Yet another ESL project

Dedicated to those large boxes at one end of the room
Post Reply
Jazzman53
User
Posts: 53
Joined: Fri Sep 12, 2014 1:00 am
Location: Savannah, GA, USA
Contact:

#1 Yet another ESL project

Post by Jazzman53 »

The speaker building psychosis got the best of me again, folks. This time it’s flat panel electrostat that’s electrically segmented for wide dispersion.

Image

This new project was not a whole new speaker but only replacement panels for a speaker project I posted on this forum a few years ago in this thread:
http://www.audio-talk.co.uk/phpBB2/view ... f=4&t=5786

The knock on flat panels is their narrow dispersion, which some describe as the “head-in-a-vise” sweet spot. My new flat panels eliminate the head-in-a-vise effect entirely and they can even be configured with switch-selectable wide and narrow dispersion modes by installing a switch that jumps over resistors in the RC network.

Flat panels and dispersion:
When a speaker’s radiating surface is wider than the wavelengths of the frequencies it’s radiating, those frequencies beam to a tightly focused sweet spot rather than spreading out. This is great for solo listening at the focal point but not so good when guests drop in.

The most common method of achieving wider dispersion from an ESL is curving the panel. A less common but more effective method is segmentation.

My segmented panel:
The stator conductors driving the diaphragm are (90) vertically arrayed, single-strand 20awg copper wires with .010” XLPVC insulation and 43% open area spacing. The wires are physically divided into fifteen 6-wire groups, which are electrically powered as eight discrete, symmetrically arranged groups-- consisting of the center 6-wire group and seven left/right paired 6-wire groups arrayed on either side of the center group.

The center wire group receives the full audio spectrum (above 200Hz where the woofer crosses in). Moving outward toward the panel edges, the left/right paired wire groups receive progressively reduced frequencies via an RC transmission line. The RC line consists of resistors inserted between the wire-groups, which couple with the wires' capacitances to form first order filters that step down the frequencies.

As driven by the segmented stators, the diaphragm radiates the top octave from only a narrow band in the center of the panel, and radiates progressively less treble moving outward from the panel centerline. Thus; the widths of the diaphragm’s radiating zones remain below the wavelengths being radiated and this results in the sound spreading out to a nice wide sweet spot.

The build process:
The copper wires were stretched on a special jig to plastic deformation (1% elongation), which renders them perfectly straight and they remain straight when the tension is relaxed.

Before stringing the wires into the stretcher jig, wax paper was laid down to prevent gluing the wires to the jig when assembling the supporting lattice. The wooden support lattice is made of interlocking red oak pieces. The lattice pieces were assembled and glued onto the wires, in the jig.

Wires in the stretcher jig:
Image

The interlocking oak support lattice (not yet glued together):
Image

Assembling / gluing down the lattice over the wires:
Image
Image

Completed stator:
Image

Bonding the front stator to the pre-tensioned diaphragm, on the pneumatic (bike tube) diaphragm stretching jig:
Image

Diaphragm bonded to front stator:
Image

Front & rear stators ready for mating:
Image

The RC network wired in:
Image

Schematic:
Image

Completed speaker:
Image
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8867
Joined: Mon May 21, 2007 9:42 pm

#2 Re: Yet another ESL project

Post by Paul Barker »

Very nice, what frequency range of the esl part?
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Jazzman53
User
Posts: 53
Joined: Fri Sep 12, 2014 1:00 am
Location: Savannah, GA, USA
Contact:

#3 Re: Yet another ESL project

Post by Jazzman53 »

Hi Paul,

The system is bi-ampd using a DSP crossover. The electrostat panel plays from the 218Hz crossover frequency to 20kHz where the DSP chops off the bandwidth.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8867
Joined: Mon May 21, 2007 9:42 pm

#4 Re: Yet another ESL project

Post by Paul Barker »

That sounds great. Shame you're in another continent. Would be well worth a listen.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Jazzman53
User
Posts: 53
Joined: Fri Sep 12, 2014 1:00 am
Location: Savannah, GA, USA
Contact:

#5 Re: Yet another ESL project

Post by Jazzman53 »

The editor of AudioXpress magazine saw my website and invited me to write an article about my latest speaker, which I did, and that article was just published in the September 2017 issue. I'm a bit disappointed my speakers were not on the cover (commercial speakers won out there) but they did a very nice 8-page layout inside. I can't show the article here of course but below is an image of the Cover and first page:
article pic.jpg
User avatar
ed
retired
Posts: 5384
Joined: Thu Jun 21, 2007 4:01 pm
Location: yorkshire
Contact:

#6 Re: Yet another ESL project

Post by ed »

Awesome

What is the stator to film gap measurement?, apologies if it's mentioned in the text, but I didn;t see it.

weren't you tempted to crossover a tad lower?
There's nowhere you can be that isn't where you're meant to be
Jazzman53
User
Posts: 53
Joined: Fri Sep 12, 2014 1:00 am
Location: Savannah, GA, USA
Contact:

#7 Re: Yet another ESL project

Post by Jazzman53 »

Hi Ed,

The d/s (diaphragm-to-stator gap) in my speaker is 0.063" (1.5mm), which is the practical minimum for a hybrid panel and gives very high efficiency for an ESL.

The choice of crossover frequency is a tradeoff between max output and EQ'ing the dipole rolloff. The dipole rolloff becomes significant for wavelengths longer than about 4x the panel width. Below this threshold frequency (wavelength), increasing amounts of EQ'ing (bass boost) is required to offset the rolloff, which in turn sets a limit for how loud the panel will play before the diaphragm is driven into the stators. It's also important to use a high quality, low inductance woofer that can play cleanly to least 1kHz in order to best match the lower Q of the stat panel.

I've heard 12 inch wide hybrid panels crossed over at 150Hz and they sounded quite anemic in the mid bass region and would not play nearly as loud as my panels before striking the stators.

BTW, complete build info, and parts lists for several ESL projects can be found on my website here:
http://jazzman-esl-page.blogspot.com/
Post Reply