hornloading / waveguiding electrostatics

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
apologies in advance for what may be a poorly formulated or stupid question, but: having begun to read some of the threads related to horns and waveguides, is there a reason why this could not work with a relatively long narrow electrostatic?

the reason i ask if the following: i have dci full range electrostatics. a while back i did a diy project related to what might be termed diffraction. i fixed some 180° pvc guttering pipe to the front and back of the speaker with masking tape (no prizes for aesthetics) in order to have a larger radius turn leaving the diaphragm. pretty much to my surprise there was a noticeable improvement in clarity and intelligibility (on speakers that have a good reputation in that area).

this made me think: i wonder how much bigger the radius could be made?
it occured to me that the larger the radius became, the more it would begin to look like the profile of a horn or a waveguide.

it also occured to me that it wouldn't be too difficult to make something more aesthetically pleasing if one could cnc wood. in the case of the audiostatic one could even make a replacement round upright structure. its true that it would be very heavy. but with an appropriate foot could actually lead to a more stable speaker (another area where i did some experiments a while back and where the speaker benefitted from greater stability).

any thoughts?

a while back i came a cross a commercial product called 'dci-wings', but i understood their purpose was to change the frequency at while the speaker begins to cancel due to 'dipolarity'. also, the 'wings' fit on the back of the speaker only as far as i know.
Thanks very much Charlie!!
The amazing coincidence of the link you sent is that these is the same brand that I have (not sure about the model).

Continuing to think aloud:
(a) presumably having an equidistant 'horn'/'waveguide' edge from the diaphragm leads to an emphasis in the cancellation effect due to 'dipolarity'?
(b) presumably if one were to have an edge that is shorter towards the bottom of the diaphragm, and further out towards the top of the diaphragm one would spread (and perhaps) lessen the emphasis.
(c) presumably, the effect of (b) would be to affect the way the speaker emits different frequencies in the vertical plane. (? a kind of vertical steering?)

I also found one comment very interesting in one the posts of lecleach. I'm not sure that i understood it, but the basic idea is that horn 'expansions' should be pursued beyond 180°. i have the impression he also said that the ideal is 360°: kind of intriguing! i'm not a mathematician but intuitively i understand that to be a kind of doughnut or torus.

i must draw some of these things, as it makes discussion much easier.

I had a look at your own project: very impressive!

what is a horn all about? Acoustic impedance matching of the usually too small diaphragm to the surrounding air. Best state of affairs is reached when the loading is ohmic. This can easily be reached with strip shape Panels of sufficient size. No Horn loading will improve things. Only a small diaphragm would profit from a horn, but at the cost of increasing building size. Increasing membrane area instead is easy and comes without the disadvantages of a Horn.

Thanks Calvin. The audiostatic is relatively narrow. I guess about 12-13 cm, off the top of my head. I have seen horn loading of fostex drivers (eg bart doppenberg) and larger bass drivers for that matter (but the bass drivers evidently don't go up so high). Could you quickly explain the idea of 'ohmic' loading, or point in the direction of an answer?

I quickly lashed up an improvised set up yesterday. The material is a laminate of foam in between two layers of card (used by model builders). The 'curve' project about 60cm from the face of the speakers, so I guesss this should beef up frequences up to 120. Interesting, apart from the same volume setting sounding louder, there was a feeling of less 'shout' listening to a big piece. I was surprised by how much richer the sound appeared to be: much more cello. That surprised me. Perhaps the cello plays higher than I thought it did, or I was hearing more overtones or something. There was something i couldn't figure out about the sound: perhaps its just a question of adjusting the ears.

Yep, Charlie, the aestethics leave something to be desired.....
but i have the gut feeling that one could
I suppose one of the questions is how big the shape should be. The 60cm/2ft that I quickly set up yesterday: it doesn't sound like much, but it did suddenly look very big........

mechanical circuits behave analogous to electrical circuits and vice versa. So You can describe mechanical components by their electrical pendants. An famous example for this are the Thiele-Small parameters.
Now a moving membrane represents a mechanical signal source with a certain impedance. As such it consist of a real (ohmic) part and a complex (inductive or capacitive) part. The impedance depends on the physical dimensions and frequency. As long as the wavelength of the signal is large compared to the membrane´s dimensions the complex part dominates. With rising frequnecy the wavelengths shorten, the frequnecy rises and eventually the membranes dimension become larger than the wavelength. In the same the impedance becomes more and more ohmic in character.
Air it self has an ohmic load-character. Energy is best transferrred from the source to the air-load, when its impedance is in the ohmic range and of similar value. So if we need to reproduce low frequencies with small membranes the impedance mismatch between source and load is large.
What a horn does is to transform a small membrane area to a larger virtual membrane area and as such it decreases the impedance mismatch.
It is possible to design a strip-like ESL (in literature You´ll find a 8:1 H:W relation) that shows an ohmic impedance character over its complete working range.
Of course do low frequencies still require large dimensions, but the clue is, that You can´t improve matters when the source shows ohmic character already. Then the membrane couples to the air in the best and most efficient manner. There´s no such as 110% here. The only use of a horn would be to lower the lower bandwidth limit due to increased ´virtual membrane size´. But the same is better done with larger membrane size in first place.

Thank you very much Calvin.

I misrepresented the Audiostatic width.
I think its more like 15cm.
I will measure the length: it would not at all surprise me if it turned out to be the 8x ratio that you mention.

Independently of the physics you've very helpfully described, I wonder then what I'm subjectively hearing then. I did a second listen yesterday with 'modified' curves (more bent, approaching 90°) and the sound was again more dynamic, more punchy, with more bass.

I'll sketch some of the shape ideas which I referred to over the next couple of days.
In fact the precise dimensions are 143 cm high (4.69ft), 17.5 cm (0.57ft) wide. That is pretty much exactly the proportion that Calvin predicted 8:1.

In the case of these speakers there is an additional factor to take into account. I remember seeing a post on Audio Asylum a long long time ago which explained that only the central part of the diaphragm actually radiates high frequencies. I guess that also affects the relationship relationship of the diaphragm to the 'waveguide'/'horn'.
Finally, a simple hand diagramme for all.

1 was the initial experiment based just on testing if it was possible to hear some kind of diffraction effect

2 was the idea of the replacement upright structure (probably totally unwieldy in wood, but possible to make out of some kind of moldable material)

3 was the idea of the larger radius

4 was the idea of the more than 130° shape

5 was the initial curve i tried

6 was the second more acute shape

7a/b is just the diagramme of how i understand the speaker emits
based on some old audio asylum posts

yep...i know i don't have a future career as an artist....... :-(


  • _1124180026_001.pdf
    24.4 KB · Views: 99
Independently of the physics you've very helpfully described, I wonder then what I'm subjectively hearing then. I did a second listen yesterday with 'modified' curves (more bent, approaching 90°) and the sound was again more dynamic, more punchy, with more bass.

I'll sketch some of the shape ideas which I referred to over the next couple of days.

Your description of the "curves" you added providing more dynamic and punchy bass makes me think that the main benefit you are hearing is from increasing the acoustic path length from front to back and delaying the onset of dipole cancellation to a lower frequency. This is exactly the purpose of the "wings" that Audiostatic sold.

If the DCI frequency response is at all like the ES-100 there is a dip in the bass response below 300 Hz that extends down to the diaphragm resonance. The dip is from dipole front-to-back cancellation for frequencies below where the wavelength is longer than the width of the panel + baffle. Audiostatic did not provide any equalization for this in the ES-100.
Thanks Belserst.
Yep, the wings were supposed to change the 'dipolarity'. The Geschka web site has pictures of them, and posts by a guy from musicology in the uk can be found. One interesting thing about the wing is that it is not a straight line but a curve. From the pics the greatest (path) length added by the wing is about 20cm and that bulge is in the middle of the diaphragm. At the top and bottom of the wing they look as if they add between 5 and 10 cm. Also, the wings were placed on the back of the loudspeakers. In my case the 'shapes' probably add about 60cm and are on the front of the speakers.

I realise too, that I made a typo in a previous post:

"4 was the idea of the more than 130° shape"
should read: "....more than 180° shape"
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.