playing bought the 1/10th as expensive used Acoustats. Happy as can be...
 

Ron Texas

Master Contributor
Forum Donor
Joined
Jun 10, 2018
Messages
6,353
Likes
9,516
Is it really as simple as something like "one or two electrostats scored poorly, and we can reliably generalize from there"?

Let's just look at radiation patterns: If one or two moving-coil loudspeakers with "medium width" radiation patterns measure and/or rank poorly, does that really inform us about how a moving-coil speaker with a radiation pattern width three times as wide would fare? How about a moving-coil speaker with a pattern 1/3 as wide? This (and more!) is the range of radiation pattern widths that exist among electrostats. Not to mention all the other variations in execution that exist among electrostats.

In other words, imo it would be a mistake to extrapolate too many generalizations from too few examples.
You already said the poor measured performance of electrostatics is attributable to the difference in how they interact with the room and they do poorly in blind testing. How does that extend to one or two electrostatics performed poorly? Your final sentence is true, though.
 

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,615
Likes
3,992
Location
Princeton, Texas
I think the only ESLs used in Harman tests were the Quad ESL63 and one of the Martin-Logan hybrids with a woofer.

Those are the only ones I'm aware of.

Hybrid electrostats like the Martin Logan are especially demanding of how they are set up. Here's one of the challenges: Sound pressure level falls off at 6 dB per doubling of distance from a point-source (which the woofer approximates), but at only 3 dB per doubling of distance from a line source (which the panel approximates). So listening distance matters, and arguably the ability to adjust the level of the woofer relative to the panel can "make or break it" with a hybrid electrostat. Personally, I would not have chosen the Martin Logan Quest (I think that was the model) as ambassador for electrostats.

Ime electrostats benefit from making good use of their backwave energy, which imo implies that their backwave energy's spectrum not be significantly altered; that it arrives neither too soon nor too late; and that the reflections be neither too loud not too soft. Imo this is just the nature of the beast. And I realize that others have reached different conclusions about what to do with the backwave energy.

I have owned both Martin Logan Quests and Quad "USA Monitors", the latter being a version of the ESL63. I didn't have much test equipment back then (30 years ago), but to my ears the Quads had a couple of tonal balance issues. I spent a lot of time and money trying to remedy those issues rather than quickly moving on to something else because imo what they did well, they did very well. I finally got good tonal balance largely thanks to an equalizer that cost almost as much as the rest of my system combined. Eventually I started looking for a more "elegant" solution, and went in a different direction.
 
Last edited:

MattHooper

Master Contributor
Forum Donor
Joined
Jan 27, 2019
Messages
7,505
Likes
12,660
I have owned both Martin Logan Quests and Quad "USA Monitors", the latter being a version of the ESL63. I didn't have much test equipment back then (30 years ago), but to my ears the Quads had a couple of tonal balance issues. I spent a lot of time and money trying to remedy those issues rather than quickly moving because imo what they did well, they did very well.

I used my ESL 63s with a CJ MV55 tube amp. Always sounded glorious in my room. Maybe it acted as an eq :)

I finally got good tonal balance largely thanks to an equalizer that cost almost as much as the rest of my system combined. Eventually I started looking for a more "elegant" solution, and went in a different direction.
I'm curious which one. Cello Palette? Z-Systems RDP-1 digital EQ? Just guessing at expensive EQs back in the day...
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,946
Likes
38,062
I used my ESL 63s with a CJ MV55 tube amp. Always sounded glorious in my room. Maybe it acted as an eq :)


I'm curious which one. Cello Palette? Z-Systems RDP-1 digital EQ? Just guessing at expensive EQs back in the day...
Yes, the CJ was a good match. I used a CJ MV50 for a time with my Quads. It was an EQ. Bass was up an amount measurable, and borderline audible. Treble was down a bit (upper mid-range, lower treble was a touch hot on 63s imo). Simply due to the output impedance. You could try a 1 ohm power resistor on the output of an SS amp sometime.
 

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,615
Likes
3,992
Location
Princeton, Texas
You already said the poor measured performance of electrostatics is attributable to the difference in how they interact with the room and they do poorly in blind testing. How does that extend to one or two electrostatics performed poorly?

Excellent question!

First let me say that imo we do not have enough data to generalize that "electrostats do poorly in blind testing." I don't think that's quite what I said, and if I did, I mis-spoke. Imo they do have set-up requirements that tend to not be met in the testing set-ups I've seen described.

By way of example that can illustrate some of the relevant issues, let's look at the Sanders Sound Model 10. This is a hybrid electrostat which combines a flat electrostatic panel with a direct radiator woofer, and it has separate amplification (and equalization) for woofer and panel. Imo this is a superb example of a hybrid electrostat.

See my preceding post for a description of the issue that arises when we combine a point-source-approximating woofer with a line-source-approximating panel. If we adjust for optimum tonal balance at the one-meter microphone location for best-looking measurements, the panels are going to be about 4 dB louder than the woofer at normal listening distances. So we have to choose between measurements that look good, and a tonal balance that sounds good at the listening position.

I would want to position the Sanders Sound speakers ballpark five feet in front of the wall, and I would not want anything to significantly alter the spectral balance of the backwave. So, no curtains or drapes behind them thick enough to attenuate the highs.

One of the main selling points of the Sanders Sound Model 10 is its incredible imaging. Would that quality come through decisively in a single-speaker blind test? I don't know.

(My recollection from eyeballing the Martin Logan's measured response years ago was that its woofer was already too low in SPL at one meter, and this discrepancy would have been even worse at a normal listening distance. Perhaps the bass could be boosted sufficently by placing them close to the wall or in a corner, but this would have reduced the path-length-induced time delay for the panel's backwave too much. Imo that particular electrostat was not a good ambassador for the breed.)

I used my ESL 63s with a CJ MV55 tube amp. Always sounded glorious in my room. Maybe it acted as an eq :)

I used a CJ amp with four EL34 tubes, I think it was a predecessor of the MV55. Was there an MV45? If not, then it was the MV50. Also tried a Clayton Audio 40-watt Class A solid state amp, and (blasphemy alert) I preferred the Clayton.

I'm curious which one. Cello Palette?

Yup! I listened to Sennheiser headphones and twiddled the knobs until the tonal balance was similar, then backed off on the bottom-end knob(s) a little because the warmth with that tonal balance was a bit much.
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,946
Likes
38,062
Excellent question!

First let me say that imo we do not have enough data to generalize that "electrostats do poorly in blind testing." I don't think that's quite what I said, and if I did, I mis-spoke. Imo they do have set-up requirements that tend to not be met in the testing set-ups I've seen described.

By way of example that can illustrate some of the relevant issues, let's look at the Sanders Sound Model 10. This is a hybrid electrostat which combines a flat electrostatic panel with a direct radiator woofer, and it has separate amplification (and equalization) for woofer and panel. Imo this is a superb example of a hybrid electrostat.

See my preceding post for a description of the issue that arises when we combine a point-source-approximating woofer with a line-source-approximating panel. If we adjust for optimum tonal balance at the one-meter microphone location for best-looking measurements, the panels are going to be about 4 dB louder than the woofer at normal listening distances. So we have to choose between measurements that look good, and a tonal balance that sounds good at the listening position.

I would want to position the Sanders Sound speakers ballpark five feet in front of the wall, and I would not want anything to significantly alter the spectral balance of the backwave. So, no curtains or drapes behind them thick enough to attenuate the highs.

One of the main selling points of the Sanders Sound Model 10 is its incredible imaging. Would that quality come through decisively in a single-speaker blind test? I don't know.

(My recollection from eyeballing the Martin Logan's measured response years ago was that its woofer was already too low in SPL at one meter, and this discrepancy would have been even worse at a normal listening distance. Perhaps the bass could be boosted sufficently by placing them close to the wall or in a corner, but this would have reduced the path-length-induced time delay for the panel's backwave too much. Imo that particular electrostat was not a good ambassador for the breed.)



I used a CJ amp with four EL34 tubes, I think it was a predecessor of the MV55. Was there an MV45? If not, then it was the MV50. Also tried a Clayton Audio 40-watt Class A solid state amp, and (blasphemy alert) I preferred the Clayton.
Yup! I listened to Sennheiser headphones and twiddled the knobs until the tonal balance was similar, then backed off on the bottom-end knob(s) a little because the warmth with that tonal balance was a bit much.
MV45 was C-J's initial EL34 amp and the MV50 was an upgraded version after a few years. The MV75 was their 6550 based amp.

The 45 and 45a had a cage on top. The MV50 and MV75 had a slab of brushed gold anodized aluminum on the entire front with a cage behind it over the tubes. Also had rack mount handles on the front which the 45 lacked.
 

Ron Texas

Master Contributor
Forum Donor
Joined
Jun 10, 2018
Messages
6,353
Likes
9,516
@Duke interesting discussion.
 
Top Bottom