ESL as a compression driver?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
In my various searchings and musings I may I had a brainstorm; I really like horns for their low distortion and high efficiency. Also I find that using a horn as an EQ device allows a boxless sound without electronic equalization. Couple with that the predictable and useful dispersion characteristics of some horn designs, I think that that is winning technology for hi fi. The problem I see may lay in the inability for most transducers, whether they be cone, dome, or compression, to produce a true planar wavefront. Most transducers save ESLs.
I would like to design a small panel, maybe 5" X 5" in an assembly such that it is purpose built to load some sort of front horn. The benefit I believe will be to apply the benefits of horn loading to what may be a textbook transducer. The design goals will be a limited bandwidth ESL capable of producing a planar wavefront with as high an spl efficiency as possible.
Because there will be no "compression" there will not be any efficiency increase from this effect. This will be an open backed driver with some absorptive material to attenuate the backwave. The ESL should have as little spacing as possible to provide system response to 300hz. To do this diaphragm tention should be high but not above 150hz, or an octave below tuning. The high voltage bias should also be as high as possible. Physical limitations of this will be the close spacing of the diaphragm and stator and the diaphragm tention. I am thinking a spacing of 1.5mm and a bias of 2kv. because of the modest excursion i think lots of diaphragm support strips or dots may be used to increase the bias further.
With the production of a planar wavefront I am thinking of loading this with an oblate spheroid waveguide whose design is to transition a planar wavefront to a spherical one.
Is this a worthwhile endeavor? I have built ESLs before, so that presents no challenge. Please, tell me your thoughts!
 
Hi,

the idea and sensitive use of a Horn is to make a small membrane virtually bigger. The better coupling to the air (acoustic impedance matching) is responsible for this. What is often forgot is, that the transformation is not lossless. A horn is neither theoretically nor practically an ideal transformer. Upsizing the membrane on the other Hand is ´lossless´. And since with ESLs it is always better and very easy to have a larger membrane.
So the horn-like contour of this baffle shouldn´t give a better efficiency than an equally big sized membrane, but rather the opposite.
You can easily shape the distribution character of the ESL by curving or electrical segmentation, so the waveguiding function of the horny baffle :clown: is not used, or shows any advantage against such a panel.
Acoustically its always better to have as much membrane as possible and as less housing, baffling, chambering as possible to avoid reflections at the Horn-/Baffle-surface.
Optically I find a transparent ESL design -even if it is large- much more pleasing than a same sized solid Baffle.

The forces acting on the membrane/air in an ESL are very small. Adding considerable air mass with the Horn will lower the bandwidth. But it is one of the major superiorities of an ESL to have wide bandwith. You´d have to add another driver and crossover which makes the things more complicated, expensive and which looses on basically every acoustical aspect without gaining on one.

With ESLs it is as such:
Think big, build Big, smile BIG -- nothing else

Just my thoughts. Maybe I´m totally wrong ;)

jauu
Calvin :D
 
IMO you might as well use around 7kv bias and 2mm spacing. This will enable the esl's drive to lower frequency's if desired while not loosing any efficiency. It should be easy to tension a small panel tight. Make the mylar as tight as possible. Bowing should not be an issue.

Don't section any more than 100 times the d/s spacing. Anything lower than a 2:1 LxW ratio and you may loose bass response.

A flat esl may beam the sound right out of the horn, i have never made a horn esl.
 
Hi,

if You want an ESL with acceptable or even good efficiency don´t don´t DON´t ever think of increasing the d/s!! 2mm is already too much for freqs above 300Hz. You will loose nearly on every point!
Dynamics, sound, livelyness, efficiency are getting poorer with increasing d/s (hence increasing voltages).
Always think of using as less d/s as possible! Nothing else! When You need more dynamics, increase the area and only the area!
If Your membrane touches the stator, increase mechanical tension by tightening the membrane or reducing the distance between the spacers. Only this way You´ll get an panel with horn-like dynamical sound.

Maybe everyone interested in building a good panel should learn this ESL-Mantra first. ;)
- build big
- use small distances
- use lowest voltages

jauu
Calvin
 
Hi,

Merhaut describes basically just what a Horn does. Nothing really special regarding how long Horns are already in use.

What he doesn´t do -imo his greatest failure- is to check wether a electrostatic device is capable to drive this load (the Horn) properly in first case! :bawling:
I´m quite convinced that this is not the case! :whazzat:
The forces acting upon the membrane are very small, so the bandwidth will be very small too (in contrary to Merhaut´s claims).
The membrane will undergo significant stress because of high pressures. I doubt, that there will be (ideally) translatory movement of the membrane. Instead there will be local bending and flexing, because of the low number of transmission channels between membrane and horn throat and the softness of the membrane.
Just waste a thought of how long the typical membrane coating might last under these conditions! :rolleyes:
The claimed raise in efficiency only applies to a small membrane coupled to the Horn. A membrane the size of the Horns mouth area will have equal efficiency (it should be even higher, because the horn transformtion is not without losses).

Funny enough that all claims of Merhaut can be reached far easier by upsizing the ESL-device to the Horn´s mouth area.


jauu
Calvin:D
 
Calvin, for sure you are not right at this moment. Josef Merhaut did that and he was a very clever guy, a university profesor of electroacoustics and so on.I was talking to guys in my country who listened to his system years ago, a they say that they were totally amazed. Using a horn in between an small electrostatic panel is a good idea, but you must be sure that you are doing it in the right way ;)

I read an article in one of our magazines from 1972, and there were plans and measured data for his tweeter system, that he designed, and I´m sure, it was a prodigy, because this type of construction solves REALLY a lot of problems. Only problem that it has, as people say, was very little maximum loudness.

They sold it to KEF the same year i think
 
I will look into that patent.

I can see the SPL response changing in Three ways; First of all, the ESL will be placed in a much larger baffle than normal, which should augment the low frequency response due to dipole cancellation. second of all, if the OS waveguide functions as I want it to it should spread the high frequencies along with the low frequencies providing constant coverage. This could cause a loss of high frequencies due to the lack of beaming common with ESLs. Third, the fact that there is a large curved baffle in front of a radiator will certainly cause internal reflections (HOM- Earl Geddes) or high frequency roll off.

I plan for the next step of my system to be a limited bandwidth transducer to go in between a midbass horn (70-300hz) and a ribbon tweeter (4khz and above); so limited bandwidth is acceptable.

The horn load is my big wonder. I know that there is a limit on the force (BL) which an E stat transducer may provide, based on arcing etc.

I suppose my main query is whether the special loading demands placed on a driver by a horn will allow some beneficial design requirements on a specially made driver such as not requiring a large xmax which will allow a small d/s. Also, I am willing to utilize the super low distortion provided by a properly designed ESl with no problems of beaming and the ability to match the radiation pattern in between the midbass horn and a tweeter.

My backup plan is horn loading an eminence alpha six midrange. this will probably be a much simpler and more rugged solution but in my quest for the best I am trying to flush out all possibilities.
Thanks!
 
Journal of the Audio Engineering Society...

See if you can find the original publication or the one from the JAES Anthology collection. That artical has all the information concerning the construction of this comp unit. I should think that given the author that this design was very carefully excuited and that it does exactly what it says it does. Merhaut has written one of the most complete chapters on ESL design in his text Theory of Electroacoustics published by McGraw-Hill ISBN number 0-07-041478-5 which went into print in 1981. I doubt that you will find anyone with a stronger theoretical/matametical understanding of electroacoustics including ESL's today
Regarding wave guides it was my understanding (and I would be happy to be corrected if in error) that the wave guide has little to no impact on the high frequency distrubution. From what I gather by the time a wave front has left the throat of a wave guide it will have almost no interaction with the guide. This would be valid at frequencies above that of the wavelength of the diameter of the exit opening of the driver. In other words at frequencies above this the horn or wave guide has no loading affect on the comp driver. Perhaps Tom Danley or Earl Geds could comment on this?
 
thats partly true, I have similar conclusions in my work on tractrix fullrange waveguide which was my bachelor degree work, but IMHO it is not completely true for "Merhautofon" as we name it in our country. Because the diaphragm is loaded in a very different manner, and IMHO the directional pattern depends almost completely on the wave guide, and its shapes and dimensions.
 
Can you expand on this?

Are you referring spicifically to Josef's design or are you including dome comp drivers? Perhaps you could start another thread and discuss the findings of your work on fullrange wave guides, that would be very interesting to hear.
I have wondered if ESL drivers could be used in a "Unity" or "Synergy" style (flat sided conical horn) multi driver horn as designed by Tom Danley. The fact that the driver would be very thin and allow the diaphragm to be mounted very close to flush to the inside wall of the horn would be a big bonus. Levels might not be state of the art for pro application but I should think that they would be more than high enough to satisfy home and HT use. Any comments on this idea?
 
Coffee can you expand on the loading of Josef's ESL comp driver?

From what I can recall of the design (I have a hard copy somewhere) Josef's design is a number or vertical but narrow ESL sections which are formed into a concave arc (one diaphragm), and each section has a rectangular slot which couples the diaphragm to the throat. Each of thse slots is the same length so the combined wave fronts all arrive at the throat at the same time. If I recall it was three or four inches high with about eight sections (plus or minus). Seems to me that this is a pretty standard approach for a comp driver (though this phase plug is easier to exicute with planar elliment than on a dome). So what is it about Josef's design that makes the disperssion of the driver so much dependant upon the horn? Has been some time since I looked at the paper so I may be forgetting something I think that the horn was exponential with a large mouth and fairly long. I do recall that the horn and driver combined did have a very wide bandwith going as low as about 1Khz (or lower) before it rolled off quickly. I remember thinking that it could mate up to a TLine and make a cool two way design. Would be interested in hearing your thoughts on the design.
 
I´m only refering to that fact, that if you are thinking coupling an electrostatic panel or a fullrange loudspeaker directly to the waveguide, both of these drivers will have very high directivity at high frequencies. But in my case ( i have examined a tractrix waveguide of mouth diameter of 50cm = 20 inches and fostex FE 206E) the directivity pattern was affected by the waveguide even at 18 Khz. Although the horn is very short (21cm) and open it has an influence.

In Merhaut´s design, the nine parts of diaphragm are coupled with the waveguide through a system of nine slots, that are 15* smaller in width then the actual width of one part of the diaphragm, so the acoustical transformation of impedances is 225*, so the diaphragm is loaded not only by the real part of exponential waveguide throat, but this impedance is multipled by a factor of 225, so the diaphragm is very heavily loaded. But IMHO that is not the fact that is talking about directivity. All of this little slots ends on the part of circle, so the waves are at this point summing together and forms a cylinder like wave. So i think only two things that affects the directivity patttern are the throat area (and its shape) and the waveguide. Doesnt matter that the panel would have high directivity alone, the wave is completely retransformed in this system IMHO.

on the picture are directivity patterns of my waveguide at 18kHz
 

Attachments

  • resize of smerovosti_18k.jpg
    resize of smerovosti_18k.jpg
    67.8 KB · Views: 443
Hi,

my comments only regarded the patent. And in this paper basically nothing useful is said about the ESL-device. Maybe the AES-print could solve some faqs?

Still though, the only reason to use a horn is to better couple a small membrane to the air. But if You don´t need to have a small membrane it´s of no positive use at all (this holds true for an ESL, with dynamic drivers it´s a completely different matter!)
It is important to understand that a larger membrane size is always beneficial with an ESL and should be one of the prime design goals!
Its imo equally important to realize, that -in opposite to dynamic drivers- the enlarging of the membrane doesn´t come with negative effects like reduced bandwidth, etc. Distribution character can still easily be shaped by electrical segmentation or curving. So having the choice between a larger ESL-Panel in first place or a smaller and inferior panel, beefed up by a lossy, distorting, reflecting mechanical device (the horn), I know what I prefer. Especially when these inferior results take up more space and need much more effort in building and cost.

I repeat: "Funny enough that all claims of Merhaut can be reached far easier by upsizing the ESL-device to the Horn´s mouth area."

jauu
Calvin
 
I think that you didn´t understand merhauts reasons for designing this blessed thing at all IMHO. As he wrote, he was not trying to build a small membrane panel that is capable to reproduce a large range of frequencies. That virtually impossible and because of well known reasons. His main goals was to produce a system that was capable of astonishing impulse response without resonances of frame, grill etc. Because of that he choose to do only tweeter system ( because of its importance for impulse behaviour of whole system) with relatively small membrane (easier to manufacture a very sturdy frame, and that area was OK for his range of operation with this kind of acoustical transformation) and to couple this system with acoustical resistance to load the diaphragm extremely heavily to minimize all kinds of parasitic resonances, to make the diaphragm move even more like a plane piston. As I remember, front part of grill was made from massive aluminium blocks that makes together the system of slots for acoustical transformation, and the rear were about 2,5mm thick steel perforated desks, supported every 15mm !!!!!!!!, it was altogether screwed. ;)

I think that this type of symbiosis of sturdy frame and multiplying acoustical loading by a factor of 225 is not common in panel devices.

Guys that I talked to, said that this system have amazing aperiodical shaped 20us square response, that is amazing and IMHO this is made by the heavy acoustical transformation.
 
Hi,

actually I can´t detect anything ´blessed´ about a device that basically puts away with everything whats good in an ESL!
To load the Horn properly You need a hell lot of force per unit area.
Dynamical tweeters with a strong motor develop by a factor of ~10^4 higher forces than an ESL. The ESL counters this with a very lightweight membrane (factor ~10^3) and larger membrane area. Alone from this a ESL device will be hardly capable of driving a Horn, let alone a compression driver with such a high compression rate!

The quality of the impulse response (step function) can be easily measured and I wonder what anyone expects to do better than an good ESL already can?? Depending on the dimensions of the panel You can make it work onto an real acoustic impedance -which is the ideal state of affairs!! You can´t be better than that, no matter which effort or which additional device You use!
The only reason for resonances are the mechanical tension of the membrane and the step in mechanical impedance at the outer rim of the membrane. You can tackle that problem easyly by introducing some damping (best : distributed acoss the membrane area).
Likewise its quite easy to stiffen the stator. A rattling flimsy resonating frame is definitively not part of a good ESL design.


Putting very high pressure on the membrane will produce distortion through heavy flexing of the membrane since the thin film has no intrinsic stability like e.g. a metal dome of a dynamic driver! It won´t work pistonical at all, thereby introducing distortion! Btw: the curved driver (at least as shown in the drawings) will under no circumstances work like a piston!
Anyway it´s already not easy to measure corrcet distortion values for a good ESL (more than often You rather measure the limits of the microphone!)

2.5mm thick perforated metal at the backside of the device, reduces the usable bandwidth considerably through the tunnel-effect. The oprenings don´t work any longer as openings but as small resonators.

jauu
Calvin
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.