Electrostatic AMT?

[DavidJanszen]

Originally Posted by Keith Taylor

David, I pressume your aversion to high aspect ratios is related to the potential for generating wideband noise due to turbulent air flow?

Not at all. Having facing areas acting against one another is going to stifle nearly all the potential motion. No motion means no turbulence. Then, the initially paltry acoustical wave will have to propagate at a right angle to the membrane motion along the pleat to the outer air mass.

To demonstrate the first effect, if you have an ordinary ESL panel around, try moving it toward a fixed, parallel surface while it is operating. Then, have a look at the formulas for sound traversing holes or slots.

Originally Posted by Keith Taylor

Building the thing would not be easy.

True, but it will be fun.

Originally Posted by Keith Taylor

Your thoughts welcome.

I'm sorry, but I can't comment further, except to say:

You might want to place the two rows of nails on separate strips, like combs with metal tines.
Use well insulating plastic, not wood.
Leave the baffle for later.

Wood will conduct away the bias voltage.

Keeping the rows of nails separate will allow you to adjust the pleat depth, which will adapt to different patterns and allow you to compensate to some extent for tolerances in your patterns.

If you face one comb up and one down, separated laterally by some distance, you can hang the film between the two rows, then cross their positions laterally to create the pleats easily.

[mige0]

Originally Posted by Calvin

Hi,


No, without polarization You generate null output!
Calvin

Not exactly as you point out further on >>>

Originally Posted by Calvin

Hi,
.....If You look at just a pair of pleats the same situation occurs as with a fixed stator asymmetrical ESL. Adjacent pleats (i.e one membrane-one stator) will be driven with signals of opposite polarity. Hence there will always be an attracting force, regardless of the polarity of each pleat. The plot of force over time would look like a fullwaver rectified AC-signal and not like a sine. This would give some output, but not what we wished.

jauu
Calvin

Yes - that's been the scenario why I said it will produce a looot of distortion.- and if I remember correctly there were some vintage ESL tweeters made that way.

Originally Posted by DavidJanszen

What Calvin said. Have a closer look at Keith's diagram and really try to get it. I was hoping my description would help you understand Keith's eureka moment -- he really had one.

I was saying that there will *not* be a lot of distortion. Keith's concept is push-pull..

Sorry, my fault !

Originally Posted by DavidJanszen

In a push-pull, constant Q ESL with appropriate membrane mechanical characteristics, force vs. displacement is nearly constant over a considerable portion of the gap width. Such constant force is unique among audio transducers. There are also practically none of the confounding dynamic effects found in electrodynamic transducers.

I partly don't agree on a more subtle level, but I agree on the more practical level under discussion here.

Originally Posted by DavidJanszen

. If you are unfamiliar with the means by which push pull operation in an ESL cancels first order force nonlinearity, please read one of the relevant texts. It is also important to appreciate the means by which constant Q (a.k.a. constant charge) operation also serves to flatten the excitation vs. displacement characteristic.
.

Well – I clearly have to admit that I was wrong that polarisation might make no sense here – shame on me!
On the other hand, I basically doubt that the "eureka" version can be seen that simple.

Forces are in labile balance IMO – meaning that adjacent pleats tend to collapse

Lets make it simple and think of three halved pleats – exactly as a conventional push-pull ESL is made.
Now make the outer parts – the stator in conventional ESL – movable.
What you think would happen?

The outer half pleats will tend to collapse towards each other every time you apply audio voltage IMO

Now lets stack with another set of two halved pleats to make for the repetitive AMT pattern.
What you think would happen?

The middle "stator" has the choice which side it wants to collapse to –
There will be some cancellation of forces – sure – but any slight misalignment will push things out of balance.

Though - possibly its not a severe issue in the end...

Originally Posted by DavidJanszen

Impressive.

Well I was flattened too when I first saw how less it takes – considering the great sonic results provided.


Michael

[mige0]

Originally Posted by Keith Taylor

David, I pressume your aversion to high aspect ratios is related to the potential for generating wideband noise due to turbulent air flow?.

Keith

There's deep suspiciousness I see in many comments about this when it comes to AMT.

There are two ways to counter that –
1.) listen to one – I guarantee – no more questioning.
2.) in case you haven't any chance to get familiar by listening to such device – read distortion plots – modern AMT's are outperforming most other speakers in this regard – but on the other hand, some say THD does not matter.


Bottom line – don't care about aerodynamics.

What can become a possible issue IMO is when depth of AMT pleats get into wavelength territory. There is a reason that diaphragms of AMT's aren't especially deep – depending on the upper frequency you want to reproduce of course.

Otherwise - at some point you would have to operate the pleats in a WALSH manner in order not to get cancellation of sound along its way along the pleats – possibly doable – possibly another *heureka moment* - possibly a reason to file another patent.

Originally Posted by DavidJanszen

Pleating the membrane of an ESL will multiply the load by increasing the amount of air load per unit area of membrane, to put it mildly. The motion of one pleat, in fact, will be in direct opposition to that of its immediate neighbor, and will tend to cancel all motion rather than pump air in and out of the pleats. Also, if the pleats are made deep enough to generate sound at reasonably low frequencies, then this will create the equivalent of a well through which the sound that remains after cancellation must travel, which will create acoustic filtering effects.
.

I'd strongly question the explanations given here except for the last part (see my thoughts above) – with the exception that the capability to reproduce low frequencies is *not* related to the depth of the pleats – at least not directly.


Michael

[DavidJanszen]

Originally Posted by mige0

Not exactly as you point out further on >>>

Please re-read and try harder to follow the discussion. Your lack of attention is getting frustrating.

Originally Posted by mige0

Forces are in labile balance IMO – meaning that adjacent pleats tend to collapse

All ESL's experience a small positional offset, not much, but some, due to the inherent instability of the situation. When the bias is set too high, there's a sudden transition from a slight offset to complete collapse.

Keith's scheme is not unlike any other, except that both the bias and driven members are flexible, so there is a lower limit on bias for a given gap, membrane tension, and membrane tensile modulus.

Originally Posted by mige0

Well I was flattened too when I first saw how less it takes – considering the great sonic results provided.

I'm afraid I was impressed not by what you think I was.

[DavidJanszen]

Originally Posted by mige0

There's deep suspiciousness I see in many comments about this when it comes to AMT.

We are discussing Keith's invention. Please start your own thread if you are interested in something else.

[Few]

Keith,
When I tried to slap together an electrostatic AMT several years ago I used a somewhat simpler construction. Since I didn't take the thing to fruition I can't offer lots of practical insights, but maybe something of use will arise.

I used rigid, nonperforated stators made of aluminum insulated by plastic shelf liner (adhesive backed plastic on a roll). Along one vertical edge of each piece of aluminum I wrapped some 1/2" wide 3M double-stick foam tape (1/16" thick). The tape ended up adopting a U-shape when I wrapped it around the edge of each stator. My intent was to wrap the coated diaphragm in a serpentine pattern around the fixed stators, and use the foam tape to maintain the diaphragm/stator gap. The foam spacer would be on the front of half of the stators and the back of the other half, in an alternating pattern. The high voltage music signal was to be applied to the stators, positive polarity stators alternating with negative polarity stators. A fixed bias voltage would then be applied to the diaphragm.

I never came up with a good way to achieve nice even tension on the diaphragm in this assembly, but also got distracted by life's demands before I had a chance to play with it very much.

The approach I described isn't the same as your no-rigid-parts design, but I thought I'd chime in just in case it triggers something useful. Perhaps my description will just be confusing without a sketch---if so let me know. In the meantime, thanks for sharing your ideas.

Few

[mige0]

Originally Posted by DavidJanszen

Please re-read and try harder to follow the discussion. Your lack of attention is getting frustrating.

Originally Posted by DavidJanszen

I'm afraid I was impressed not by what you think I was.

Originally Posted by DavidJanszen

We are discussing Keith's invention. Please start your own thread if you are interested in something else.

Ahh - mmmmhh ?
Any particular passage or tone in my postings to trigger your cynical patronising reflex?
You 'PM is welcome any time not to load this thread with personal stuff...

Michael

[Keith Taylor]

Hi all, and FEW, in particular, I am a bit reluctant to post these links to the European patents as they are NOT what we are currently disscussing.
esp@cenet — Bibliographic data
esp@cenet — Bibliographic data
Not certain why the URL's, which I entered by hand, have morphed into something different, but it seems to get you there. Drawings are on the "mosaics" tab.

It is well known that "too hard basket" problems, such as connecting to the pleats become clearer when one is forced to write about them or draw them. Things get clarified in the attached file, where making connections via the posts seems the way to go, but maybe not for a rough version.

Keith

[DavidJanszen]

Originally Posted by Keith Taylor

Hi all, and FEW, in particular, I am a bit reluctant to post these links to the European patents as they are NOT what we are currently disscussing.
esp@cenet — Bibliographic data
esp@cenet — Bibliographic data

As a first pass, I think that if you were to prototype something like a section of the second one, but using membranes in place of the stators, you would have a practical test bed for your idea. In other words, instead of going straight to trying to create a serpentine arrangement with a patterned membrane along with all the challenges this entails, you could stretch a series of independent membranes in a row/stack.

Each membrane would be thus electrically isolated and easily energized independently. You would create pseudo-pleats by taping off the edges of the gaps on alternating sides or just introducing a bit of filler between the membrane supports.

This apparatus would predict whether sound can emanate from your invention, and if so, help you determine how its spectrum and amplitude are affected by changes to gap width and pleat depth. If the results are promising, it could even serve to help you adjust bias voltage and membrane tension, mass and modulus. Don't worry about baffling it; at the high frequencies at which I think this has a chance of operating, there will be no significant front-to-rear cancellation.

[mige0]

Hi Keith

David's posting made me think about I possibly have hurt your feelings by not seeing the beauty of your attempt at a first glance.

If so I'd really want to apologise for my – what I thought –harmless joking around.

That hopefully out of the way I'd like to express that I find your idea the better the longer I think about it - though - naturally – have some reservations (due to reasons outlined) – but I certainly want to encourage you to give it a go!


######

One of the bigger problems doing a prototype will be to keep good control over foil tension (you know the collapsing issue I raised).

Thinking about tensioning the foil I basically came to the same conclusions as David first – but would not recommend his taping variant.

If you make sliced curtains like suggested by David and apply tension only in the vertical direction you probably run into problems with taping off the gaps *if* it was meant to be in meander form as in usual AMT's

This kind of "sealing" most probably is not good enough for one (AMT's do have a pretty stiff foil in contrary to the extremely thin and weak Mylar you will have to use !) and secondly the tape would possibly wrap uncontrollable when adjacent pleats try to close - generating a lot of unwanted noise.

You possibly could instead fold your pleats in an sharp (saw tooth) zig zag form.
That the pleats are no longer in parallel should do no harm as long as the even and odd pleats are roughly parallel to each other.
Regarding force linearity it does not matter *where* in the electric field the constant Q actually is.
This form of folding – I admit – isn't the best with respect to discharge issues.

Going back to your suggestion and apply tension only in the horizontal direction:
if you do the posts with strait nails as you outlined earlier there is a good chance that you get good tension control at the upper and lower area of the foil (where the nails are supported by the baffle they stick in) but you may get weak tension in the middle of the foils where the nails tend to bend.
Pre-bending of the nails would be a solution but you would have to make sure they don't turn around (using a rectangular cross section for example).


Some thoughts about pleats depth:

If you want to set a upper FR limit of – say like a mid rang unit – I guess you should make pleats depth no bigger than around three times the usual AMT depth of roughly 5-7mm 2/10"-3/10" resulting in 15-20mm 1/2"-1" pleat depth.

As for the spacing of the pleats my guess would be that 1mm / 50/100" could make a good start but I think David could possibly make recommendations based on a lot more experience here.


Michael