Audiostatic baffle step filter

[bolserst]

Originally Posted by Bazukaz

Perhaps it depends a lot on specific implementation.

Maybe Calvin's point it that compared to large unsegmented panels the dispersion is so much better anyways, that worring about coupling thru the crossbars is not that important. I can't argue with that logic. For a given amount of capacitive coupling thru the crossbars, the impact on off-axis response is seen more for designs with lower segment capacitance and higher value of ladder resistors. In general this means it will be more important for FR designs than for hybrids.

It turns out that each wire can have about 0.6 pF capacitance to one supporting strip(simplification, but perhaps conservative due to assumptions in 3).

Due to fringing between wires and cross bars, you will find capacitance to be 2x to 3x higher than estimated with formula for parallel plate capacitance. Formula for capacitance between wire and plate would be more appropriate. If spacing between the wires is < OD of the wires the capacitance is very near what you would calculate using the segment width in the parallel plate formula. If you have an LCR meter you can easily measure with the ladder resistors removed.

The 2-3 segments are bypassed by far larger capacitance of about 230 pF.

Don't forget that there will also be coupling directly from 1-3!
Trying to model in Spice for more than 3 segments and you wind up with a web of capacitances

It turns out that under these assumptions segmentation resistors are bypassed with significant capacitance which can have even lower reactance than resistance of resistors.

Absolutely. As an example, the attached plot shows expected and measured voltage driving the outter segments in an ES-100 style ESL. Measured capactive coupling was about 200pF! The level at which the attenuation flattens out is dictated by the capacitive voltage divider made from the segment capacitance and the coupling capacitance. This amount of response from the outer panels for f>10Khz negatively impacted the off-axis response. Remember plot shows electrical signal applied(not SPL) to outter segments which are larger in area than the center segment.

Originally Posted by Calvin

Panels are working well even after 20 years. So far the only thing I noticed every now and then is, that the metal bars aquire a bit of charge over time, that faintly zapps you on touching them. But the zap is even less than what happens walking over a carpet.

The charge you mention is the static charge stored in the (wire-pvc-crossbar) capacitor that Bazukas described. The HV supply connections to this capacitor are thru direct connection to wires, and surface leakage path from diaphragm contact to the crossbars.

If you try grounding the crossbars to avoid the charge buildup you will notice two things:
1) Increased capacitive load on the step-up transformer since it now has to drive this capacitance in parallel with the stator-to-stator capacitance.
2) If you are using a QUAD style neon light charge indicator, the flash rate will increase from current thru the grounded leakage path. This is most noticeable when connecting and disconnecting the grounding wire on humid days.

[AAaF]

As you can see in the first post, there is a big resonance at approx 30Hz. Is it any way to modify the system to get rid of this?

I had a hypothesis that this would improve by removal of earlier mentioned series resistor, something that was obviously wrong.

[bolserst]

Originally Posted by AAaF

As you can see in the first post, there is a big resonance at approx 30Hz. Is it any way to modify the system to get rid of this?

The diaphragm resonance is most easily damped by application of acoustic resistance to one or both of the stators in the form of silk screen mesh or thin felt.
This will ruin the see thru optics of the ESL panel(this is very important to some people), but the bass response is much improved.

I posted some measurements of this damping technique here:
Mechanical Sectioning .vs. Silicon dots for resonance control

There had been a thread on another forum detailing the application of silk screen mesh damping to an ES-100. But it looks like the link no longer works. Attached are two pics I had saved from the posting before it disappeared. You can see that they glued silk screen mesh to a perforated metal sheet with square openings and then mounted it behind the rear stator.

Forum posting(link appears dead now): http://forum.puresound.be/viewtopic.php?f=9&t=447
Company that did the work: Modificatie Audiostatic door Audio 4 elektrostaten


One other alternative is to place on electronic notch filter in line between your preamp and power amplifier.

[esltransformer]

Maybe this link will help to
http://www.wiremeshok.com/Polyester-...nting-mesh.htm

mesh 195 (inch) or 77(metic) will be probably good

Originally Posted by bolserst

The diaphragm resonance is most easily damped by application of acoustic resistance to one or both of the stators in the form of silk screen mesh or thin felt.
This will ruin the see thru optics of the ESL panel(this is very important to some people), but the bass response is much improved.

I posted some measurements of this damping technique here:
Mechanical Sectioning .vs. Silicon dots for resonance control

There had been a thread on another forum detailing the application of silk screen mesh damping to an ES-100. But it looks like the link no longer works. Attached are two pics I had saved from the posting before it disappeared. You can see that they glued silk screen mesh to a perforated metal sheet with square openings and then mounted it behind the rear stator.

Forum posting(link appears dead now): http://forum.puresound.be/viewtopic.php?f=9&t=447
Company that did the work: Modificatie Audiostatic door Audio 4 elektrostaten


One other alternative is to place on electronic notch filter in line between your preamp and power amplifier.

[Bazukaz]

Here is another idea how silkscreen mesh can be fixed to a wire stator with paper clips.
The metal handles of each clip are removed after.

[esltransformer]

That's really ugly.... try plastic U profile.

Originally Posted by Bazukaz

Here is another idea how silkscreen mesh can be fixed to a wire stator with paper clips.
The metal handles of each clip are removed after.

[Bazukaz]

Originally Posted by esltransformer

That's really ugly.... try plastic U profile.

Why? The speaker is covered with cloth, and damping is on the rear side. Paper clips are good for experimentation - easy to remove and hold well.

[esltransformer]

Ok, just for a quick experiment it will do.
The tighter and closer the mesh is to the foil the better it works.

Alfer* deliver U profiles which exactly fit the Audiostatic aluminum rods.


*available at Hornbach (German and Netherlands)

Originally Posted by Bazukaz

Why? The speaker is covered with cloth, and damping is on the rear side. Paper clips are good for experimentation - easy to remove and hold well.

[AAaF]

I would have believed that a cloth would damp higher frequencies?

If so, one alternative would be to just add on backside. If it influence on higher freq, it will also influence on spread pattern, but for all we know it might improve interaction between speaker and room.

I can implement a notch filter, but in general, I believe in "making right from the start" to get the best sound quality.

Brilliant idea regarding the paper clips. We have exactly the same type at work

Really appreciate your input guys!

[AAaF]

Just to verify; you basicly need to add some kind of air flow resistance to dampen resonance? Sounds logical when I think of it