Sealed box and ESL

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Well, it depends on what you mean by "better". There is no guarantee about it being better, as there are so many variables at play. Certainly dipole cancellation is eliminated from the bass frequencies, and that’s a start. The potential for "better" bass exists, but also the potential to get an under-damped bass bump. Careful design and adjustment of driver and enclosure parameters could make the difference. I believe that more work needs to be done in general with ESLs in enclosures, or better yet, what I've been toiling on, ESLs covering all faces of a closed cylinder.
 
Sure, it acts like a closed box, but instead of side walls which flex and resonate, the whole cylinder expands and contracts ("breathes"). Absorptive damping material must fill the inside, of course. There are many advantages to this kind of approach. Peter Walker of Quad fame built experimental cylindrical ESLs many years before introducing the well-known planar models to the market. Let us know how the two models work out for you.
 
Christmas is comming

Brian: a whole bunch of years ago I built up an ESL which was to be conical. I built the long triangular panels out of perf steel with a nylon powder coat dielectric. Due to other concerns I had to stop panel construction with only half of the necessary panels built. Since I had the frames all ready built I set them up with just half of the panels each. The resulting effect was much like that of the big MBL radial speaker. There was excellent stage width and depth with the interesting effect that as you walked toward the speaker the image stayed the same but moved forward an equal distance to that you had moved closer. The other nice effect was that as you walked back and forth across the back end of the listening room the image stayed solid but your perspective shifted as it would if you were to move across the front of a stage. In other words the players did not move from thier aparent positions but the balance of the presentation shifted as you would expect to have it do in a live situation.
I never got the time or chance to finish the project. I had thought that the conical shape would have less problems with internal back wave cancelation (as compared to a cylinder) as the presure wave front would have the natural tendency to vent out of the open bottom of the cone shaped frame. I did not get the chance to prove or disprove this theory as I only got the half cone versions up and running. Thought that this might be of some interest. Best regards Moray James.
 
I did place a large sewer pvc pipe behind the esl (cut the pipe half). I kept the upper and bottem open. I expected more extended and powerful bass which was not the case. Maybe I got other results with a larger enclosure but I didn't try.
Even the hudge beveridge are compromised in bass extension so I stick to dynamic woofers for bass.
 
Moray,

Interesting project. Do you have any pictures? Relative to your open-end venting comment, I would tend to think of it as depending on wavelength. At low frequencies the open cone end is a large vent, so below some frequency, you have "dipole cancellation" just as if it were a regular planar design. (This is also why MJ Dijkstra didn't achieve deeper bass.) At mid and high frequencies, the wavelengths "see" the other side of the cone as a reflecting boundary, and this distance would vary up and down the cone, distributing the comb filtering effect of the reflections. In my experiments, the configuration is a closed (top and bottom sealed) cylindrical drum with absorptive material inside to swallow the reflections. If the cylinder is tall, from floor to ceiling, you get an ideal line source, with a radiation pattern independent of frequency, and with the potential for pretty deep bass, although the last part is tricky as my efforts have shown me. I will report on this if/when I have something to show and tell.
 
no pics

Sorry Brian: no pictures, don't think digital cameras were around back then. You are right about the cone spreading out the reflections that was part of the intent. The fact that the diaphragm was also shifting from wide at the bottom and narrow at the top was also intended to sperad panel resonances. The panels were about a hand span wide at the base and 1.25 inches at the top and about 40 inches long. Even with only half of the intended panels there was lots of surface area and they were quite dynamic. Stator to diaphragm was just under 1/16 inch. I used an Acoustat interface to drive them.
I must say that I have held doubts all these years about a mono pole design wondering if one could possibly absorb backwave over anywhere near a wide enough bandwidth with uniform absorption. I feel that dipole structures are the way to go. Letting a dynamic driver do what it can do best for the bottom end. I think that if recordings are made with directional microphones then playback ought to be directional as well. No proof about that it's just my impression. Good luck with the experiments. Regards Moray James.
 
Hi,

imo a cylindrical ESL won´t work well. The reason is, that the soundwaves emitted from the inner side of the diaphragm interfere with the ones outside. They are not(!) reflected and kept inside the cylinder, even not at high frequencies. Think of the sealing membrane of Quad. The very thin diaphragm is acoustically nearly transparent! Filling the tube with fiber doesn´t solve the prob, but just reduces it slightly -and only at higher frequencies.
Putting the ESL into a CB can be a solution, but is imo far from optimum. Indeed does it help in reducing phase cancellation, but it should be constructed such way, that the sound from inside is absorbed completely, because every reflected sound will pass through the membrane to the outside -with accompanying interference probs.
So a classical box won´t be best. As Mr. Walker described, a very deep and heavily dampened compartement should be fine. But beeing closed, this box will still suffer from the problem, that almost all panels exhibit a very high Q-resonance with a long decay (group delay). This Q is even raised in a CB. A solution could be to reduce the input signal to the panel with a notch-filter. A measurement that´s used by many designers anyway. Another one is to use an open box design instead of CB. A BR-design for i.e. could be tuned such that the resonance frequency is dampened mechanically. By appearance this design will probabely look more like a tapered TML.
You have still to keep in mind, that with any box the problem of acoustic coupling is enhanced in hybrid-designs.
Too the compartement will always be quite voluminous.
I´d prefer to use a panel with a greater width in a hybrid design, without sealing the panel. Wether driven passively or actively its very easy to equalize the resonance and to correct for the phase cancellation. Depending on the build geometry of the panel the correction can be quite subtle. As I pointed out in an earlier thread ("Please rec. step-up transformers for ESL"), a combination of a dampened second-order-HP, together with a Notch-filter can equalize many panels perfectly.

jauu
Calvin
 
Calvin,
By citing the obvious challenges of backwave absorption, IMO you are too readily dismissing the many problems with dipolar drivers that potentially could be eliminated with a 360 degree cylindrical ESL. We live with these problems and we are too ready to forgive them as something we just have to live with, or go to great lengths to compensate for. Before I go on, please know that my beloved Quad ESL-63s are dipoles, and represent a high-water mark in dipolar design that I respect, but Walker’s efforts to reduce fundamental dipole problems were heroic, and not simply implemented. If the backwave reflection could be stifled, then we are ready to remove a number of dipolar problems. Although I am not prepared to disclose all my work yet, let me just say that there are other absorptive techniques besides fiber fill “stuffing”, including time and frequency selective mechanical dissipative filters that can be used inside the cylinder. Walker started looking at these approaches decades ago.

I was motivated to consider a 360 degree ESL when I was stunned by the sound from the MBL 101 speaker, a 360 design that “shocked and awed” listeners at a Stereophile show with its panoramic imaging and soundstaging.

Here are some dipolar problems that *potentially* could be solved by a tall cylindrical drum with ESLs on all surfaces:

A dipole will exhibit a wavelength/frequency dependent radiation pattern which is a function of panel dimensions. On the “boresight” axis, the highs will rise, and this increase is a complex interaction of the wavelength with both the vertical panel dimension and the horizontal panel dimension. Phase response will not be flat if the frequency response is not flat; so a square wave or impulse transient is smeared in time. Segmentation or electrical filtering becomes necessary, but these measures pose new problems, and rarely are exactly compensatory for the original problem. A cylinder approaches the ideal of a frequency-independent emitter at any listening angle. Phase response and resultant transient would be outstanding. All ESL drivers contribute to a effective pulsating line source.

Dipolar off-axis frequency/phase response will be variable at every angle, in both vertical and horizontal directions, and zero at degrees (a potential advantage). The inevitable room reflections will have a different character than the on-axis response. Not so with a 360 cylinder.

Dipole cancellation kills the bass as a function of wavelength interaction with both panel dimensions. Efforts to boost the bass to compensate for this cancellation can result in large ESL diaphragm excursions that invite overload and arcing. Magnetic driver subwoofers are often employed (as I do with my Quads) but there are new sonic compromises with this approach. A cylinder pressurizes the air in one polarity at every angle. The speaker doesn’t have to work so hard. With either type of speaker, dipole or cylinder, the fundamental panel resonant peak has to be dealt with. Electronic EQ can be used, and if the fundamental equations are understood, can exactly compensate for the bump, at least in theory, the same as can be done with a cone in a box.

The whole dipolar speaker vibrates in opposite reaction to the diaphragm’s forces on the air. This is a big "hidden" problem. Consequently, there is secondary emission from the whole speaker, and frequency dependent Doppler intermodulation, since the diaphragm vibrates within a moving reference system. Mass loading helps, but the dipole is fundamentally imbalanced. All radial force vectors in a cylindrical driver cancel to zero, so there is no net force to push the speaker in any direction in reaction to the diaphragm’s motion. Forces exist perpendicular (radially) to the center line (“squeezing” and “negative squeezing”), but it is possible to brace for these relatively easily compared to the dipole.

I’m not saying it’s easy, but the potential to eliminate the above dipole problems (and these really are BIG limitations) makes the cylinder worth considering, and makes trying to find answers to the backwave absorption a worthy endeavor, IMO.
 
I think that if recordings are made with directional microphones then playback ought to be directional as well. No proof about that it's just my impression.

Moray,

I've heard this belief stated occasionally through the years. IMO, I don't think it holds water. If it were true, you'd need a cardioid speaker for some recordings, and an omni for others. True dipolar (“figure of eight") microphone patterns are less commonly used than cardioids in commercial recordings. Recordings are often made with all kinds of microphones. Multi-tracked studio recordings often combine several microphone polar patterns at the same time. And some the finest classical recordings ever made used three spaced omnidirectional mikes.
 
Dispersion pattern is the function of output vs angle.
If we are listening at a fixed position (which is mostly the case) the dispersion pattern doesn't matter. We can't hear it. We only encounter the secundairy effects of it which is more or less reflection of the walls, a rising freq. response related to beaming effects.
Disadvantages (probably more): A lot of indirect sound (like BOSE), more area which has to be driven by the amplifier.
You have to absorb the sound in the middle of the cylinder because sound of the back of the cylinder will get through the front adding a delayed sound which is absent in the original sound.
 
You have to absorb the sound in the middle of the cylinder because sound of the back of the cylinder will get through the front adding a delayed sound which is absent in the original sound.

Of course! That was the whole point of the discourse above. (Did you read it?)

If we are listening at a fixed position (which is mostly the case) the dispersion pattern doesn't matter. We can't hear it. We only encounter the secundairy effects of it which is more or less reflection of the walls, a rising freq. response related to beaming effects.

Again, you are certainly right, but those secondary effects are not so nice and are worth trying to eliminate. And, at least in my case, I want a broad sweet spot to allow more than one listener at a time. Beaming dipolars won't allow that. Just last night, three of us were listening to old jazz and blues recordings on a pair Quad 63s. We had to keep swapping the sweet seat among us. Besides, I don't like to be locked to any one spot even when listening alone. I have been known to move around playing "air guitar" when no one else is looking.:) It would be nice to have full spectrum response around the room.
 
I like ML speakers very much too. If the Quads didn't exist, I would be looking seriously at current ML models. But the partial cylinder approach that ML (as well as Sound Labs) uses only partially addresses the grievances that I listed above for dipoles. While highs are better dispersed, there is still frequency dependence in the radiation pattern that has had to be compensated for (even for sitting on axis). Dipole cancellation still exists. Reactive forces still exist in the frame and stators. The curvilinear approach, with diaphragm tension much higher in the vertical direction, causes the diaphragm to want to cave inward. ML compensated for this with many support strips and doesn’t allow excursions to become too high (bass is made by cones in most models).
 
Hi,

As you said, marton logan avoids some problems by choosing for the hybrid concept.
The reactive force problem is probably encountered by clamping the durved panel firmly in the frame, mimicking the strength of a full cylindre. Unfortunately most of the logan vibrations are caused by the woofer.
While a cylindre can solve some reactive force problems, vibrations across its surface can still be present. Since bracing of a cylindre is more complicated you will probably have to consider this. A large metal sheet with poor internal damping is a problem wether being a cylindre or not. I think the concept of a cylindre will have its benefits but will be a very difficult to make. I've never seen such a thing. Maybe building some extra rear- or side-firing panels will suit your taste? ( a bit soundlab like)

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