I just did an exhaustive search of the web for any sort of hard science on electrostatic speakers, and except for a few blogs of variable quality, the only thing I consistently hit was an advertising blurb from Martin-Logan. For obvious reasons, I don't think their take on the subject is unbiased.
What I'm curious about is whether there has been a decent write-up anywhere talking about some of the claims made about electrostatics, and any basis in fact associated with those claims. One thing which comes up with maddening regularity is the so-called "massless" nature of the driver. Obviously, this is pure hype because even the thinnest, most flimsy LDPE or Mylar films have a density of tens to hundreds of times that of normal atmospheric air. Combined with the fact that the force behind the ESL "motor" is lower than that of a typical electrodynamic driver, some of the claims about the system seem questionable. There's no doubt that a good speaker can be made with the technology, but the hype seems to over-reach the actuality.
It seems that so many people repeat the myths, any research, design or evaluation information has been buried under them. Does anyone know of a good unbiased treatise on electrostatics which addresses some of the pros as well as cons of the type, and backs them up with measurements and logic?
A good start:
https://en.wikipedia.org/wiki/Electrostatic_loudspeaker
It's as valid a speaker design as any. My experience is that they need a big room and lots of space around them. They don't sound good against a wall or in a corner. Often best served by a high-current amp. Probably need a sub to get that bottom octave, though M-L does make some nice hybrids.
Check other manufacturers (JansZen, Sound Lab) and read some reviews. No substitute for hearing a pair yourself.
I haven't ever heard electrostatics, but I did once get to hear some ribbon loudspeakers, which have some aspects in common as regards how you set them up in a room. They were Apogees (either Centaurs or Centaur Minors, I forget), and made some very nice sounds. However, like (most) Martin-Logans, their cone woofer/sub was part of the same cabinet as the mid/tweeter, which doesn't necessarily always make for perfect placement of either element.
If you want proper massless drivers, look up ionophone/plasma speakers... and all the problems that come with trying to operate them effectively while staying sane and un-electrocuted.
There is some good info in the Wiki article; I was just hoping for more direct details and numbers. What it has is fairly vague, though at least it sort of addresses a few things like air load. I have heard a fair number of electrostatics in my travels, and the main problem seems to be building large panels in such a manner that you don't get excessive amounts of comb filtering (which is a characteristic of the Martin-Logans, among others). The low mass is definitely good in some ways, although no one seems to talk about the limitation - which is that they can easily be over-damped if the electrostatic force generated is insufficient. But numbers are hard to find, and I was hoping someone had done the math already and shared it.
Actually, you want a high voltage amp for electrostatics, since current demands by the driver are (or should be) minimal; if a transformer is used, in fact, stepping up the voltage to the level required by the diaphragms reduces the current delivered to the speaker by a comparable amount. The problem is usually the reactive load causing heating in the output devices, and a higher current amp is more likely to be designed to tolerate that or at least have better protection against reactive loads, such as a Zobel network.
And it's not so much that I'm looking for a massless speaker as hoping to cut through all the audiophool and marketing gibberish about electrostatics. They definitely have advantages, but also sonic disadvantages, and there don't seem to be many people willing to call them out on that. Mostly it's manufacturers trying to sell them.
ESLs are tricky.
In most cases, the impedance varies from a high in the bass to a low in the highs by about a factor of 9 or 10 to 1. They are fundamentally different from a speaker in a box, whose impedance curve is a pretty good map of its efficiency (hence in many cases best driven by a voltage source for flat frequency response). The impedance curve of an ESL is not based on the resonance of a driver in a box; essentially its based on a capacitor; so if one is to put out a given x sound pressure at 30Hz, the speaker needs the same power to do that as it does at 10KHz.
This means that ESL manufacturers have to resort to some tricks depending on what part of the market they cater to. Martin Logan for example keeps their impedances low- about 4 ohms in the bass, resulting in about 0.5 ohms at 15 or 20KHz. The latter impedance is so low that the speaker cable often becomes a factor in the driving impedance, which I suspect is by intention, in order to limit the current that the amp makes so as not to be as bright. If the amp driving the speaker is a voltage source and is able to act that way when driving a Martin Logan, you can see that there will be too much energy at high frequencies. Since a lot of solid state amps run into troubles with these impedances they get away with it; its helpful that there is not a lot of energy up there.
Sound Labs, at the other extreme (and capable of true 20-20KHz operation) are as much as 30 ohms in the bass, and to drive them with solid state takes a rather large amp despite the speaker being relatively efficient (you can see that a mic placed at 1 meter from the speaker will not pick up all that the driver radiates; for such speakers a more accurate value will be the measured sensitivity + 6db, which is experienced at a greater distance from the speaker). For example a 600 watt amp can only make about 150 watts in the bass on a Sound Lab- and not surprisingly, a lot of solid state amps tend to be bass shy on this speaker. Tube amps on the other hand can make power into loads like that so its not uncommon to see a 150 or 200 watt tube amp keep up easily with a 600 watt amp in this case.
This is not the case with ML and a number of others as their impedance does not favor tube amps at all. With the higher impedance ESLs it is common knowledge that you need an amp that can make voltage. At this link you can view the impedance curve of the old ESL57 (2nd post):
http://audiokarma.org/forums/index.php?threads/quad-esl-57s-destroyed-a-carver-tfm-55x-amplifier.785429/
With most speakers employing a permanent magnet, the magnetic field can sag a bit as current is put through the voice coil. According to a friend in the industry (who makes drivers for a lot of OEMs in the US) Alnico, while the weakest magnet used, also sags the least so has a certain following amongst audiophiles. An ESL's motive force is unaffected by the amplifier's current to move the diaphragm (since it derives from a high voltage power supply), and so the argument is that the driver is thus faster and lower distortion. Its certainly not about the diaphragm being as light as air, although because of their size, they don't need much excursion, like any planar.
The old Quads made by Peter Walker were reputed to be one of the lowest distortion speakers made for many years.
Some ESLs are quite fragile- the old Quad ESL57 could only handle about 35 watts and if you drove it too hard it would arc and damage the panel. Modern Quads have protection circuits. Overall ESLs can enjoy quite a loyal following so its common to hear about their owners going through quite a bit of hell to keep them alive and playing.
I got a pair of the less expensive ML's that I got a good deal for years ago. The thing that most jumped out at me is not exactly the sound per se, which is just fine, but they have very narrow dispersion, so much so that moving your head just above the speaker panel's height, at fairly far away, will drop off treble noticeably. Just moving your head a few inches up and down you can easily tell the difference. The base even is angled like a wedge, and you can rotate it 180° to angle the speaker upwards a bit because of this. So, if I'm walking around the apartment, I don't use those speakers for music, they're my home theater mains.
Also the sound doesn't come from a point like regular tweeter speakers, it "shoots" from the whole surface of the panel, so I think (and others can correct me if I'm wrong) that volume should fall off with distance differently than a regular speaker.
http://www.princeton.edu/3D3A/Directivity/Essence%20Electrostatic%20Model%201600/index_H.html (http://www.princeton.edu/3D3A/Directivity/Essence%20Electrostatic%20Model%201600/index_H.html)
(http://www.princeton.edu/3D3A/Directivity/Essence%20Electrostatic%20Model%201600/images/Plots/Horizontal/Essence%20Electrostatic%20Model%201600%20H%20Polar%20Plot.png)
Placement in the corner? My Staxes do not sound too impressive when I put them i the corner, no. (To those who didn't catch it: they are headphones ;-) )
So obviously I am not going to fill this post with the hard science or measurements the OP wanted (and by the way I think the Wikipedia article suffers from even more [citation needed]'s than is even indicated in the text). But this part:
One thing which comes up with maddening regularity is the so-called "massless" nature of the driver. Obviously, this is pure hype because even the thinnest, most flimsy LDPE or Mylar films have a density of tens to hundreds of times that of normal atmospheric air. Combined with the fact that the force behind the ESL "motor" is lower than that of a typical electrodynamic driver, some of the claims about the system seem questionable.
Is it really so relevant to directly compare mass density to that of air? One thing is "density" vs "total mass", I mean, if you had the same mass but just made it thinner, you would inflate density, and ... that would not matter to the argument, ceteris paribus.
And it is not the heaviness of mass either - you wish to accelerate the driver [film|cone], so it would be inertia? But not directly compared to air; if one attempted to ask whether the inertia of cone / inertia of air relation is troublesome, and the inertia of film / inertia of air is small enough for that problem to be practically solved - then one would make the error of presuming that we want to
move a lot of air as if it were a rigid body. (Gases are not - air waves are generating by attempting to
compress it, right?)
Connected to that: a cone is "driven" where it is glued to the voice coil, and relies on sufficient rigidity for the surface. Electrostatic film is certainly not much rigid, and is driven everywhere.
But of course, this consideration is a matter magnitude that should be measured: if cone speakers manufacturers have solved the rigidity/inertia trade-off to an insignificant order of magnitude, then the return on removing the issue would not be too great.
As a Quad 57 owner, I found these speakers to be the best I've heard for acoustic music. They lack power and do better in smaller rooms and have a small "sweet spot". If the felt from the rear of the panels is removed, they are louder, but they then need to be placed further from walls or corners to maintain the bass. The midrange seems lifelike, subjectivily speaking. As for the mass, my unprofessional understanding is that when the panel is electrically charged, the mass / force ratio is the best of any driver, however, I have no reference for this info.
As well, I have had to replace arced tweeter panels on several occasions.
I got a pair of the less expensive ML's that I got a good deal for years ago. The thing that most jumped out at me is not exactly the sound per se, which is just fine, but they have very narrow dispersion, so much so that moving your head just above the speaker panel's height, at fairly far away, will drop off treble noticeably. Just moving your head a few inches up and down you can easily tell the difference. The base even is angled like a wedge, and you can rotate it 180° to angle the speaker upwards a bit because of this. So, if I'm walking around the apartment, I don't use those speakers for music, they're my home theater mains.
Also the sound doesn't come from a point like regular tweeter speakers, it "shoots" from the whole surface of the panel, so I think (and others can correct me if I'm wrong) that volume should fall off with distance differently than a regular speaker.
The sound pressure falls off slower.
Some ESLs have curved panels (like the Quads) and so have less of that 'head inna vice' thing going on. The Sound Labs are so tall its a non issue; they are curved though so the radiation pattern is fairly wide.
"Why the Quad ESL is able to do so is simply because of the moving part of the Quad ESL, the electrostatic diaphragm, which is a thin layer of stretched Mylar, ten times thinner than a human hair and so light that it approaches the mass of air to which it is coupled." - http://quad-hifi.co.uk/product.php?cid=5
Also the sound doesn't come from a point like regular tweeter speakers, it "shoots" from the whole surface of the panel, so I think (and others can correct me if I'm wrong) that volume should fall off with distance differently than a regular speaker.
I'd guess that a speaker of the shape of a Martin-Logan might have the same properties as a more conventionally-driven "vertical line array" speaker like the ones made by Townshend, Scaena. or even the Grateful Dead's legendarily humungous 'Wall of Sound'. At least when it comes to volume/distance matters.
Placement in the corner? My Staxes do not sound too impressive when I put them i the corner, no. (To those who didn't catch it: they are headphones ;-) )
So obviously I am not going to fill this post with the hard science or measurements the OP wanted (and by the way I think the Wikipedia article suffers from even more [citation needed]'s than is even indicated in the text).
Yeah, it reads like sort of vague advertising copy - very few of the assertions made in the article are backed up with facts.
1338]One thing which comes up with maddening regularity is the so-called "massless" nature of the driver. Obviously, this is pure hype because even the thinnest, most flimsy LDPE or Mylar films have a density of tens to hundreds of times that of normal atmospheric air. Combined with the fact that the force behind the ESL "motor" is lower than that of a typical electrodynamic driver, some of the claims about the system seem questionable.
Is it really so relevant to directly compare mass density to that of air? One thing is "density" vs "total mass", I mean, if you had the same mass but just made it thinner, you would inflate density, and ... that would not matter to the argument, ceteris paribus.
And it is not the heaviness of mass either - you wish to accelerate the driver [film|cone], so it would be inertia? But not directly compared to air; if one attempted to ask whether the inertia of cone / inertia of air relation is troublesome, and the inertia of film / inertia of air is small enough for that problem to be practically solved - then one would make the error of presuming that we want to
move a lot of air as if it were a rigid body. (Gases are not - air waves are generating by attempting to
compress it, right?)
Connected to that: a cone is "driven" where it is glued to the voice coil, and relies on sufficient rigidity for the surface. Electrostatic film is certainly not much rigid, and is driven everywhere.
But of course, this consideration is a matter magnitude that should be measured: if cone speakers manufacturers have solved the rigidity/inertia trade-off to an insignificant order of magnitude, then the return on removing the issue would not be too great.
[/quote]
Right; that's part of what I'm trying to determine. You're moving a larger area but with (I suspect) much less force, and linearly over the vast majority of the surface but with a fraction of the excursion. With a typical cone, the mass ratio is higher and the force much greater, but you have to have a greater excursion to achieve the same output level. I was hoping to find some numbers to put that into perspective, so I could see whether the lower mass was a significant advantage or whether the ratios still came out similar.
I've heard some cone-type speakers which do a great job of reproducing transients and have the same sort of "hear-through" illusion that an electrostatic provides, so I have long felt it's not the transducer per se, but a combination of good drivers and crossover design (of course) plus other factors: room interaction, reflection/diffraction, and radiation pattern - all of which have common progenitors in the mechanical layout of a speaker but are usually not all well-behaved.
You can see in AJ's graph that the speaker represented has a lot of comb filtering at higher frequencies, and I'll bet there's a sort of "venetian-blind" effect from moving the listening position - plus probably timbral changes since the different frequencies don't change similarly at the different angles. OTOH, they indicate a dipolar pattern which could be beneficial to eliminating room modes if positioned carefully.
http://www.princeton.edu/3D3A/Directivity/Essence%20Electrostatic%20Model%201600/index_H.html (http://www.princeton.edu/3D3A/Directivity/Essence%20Electrostatic%20Model%201600/index_H.html)
(http://www.princeton.edu/3D3A/Directivity/Essence%20Electrostatic%20Model%201600/images/Plots/Horizontal/Essence%20Electrostatic%20Model%201600%20H%20Polar%20Plot.png)
What spoke to me is this plot for the same speaker:
(http://www.princeton.edu/3D3A/Directivity/Essence%20Electrostatic%20Model%201600/images/Plots/Horizontal/Essence%20Electrostatic%20Model%201600%20H%20Contour%20Plot.png)
IME this is not atypical.
What I see is a speaker that is the opposite of something that approximates constant directivity. It's beamy at high frequencies and compounds that failing with a lot of side lobes.
You can see in AJ's graph that the speaker represented has a lot of comb filtering at higher frequencies, and I'll bet there's a sort of "venetian-blind" effect from moving the listening position - plus probably timbral changes since the different frequencies don't change similarly at the different angles. OTOH, they indicate a dipolar pattern which could be beneficial to eliminating room modes if positioned carefully.
Well, that's a good start, as I saw folks "guessing" and posting conjecture even after I posted that link. If you go to it, you'll find far more data (there is a <> nav bar above the graphs) on this and a couple other panel/stat type speakers.
There are no mysteries for non-audiophile folks like me. What you see is what you get. The polar pattern is indeed chaotic at best and quite different across the spectrum. Many, with rare exceptions like the huge Soundlabs, are actually dipolar panels atop monopole boxes. There is no question there are alluring aspects to the kind of soundfield they generate, especially to untrained ears, hence their popularity amongst a subset of audiophiles. The chaotic nature of the HF is actually quite ideal
for the rear radiation. That part it good! (even if listeners have no clue why).
An ideal speaker for generating "realistic" soundfields would have indirect radiation that is highly diffuse, i.e lacking "leading edge" wavefront info for the ears to have directional cues to lock onto.
Unfortunately, stats also radiate this way forwards, with the direct field.
"Why the Quad ESL is able to do so is simply because of the moving part of the Quad ESL, the electrostatic diaphragm, which is a thin layer of stretched Mylar, ten times thinner than a human hair and so light that it approaches the mass of air to which it is coupled." - http://quad-hifi.co.uk/product.php?cid=5
From a scientific veiewpoint this is meaningless, until the repeal of the Law of Newton we often refer to as F = MA. This means that you can manage high mass quite effectively if you can come up with a correspondingly high force.
Of all common formats of loudspeaker, electrostats have a track record for having problematically limited sources of force.
The simple fact is that obtainin lots of dynamic range is no problem for conventional designs. but are mission impossible for 'stat's. If diaphragm acceleration were the most important thing, compression drivers would rule.
I have been listening to music/tv via an old pair of ML (Aeos I think) purchased second hand around 15 years ago and am happy with the sound which has somewhat improved since I moved a large piece of furniture from the lounge room. As with many I have to share my "listening space with the furniture and items of everyday living plus there is a kitchen off to the right which makes the space asymmetrical. But having said all that, when I do get the chance to arrange the room and listen from a near centre position I am still pleased with the SQ on a non-scientific basis. What I would like to note is that the membrane in the panel is not stretched and had thought about this for quite a while as to how it could possibly be stretched across a panel that is curved without touching and being supported by the panel itself. I pulled one of the speakers apart about a year ago for minor repairs and found the front and rear panel that supports the membrane is separated by what is now deteriorating double sided tape and a thin strip of rubber and it did confirm my thoughts that it is indeed not stretched at all.
The best study about electrostatic loudspeakers is the chapter written by PJ Baxandall in Loudspeaker and Headphone Handbook by John Borwick. (https://books.google.fr/books?id=lPd2OucRfu8C&pg=PT5&hl=fr&source=gbs_selected_pages&cad=2#v=onepage&q&f=false)
Theoretically, a constant charge push-pull electrostatic (Hunt's model) is a linear systems with no distortion for any displacement (contrary to electrodynamics, ribbon, BA,...)
But the major problem of electrostats is due to the big membrane surface that gives uneven radiation.
Quad ESL63 was an interesting step to lessen directivity problems due to radiation.
I have been listening to music/tv via an old pair of ML (Aeos I think) purchased second hand around 15 years ago and am happy with the sound which has somewhat improved since I moved a large piece of furniture from the lounge room. As with many I have to share my "listening space with the furniture and items of everyday living plus there is a kitchen off to the right which makes the space asymmetrical. But having said all that, when I do get the chance to arrange the room and listen from a near centre position I am still pleased with the SQ on a non-scientific basis. What I would like to note is that the membrane in the panel is not stretched and had thought about this for quite a while as to how it could possibly be stretched across a panel that is curved without touching and being supported by the panel itself. I pulled one of the speakers apart about a year ago for minor repairs and found the front and rear panel that supports the membrane is separated by what is now deteriorating double sided tape and a thin strip of rubber and it did confirm my thoughts that it is indeed not stretched at all.
The way M-L makes their panels curved is to turn them into isolated horizontal strips by running the supports across the panel at intervals; this supports the diaphragm in segments and if the intervals are small enough, allows the segments to be curved without collapsing. I'm sure there is tension on the diaphragm before it is mated with the isolation strips, though it may not be as extreme as that for a single panel. The tension is what determines the resonant frequency of the diaphragm, and you would want it either well above or below the active bandwidth of the panel.
I have never measured one with a calibrated mic up close, but it wouldn't surprise me if the many in-phase segments are what interfere to cause the rather bumpy frequency response (and vertical lobing) of a M-L panel.
I once auditioned a pair of Sanders electrostatic speakers, properly set up in a home. The midrange, in particular, was breathtaking.
The interview with John Sanders below is quite enlighting, along with the many technical papers he has written on this subject that are available on his website. What strikes me about the interview is that he seems very clued on audio engineering, dismissing a lot of audiophile myths.
http://www.monoandstereo.com/2013/11/interview-with-roger-sanders.html
What strikes me about the interview is that he seems very clued on audio engineering, dismissing a lot of audiophile myths.
The reason that electrostatic speakers are so good is that they are the only type of midrange driver that has essentially no moving mass. Magnetic speakers simply cannot match the performance of electrostatics in the midrange because they are heavy so cannot be accelerated quickly and accurately at treble frequencies.
::)
I was thinking more about his general discussion around digital media, amplifiers, measurements, double blind vs sighted testing and so on.
The reason that electrostatic speakers are so good is that they are the only type of midrange driver that has essentially no moving mass. Magnetic speakers simply cannot match the performance of electrostatics in the midrange because they are heavy so cannot be accelerated quickly and accurately at treble frequencies.
::)
Can you please elaborate about the math ? Arny already stated F = MA. This Wikipedia article (https://en.wikipedia.org/wiki/Electrostatic_loudspeaker) states:
...typical dynamic speaker drivers can have moving masses of tens or hundreds of grams whereas an electrostatic membrane only weighs a few milligrams, several times less than the very lightest of electrodynamic tweeters. The concomitant air load, often insignificant in dynamic speakers, is usually tens of grams because of the large coupling surface, this contributing to damping of resonance buildup by the air itself to a significant, though not complete, degree.
According to this Stereophile article (https://www.stereophile.com/content/quad-esl-63-loudspeaker-martin-colloms) the mass of a Quad ESL63 diaphragm was 3 mg.
How relevant is the driver mass in the comparison between speaker types ?
The reason that electrostatic speakers are so good is that they are the only type of midrange driver that has essentially no moving mass. Magnetic speakers simply cannot match the performance of electrostatics in the midrange because they are heavy so cannot be accelerated quickly and accurately at treble frequencies.
Can you please elaborate about the math ?
He repeats an audiophile myth, not math. There is no "performance/accuracy"advantage to stats in the midrange due to F=MA, that is pure nonsense. The difference in sound is due to many other real factors, like polar radiation pattern, subsequent reflections, lack of or increased resonances, etc, etc, etc....that have zero to do with diaphragm mass.
How relevant is the driver mass in the comparison between speaker types ?
Zero, unless pathological.
The "mass" of an ESL diaphragm has zero to do with why he and audiophile believers prefer their mids...or anything else.
Btw, the Mms of my Audax HM100zo is 2.5g. But the concentration of magnetic strength around the coil is higher than a ESL like the Quad, so the radiation efficiency is higher. But again, this is largely irrelevant info to soundwaves>ears...unless a believer.
cheers,
AJ
I was thinking more about his general discussion around digital media, amplifiers, measurements, double blind vs sighted testing and so on.
Yes, he seems to be spot on there. It was just ironic that he repeated an audiophile myth about driver mass right off the bat. That claim is pure nonsense.
Can you please elaborate about the math ? Arny already stated F = MA
So what? It's a linear equation. If you have two speakers producing (roughly) the same sound and at the same volume, they've both got to be imparting similar forces on the air. I could be wrong but I imagine this is part of why electrostatics have poor bass, in order to achieve the same force in that range, A would have to be quite large, and it's going to be limited by the materials and presumably a bunch of other factors.
I was thinking more about his general discussion around digital media, amplifiers, measurements, double blind vs sighted testing and so on.
Yes, he seems to be spot on there. It was just ironic that he repeated an audiophile myth about driver mass right off the bat. That claim is pure nonsense.
As I read my post, I said pretty much what you did, just up thread. I read your post as debunking the myth that minimizing diaphragm mass gives an insurmountable advantage. I agree. and said exactly that in my way with the comment about F=MA. Your comments about differences in radiation pattern were addressed by my posting of some radiation patterns for 'stats. Are we agreeing noisily or what?
Can you please elaborate about the math ? Arny already stated F = MA
So what? It's a linear equation. If you have two speakers producing (roughly) the same sound and at the same volume, they've both got to be imparting similar forces on the air. I could be wrong but I imagine this is part of why electrostatics have poor bass, in order to achieve the same force in that range, A would have to be quite large, and it's going to be limited by the materials and presumably a bunch of other factors.
Since people are not getting my intent. Please let me expand on the mass issue. Let's say that we have two speaker diaphragms, one with low mass and one with high mass. Either can be accelerated as desired by providing sufficient acceleration, or force. Either can have the same rise time or impulse response if enough force is available. Coming up with enough force is simply an engineering problem, subject to the inherent properties of the materials involved.
Electrostatic drive in practice tends to be limited in terms of available force. Electromagnetic drive is also limited, but in the real world we observe that it is common to have far, far higher amounts of force available with electrodynamic speakers. The ready availability of cheap good clean amplifier power helps. This shows up as much higher dynamic range. The ESL63 addresses some of the directionality issues with a novel diaphragm design but it is brutally limited on the dynamic range issue. The highest dynamic range drivers are based on the so-called electrodynamic compression driver that is actually limited by the nonlinear compression of air at the extremes.
To expand on the electrostatic dynamic range limitations. Electrostatic diaphragms are like any diaphragm, they need to have working room to move. In an electrostatic, one obtains that working room by increasing the spacing of the diaphragm and drive elements. But increasing the spacing decreases the electrostatic field gradient and thus decreases efficiency. You can increase the electrostatic field gradient by applying more voltage but above about 10,000 volts you start ionizing the atmosphere that the speaker is operating in.
There was one attempt that I know of to address this problem of air ionization by putting the whole speaker into a gas chamber, and this was the legendary Dayton-Wright. I don't know why they went out of business, but I'm guessing that the insulating gas (Sulfur Hexaflouride?) leaked away and when the speaker turned itself back into an air-dielectric speaker, it failed fatally.
Reference: https://en.wikipedia.org/wiki/Dielectric_gas
It is possible that since one of the consequences of arcing in Sulfur Hexafluoride can be the generation of Phosgene, a well-known highly poisonous gas with possible military significance, safety concerns may have killed the Dayton Wright.
The "problem" about "F = ma" ("in singularis", just one equation and just one force) is the simplification to one representative body. For a dynamic-speaker cone, the forces at the outer rubber suspension are not the same as the forces at the point it is glued to the voice coil unless the cone is rigid. For an electrostatic, the idealized model would be expected to be much closer to the truth. If that matters.
To get from F = ma to something more realistic like a(x) = F(dx) / m(dx) (where x = location on the diaphragm) and then aggregating up, isn't witchcraft, but it surely isn't handwaving. And to get from there to actual sound at listening position? Not "sound or not", but "better or worse"? If you pretend that you are doing science, then don't show me the my high school spherical cow in the vacuum. Show me model-predicted reproduction vs actual measurement.
Are we agreeing noisily or what?
I guess, since I'm talking about
Sanders statements...not yours that Kees quoted.
Are we agreeing noisily or what?
I guess, since I'm talking about Sanders statements...not yours that Kees quoted.
Glad you helped me with that!
The ready availability of cheap good clean amplifier power helps. This shows up as much higher dynamic range. The ESL63 addresses some of the directionality issues with a novel diaphragm design but it is brutally limited on the dynamic range issue.
The Sound Labs seem to solve this problem. They run pretty high voltages (I seem to recall about 10KV) and for many years they did have arcing problems. That seems to be solved with a new diaphragm material they call the 'PX technology' .
They do need a bit more power, but with 100 watts on them they seem to fill an average room easily enough.
There is no "performance/accuracy"advantage to stats in the midrange due to F=MA, that is pure nonsense.
It's not that I don't believe you, but I was actually hoping for some links to papers, or formulas that show that the 800x mass difference (3mg ESL 63 diaphragm vs a 2.5g typical driver like your Audax Mms) is non significant.
There is no "performance/accuracy"advantage to stats in the midrange due to F=MA, that is pure nonsense.
It's not that I don't believe you, but I was actually hoping for some links to papers, or formulas that show that the 800x mass difference (3mg ESL 63 diaphragm vs a 2.5g typical driver like your Audax Mms) is non significant.
Kees, here is the quote again:
Magnetic speakers simply cannot match the performance of electrostatics in the midrange because they are heavy so cannot be accelerated quickly and accurately at treble frequencies.
There is nothing for me to link or show. It's up to Sanders to back that claim that "Magnetic" speakers (he probably means moving coil/dynamic) can't match the "performance" of electrostats due to diaphragm mass.
What "performance"?
There is no "performance/accuracy"advantage to stats in the midrange due to F=MA, that is pure nonsense.
It's not that I don't believe you, but I was actually hoping for some links to papers, or formulas that show that the 800x mass difference (3mg ESL 63 diaphragm vs a 2.5g typical driver like your Audax Mms) is non significant.
Barring a reliable source substantiating that the entire ESL63 diaphragm weighs a mere 3 mg, I'll call BS! I searched, but I found no such thing. 3.5 mg per square inch, I'll believe. ;-)
Barring a reliable source substantiating that the entire ESL63 diaphragm weighs a mere 3 mg, I'll call BS! I searched, but I found no such thing. 3.5 mg per square inch, I'll believe. ;-)
Yeah, 3.5 mg per square inch, that matches very well - assuming Stereophile's thickness figure is off by a factor of 10 (and I manage to get the number of zeroes right!).
According to various sources (e.g., like this (http://www.estatsolution.com/elements.html)) the ESL63 uses a 3.5 µm thick mylar film. (Stereophile's figure translates to ten times as much). Googling a couple of sources for the weight of mylar:
http://www.pauzuolis-rc.com/rc-shop/building-materials/mylar-film/mylar-film-3-mk
http://www.homefly.com/reference/Covering%20Weights.htm
A 3.5 µm mylar film would be around 5 grams per square meter. ~5000 mg/m^2 * (2.54/100)^2 (in/m)^2 translates to ~3.2 mg/square inch. (Edit: corrected m and µ this line.)
Barring a reliable source substantiating that the entire ESL63 diaphragm weighs a mere 3 mg, I'll call BS! I searched, but I found no such thing. 3.5 mg per square inch, I'll believe. ;-)
Yeah, 3.5 mg per square inch, that matches very well - assuming Stereophile's thickness figure is off by a factor of 10 (and I manage to get the number of zeroes right!).
According to various sources (e.g., like this (http://www.estatsolution.com/elements.html)) the ESL63 uses a 3.5 µm thick mylar film. (Stereophile's figure translates to ten times as much). Googling a couple of sources for the weight of mylar:
http://www.pauzuolis-rc.com/rc-shop/building-materials/mylar-film/mylar-film-3-mk
http://www.homefly.com/reference/Covering%20Weights.htm
A 3.5 µm mylar film would be around 5 grams per square meter. ~5000 mg/m^2 * (2.54/100)^2 (in/m)^2 translates to ~3.2 mg/square inch. (Edit: corrected m and µ this line.)
In comparison, the mass of the diaphragm of a candidate tweeter is given as 0.21 grams.
https://www.madisoundspeakerstore.com/ring-radiator-tweeters/vifa-ot19nc00-2/3-ring-radiator-tweeter-4-ohm/
In comparison, the mass of the diaphragm of a candidate tweeter is given as 0.21 grams.
About 86 times as much per area unit, in order to keep the diaphragm rigid. But that says virtually nothing about whether it is "rigid enough", whatever that means.
When it comes to just the quantitative stuff, it is about as efficient as the ESL63.
There is nothing for me to link or show. It's up to Sanders to back that claim that "Magnetic" speakers (he probably means moving coil/dynamic) can't match the "performance" of electrostats due to diaphragm mass.
What "performance"?
I remember being intrigued by Sanders' articles in The Audio Amateur when he published the details of how to build his direct-drive amp and electrostatic speaker combo - but he has always seemed to have a rather tenuous grasp of theory. Even then you could tell by the handwaving that he accomplished most of his designing by trial and error.
Somewhere on his site he gives a putative explanation for the higher instantaneous power capabilities of tube amps, and he describes the phenomenon we know as "space charge" around the cathode providing a huge current reservoir to the amplifier. Right then I realized that he's just trying to pump up the audiofools. I wouldn't take any of his claims about loudspeaker performance seriously unless I read them from someone with a little better grounding in physics.
In comparison, the mass of the diaphragm of a candidate tweeter is given as 0.21 grams.
About 86 times as much per area unit, in order to keep the diaphragm rigid. But that says virtually nothing about whether it is "rigid enough", whatever that means.
When it comes to just the quantitative stuff, it is about as efficient as the ESL63.
I love people who cherry pick their facts.
The use of "mass per unit area" is entirely misleading as there is no need for the ring tweeter to have anywhere near the same diaphragm area as a comparable 'stat.
The ring tweeter's small area is a well-known technical advantage.
Comparing the size, complexity and cost of the 'stat with the ring tweeter or typical speakers containing the ring tweeter would be instructive, but also devastating to the 'stat. Are there any 'stats with comparable dynamic range and power bandwidth for less than $10K? A speaker with a ring tweeter and suitable woofer could run less than $200.
Comparing efficiency and ignoring all of the other disadvantages of the 'stat, or if you will the cost of elaborate and expensive efforts to attempt to overcome the inherent disadvantages of stats, is another example of cherry-picking the facts.
So you want to talk about F but refuse to talk about F*s, and then you speak about cherry picking?
So you want to talk about F but refuse to talk about F*s, and then you speak about cherry picking?
Please expand on that. I have no clue about what you are trying to get at.
I realise the thread is a bit old, but just in case someone was interested the http://massless.info site debates the use of the term with respect to plasma speakers, tweeters, ionophones that sort of thing.
the http://massless.info site debates the use of the term with respect to plasma speakers, tweeters, ionophones that sort of thing.
The argument made there is very centered on how bad the inertia of everyday speakers is, as it creates a mass-damper system. For example
All of the above have a common issue - they use a solid mass to drive the air. As a result of this there will always be some distortion present as the solid mass acts as a spring-mass-damper system and never truly follows the original signal. Of course this distortion is of little consequence for a wide variety of uses, however if we are trying to approach perfect sound reproduction then it is an area that needs to be addressed.
What it fails to mention is that as far as I know most plasma drivers have some sort of horn, which encloses an amount of air, and therefore also create inertia, and a mass-spring-damper system. However, while in an electrodynamic transducer you can control the mass, spring and damper, the air in a horn has fixed properties.
Other advantages of having no speaker diaphragm include: the lack of resonant frequencies
But then the horn has resonances, which makes this point moot.
Cool, necrothreading.
Well, if it resurrects again in 2026, hopefully this link will still work https://adireaudio.com/wp-content/uploads/2018/12/Adire-Audio-Woofer-Speed-by-Dan-Wiggins.pdf (https://adireaudio.com/wp-content/uploads/2018/12/Adire-Audio-Woofer-Speed-by-Dan-Wiggins.pdf)
Hello!
I am reopening this topic because, having read it, it does not seem to me that the assertion of dynamic/cone inferiority (to ESLs) due to their being spring-mass-damper systems was addressed.
This page (https://janszenaudio.com/pages/what-makes-electrostatics-better) describes it well:
"Interestingly and importantly, the way an ESL converts an audio signal to sound is the exact inverse of how a recording microphone converts sound into an audio signal. In a microphone, pressure creates voltage, and in an ESL, voltage creates pressure. This contributes to the exceptional accuracy of ESLs. It is not the case for cone speakers, where electrical current supplies non-linear force to a multiple spring-mass-damper system.
...
The situation is very different with cone speakers. One generally hears substantial contrasts in tone when listening to different designs. A large part of the explanation for this is that a heavy, spring-loaded coil pushes and pulls the center of a high mass radiating surface. There are practical limits to how accurately the coil movement can follow an audio signal, and then to how well the cone movement can follow the coil movement.
A cone speaker's resulting lags, displacements and resonances alter the sound in several ways, and add sound that was not part of the original recording. These effects can be minimized, but not eliminated or even made inaudible."
It is a fact that all cone speakers do operate as spring-mass-damper systems, and also a fact that such systems inherently lose information.
Intuitively, this strikes me as a serious weakness; it is intuitively obvious that this characteristic is going to result in substantial loss of information vis a vis an ESL, even with their other advantages (no crossover, uniform drive) being left on the table. We can also suppose that the advantage of the super-light membrane is not really an advantage, for the reasons discussed here - i.e., that heavier mass simply requires greater force and thus there is no inherent advantage. (FWIW, there's still an advantage in practice there.)
Is there anyone present here who would take a stab at "debunking" this supposed critical weakness of cones, and thus advantage of ESLs?
What counts as loss of information? How much of it is too much?
If there is an imperfection that can be near perfectly negated by some equalization, is that also loss of information?
I would trust measurements more than intuitive reasoning. Are there any such measurements to back up this argument?
Is there anyone present here who would take a stab at "debunking" this supposed critical weakness of cones, and thus advantage of ESLs?
Sure. I'll start with a simple omission in one of the statements:
"Interestingly and importantly, the way an ESL converts an audio signal to sound is the exact inverse of how a recording microphone converts sound into an audio signal. In a microphone, pressure creates voltage, and in an ESL, voltage creates pressure. This contributes to the exceptional accuracy of ESLs. It is not the case for cone speakers, where electrical current supplies non-linear force to a multiple spring-mass-damper system.
This ignores that electrostatic speaker membranes are usually rather large. Microphone membranes are often less than a centimeter (roughly half an inch) in diameter, and with larger membranes one often sees a roll-off in the high frequencies. The size of the membrane matters, so one cannot say an ESL is a direct inversion of a microphone.
More in general: Sure, electrodynamic speakers are not perfect. Tuning a mass-spring-damper system is a compromise, that is correct. But in return, electrostatic speakers are large and thus deviate from the ideal point-source. The lack of a crossover actually works against electrostatics in this respect: the higher the frequency, the more the size of the driver matters. So, a two-way speaker system can use a small element to come as close as possible to a point source for those high frequencies that need it the most, while using a larger driver for low frequencies that have long enough wavelengths for the size not to matter too much.
As with most technologies, there is compromise. Just because one technology outperforms the other in one aspect doesn't mean it is better, as it might be flawed in another.
Finally: Almost all nearfield studio monitors, which are built for the sound production with the highest accuracy, are electrodynamic, not electrostatic. That is not a coincidence. You could also say: the recording you're listening too has probably been approved by the musician while listening to an electrodynamic speaker. There is no reason to assume electrostatic loudspeakers will be closer to the artists intent if that artist worked on the recording using electrodynamic speakers.
What counts as loss of information? How much of it is too much?
If there is an imperfection that can be near perfectly negated by some equalization, is that also loss of information?
By definition, lost information cannot recovered.
The amount of loss of information that is tolerable is subjective.
But the topic here is the supposed debunking of objective electrostatic advantages.
We are talking here, mainly, about fine detail - decay trails fading to nothing, atmospheric nuance, etc. - these things live entirely outside the domain of EQ, which exists mainly to correct frequency response anomalies.
I would trust measurements more than intuitive reasoning. Are there any such measurements to back up this argument?
Yes. They are very easy to find - almost any ESL has far lower distortion in the midrange than almost any dynamic cone speaker, typically by an order of magnitude or better, even when the dynamic speaker is far more expensive.
In addition to its (virtually literally) perfect time-domain performance, the ancient Quad 63 & derivatives are some of the very lowest-distortion transducers made, at any price.
https://www.stereophile.com/content/quad-reference-esl-2805-loudspeaker-measurements
Measurements show that the best (under ~$50K) dynamic speakers such as Wilson Sasha and YG are around .5% distortion (500-1000 Hz) while Quads are .05-.1% - around 10% as much distortion.
The best dynamic speaker I could readily find measurements for compared to the Quad:
https://www.stereophile.com/content/wilson-audio-specialties-maxx-loudspeaker-measurements-part-2
The Quad is still 10 dB better than the incredibly expensive Wilson "MAXX" in the critical midband (1 Khz), and much better at lower frequencies. (S-phile summed it well: "The Quad was always a low-distortion speaker—I remember being with Martin Colloms when he measured an ESL-63 and found that it produced 0.1% THD or less over most of the audioband!")
The only dynamic speaker I have found that measures as well as even that old Quad 63 in the midrange was an active (active drive has advantages in driver control, of course) ATC three-way - it got very close. IIRC that speaker was well over $50K.
---
These measurements reinforce what my ears tell me. I have been going back to the Quad 63/988/2805 after trying many other speakers over the last decade or so because none of them give that sense of extreme purity, especially in the midrange. (I think the lack of any crossover and the resulting almost perfect phase performance are important ingredients as well.)
More in general: Sure, electrodynamic speakers are not perfect. Tuning a mass-spring-damper system is a compromise, that is correct. But in return, electrostatic speakers are large and thus deviate from the ideal point-source. The lack of a crossover actually works against electrostatics in this respect: the higher the frequency, the more the size of the driver matters. So, a two-way speaker system can use a small element to come as close as possible to a point source for those high frequencies that need it the most, while using a larger driver for low frequencies that have long enough wavelengths for the size not to matter too much.
As with most technologies, there is compromise. Just because one technology outperforms the other in one aspect doesn't mean it is better, as it might be flawed in another.
You are certainly correct that there is no perfect speaker.
(It is interesting that one can literally spend $100K on a speaker system and immediately notice many faults. Audio shows are good for this type of education.)
You are also correct that the dispersion characteristics of large panels are a negative. The Quad 63 & friends get around this with that ingenious delay system. Trade-offs there too? Certainly.
In any case, because I came across this thread and found it interesting, I wanted to see what people here had to say about the mass-spring-damper issue. The stated purpose of the thread was to debunk the notion that electrostatics are generally superior to dynamic speakers, yet what is assuredly the biggest weakness of the latter was not discussed.
Finally: Almost all nearfield studio monitors, which are built for the sound production with the highest accuracy, are electrodynamic, not electrostatic. That is not a coincidence. You could also say: the recording you're listening too has probably been approved by the musician while listening to an electrodynamic speaker. There is no reason to assume electrostatic loudspeakers will be closer to the artists intent if that artist worked on the recording using electrodynamic speakers.
One need only take a look at any recording studio to realize that using ESLs as monitors is completely impractical for physical reasons. :)
The active ATCs I mentioned previously are extremely popular in studios, and, as noted, they are literally the only dynamic speaker I have ever been able to find measurements for that can compete with even the 50+ year-old Quad 63 design in midrange distortion.
They - the ATC three-way actives in particular - are also extremely expensive.
And, to my ear, still inferior to even the original Quad 63. I've owned several pairs of ATCs, though not the active 3-way models, but I have listened to those extensively at shows. I'd also climbed the Wilson ladder up to Sasha, but they, for their charms, can't touch the old (properly rebuilt) ESLs for purity, transparency, and naturalness - IMHO. (And going higher up that Wilson ladder is big $$$ indeed.)
FWIW, if anyone is interested, this is what I have hit upon, what has proven unbeatable for me after 15 years of being deep in the audiophile hobby, owning over 50 pairs of speakers, from single-driver backhorns, to conventional multiway dynamics, to cones (including with field-coil motors) on open baffles, to - probably the most esoteric - front horns with field-coil compression drivers (Cogent):
Quads, positioned semi-nearfield (8') in a large room, actively crossed at 24 dB/octave to dual 15" sealed subs.
Again, for sure there is no such thing as any home audio/speaker system that is without flaws, but I cannot find anything better, at least for under six figures. And this can be put together for far less.
But, again, IMHO and YMMV. It's all quite subjective.