Everything else is two Dixie cups and a string :-)

















Greetings all from the Jazzman,
I'm here to share my passion for jazz and hi-fi and do-it-yourself electrostatic speakers.  My electrostat panel and electronics are simple and the beam-splitter design combines a line source electrostat with transmission line bass in a compact, affordable package.        

It still amazes me that an average Joe with no electronics experience can build a speaker on the kitchen table that rivals the $20k commercial offerings.  And building the actual drivers from scratch takes the cool factor to a whole new level.       

The links at the bottom of the page provide detailed build instructions, component sources, a newer version of the beam splitter design and another design called the Eros Clone.   

I am most happy to share with you my DIY electrostatic loudspeaker project.


Charlie Mimbs
Savannah, GA
jazzman1953@gmail.com


How does it work?


















If your hair has ever stood on end while unloading a clothes dryer, then you've felt the same force that drives an electrostatic loudspeaker (ESL).

An ESL is a push/pull motor powered by electrostatic attraction/repulsion.  The heart of the motor is a plastic diaphragm suspended between two conductive screens (stators).  A separate power supply puts a fixed DC bias voltage on the diaphragm, and the music output from an audio amplifier, routed thru a pair of step-up transformers, puts the driving AC voltages on the stators.  And the ultra-light diaphragm responds with instant precision to reproduce the music with exquisite fidelity.

The woofers and stat panels are controlled and powered separately using an active crossover upstream of a pair of stereo power amps.

Warning:

The charges on the stators are dangerous. If the insulating coatings are not essentially perfect, touching both stators while the panel is playing could stop your heart.  Also, the high-voltage bias supplies should be enclosed within the speaker cabinets for safety. With children or pets in the home, protective speaker grills are a practical necessity. 


How it all began:
My fascination with ESL's started in 2006 when I auditioned a pair of Martin Logan Summits. Their phenomenal clarity and speed seduced me instantly but their price was light years beyond my meager budget. Like Gollum coveting the ring, I was ruined thereafter; not even wanting to power up my home system again... it was a dark time.


New hope came while browsing the DIY Audio Forum, where I found guys rolling their own electrostats for cheap.  So I got myself a copy of Roger Sanders’ Electrostatic Loudspeaker Design Cookbook and dived in.  After some weeks of study my hybrid design emerged, with a 10” woofer in a unique beam splitter transmission-line, and a 12"x 48" flat panel electrostat above the woofer.

Beam-Splitter Transmission Line ESL (original version):

The Beam Splitter design:
Transmission line bass requires a huge woofer box, and the challenge was cramming all that volume into a small footprint and mating it to a tall stat panel-- all in a slim package.  Forming the woofer box into a V-shaped beam splitter allowed extending its volume upward, behind the stat panel, as the splitter deflects the panel's rearward sound out the open sides of the speaker rather than bouncing it back to the diaphragm.  

My criteria for a woofer with the speed to blend with the ultra-fast ESL were low QTS, low moving mass and above all a strong magnet coupled with a low-inductance voice coil.  I chose the Aurum Cantus AC250 MkII. 

The woofer is angled 8 degrees upward and couples to 9-foot transmission line.  Behind the woofer, the line transitions to vertical in a gentle curve that prevents resonances rebounding to the cone.

Power and Control:
Passive crossovers are not suited to an ESL's capacitive, frequency-dependent impedance so I recommend active crossovers exclusively.  Besides being superior in all respects, an adjustable active crossover makes it easy to precisely balance the woofer and stat panel outputs in real time.     

My speakers are vertically bi-amped using a Behringer DCX2496 digital crossover feeding a pair of vintage Carver TFM25 power amps.  The crossover frequency is 220Hz via the 48db/octave Linkwitz-Riley filter.

The Moment of Truth:
I fired these babies up on July 4, 2008 with my goddess Diana Krall singing Nat King Cole's "You're looking at me". The illusion of Diana performing live in my living room was so spooky-real I could almost smell her perfume. Finally I have speakers worthy of Diana, who now sounds as good as she looks :-)

The electrostats are breathtaking and the bass is full and deep with gorgeous tone. In the second tune (“Peel Me a Grape”), the bassist does this wonderful riff with a long first note sliding upscale followed by two short notes stepping down, and down again, and the Aurum Cantus woofers were clean and tight, all the way down... it was just jaw-dropping.

Observations:
The flat-panel electrostats are incredibly fast and detailed.  Even at very low volume, every nuance is heard and they can play ear- splitting loud with no loss of accuracy or hint of distress. They are, however, ultra-directional, with a sweet spot only about a foot wide; so, not a good choice for party speakers.  Within their focal sweet spot there is surreal imaging, exquisite detail, and stunning speed.  Moving out of the sweet spot, the beaming highs fall off progressively.  My friend Jason describes them as “remote headphones”.  A woman’s voice thru these speakers just takes your breath away.

A flat-panel ESL's beaming certainly constricts the sweet spot but it's this very characteristic that projects a fully-coherent wavefront, with practically none of the energy  bouncing off walls to interfere and confuse the imaging.  So while beaming may be considered a fault, the sweet spot is magical indeed.    

Conclusion:
No conventional speakers I’ve heard can match these for holographic imaging, clarity and speed. Not even the ML Summits can match their imaging. Moving outside the sweet spot, they are still clean and listenable but progressively less magical. Ultra-directional speakers don't suit everyone, but I live alone, so I'm free to hog the sweet spot, and I love their sound. Now, "everything else is two Dixie cups and a string".

Thank you:
Roger Sanders for your marvelous Cookbook.
Mark Rehorst and Sheldon Stokes for your inspirational DIY ESL websites.
GM at the DIY Audio forum for steering me away from my original, poor bass box design.  Calvin, Few, Jer and Bolserst at the DIY Audio Forum for sharing your knowledge with me.


Charlie Mimbs
Jazzman1953@gmail.com


Videos:

My article in AudioXpress magazine:
(when the sign-up box opens, scroll down and click the
"free preview" link to open the magazine, then tab to page 14
to view my article)


http://jazzman-esl-page.blogspot.com/2009/11/new-beam-splitter-esl-aug-2012.html

Building the Stat Panels



Exploded view of ESL panel (click to enlarge):





Corona:   


The photo above shows a metal pinwheel charged to a high voltage, and the resulting corona streaming off the sharp points.  Likewise; any sharp points on the edges of a charged stator would exhibit the same corona effects, which would quickly burn through the paint coating and create an arc path to the mating stator-- causing the panel to short-out and fail. This is why we must use great care to remove all sharp points along the trimmed edges of our metal stators!             

Choosing perforated metal for the stators:
I prefer steel over aluminum because the electrical leads can be soldered to it, it's extra weight makes it less prone to ringing and it costs less. When choosing hole size and open area, consider how the stators will be coated: A spray-painted stator with 12 mils of paint on its faces will likely have only about 3 mils on the inside hole surfaces because a paint gun can't be oriented perpendicular to the inside-hole surfaces. Powder coating uses electrostatic attraction to place the coating, so there would likely be more coating in the holes compared to painted stators. For that reason, perf with higher open area (50% minimum) is needed for powder coated stators.

Stator coatings:
Powder coating is generally considered the best option but painted coatings (laquers, polyurethanes, epoxies, i.e. Krylon, Rustoleum, etc...) have been used with success. I think most any non-metallic paint would work if applied in sufficient thickness.

I opted to spray coat my stators with automotive polyurethane because it chemically sets and I had equipment and experience to apply it myself.  I saved myself some time and hassle but automotive polyurethane isn't cheap ($125+ with cleaning solvents), so I probably didn't save any money over the cost of powder coating.

If spray coating the stators, the electrical leads can be soldered on before painting, as I did.

If powder coating the stators, the 400-degree baking step would melt a solder connection. A metal tab connector could be brazed on before powder coating or a tab could be mehanically attached [before or after powder coating], as shown in this thread: mechanical connection

The instructions below are for spray-coating steel perf stators with pre-soldered leads.

Guidelines:  Steel Perf stators should be 16-20 gauge (.063-.036) thick with 40%-60% open area and hole diameters at least 2X the metal thickness but not larger than 3/16".

Stators with smaller holes are more efficient because they produce a more uniform and denser electric field. However, smaller holes require thinner stators (1/2 hole diameter, max); otherwise the sound sees the holes as cylinders, and the air within the cylinders has an impedance that acts to roll off the treble response. So, ultimately, the stator thickness sets the lower limit for hole size. 

Stators also vibrate when playing (especially flat stators), and I suspect thin stators may exhibit audible ringing and possible distortion at their resonant frequency. So, for flat stators, I prefer to go no thinner than 20 gauge (.036") steel. 18 or 20 gauge is my preference.

There is also much debate over how much open area is best. I have only used 40% and 51% open stators and I don't hear a lot of difference-- the 51% panels seemed slightly brighter and the 40% panels were perhaps a bit softer (it could be that less open area constrains the air flow enough to dampen the diaphragm's fundamental resonance and harmonics but I can't say for sure). Some of my stator photos show 51% open perf but I currently prefer and use 18 gauge 40% open steel perf with .125 diameter holes. Again, for powder coated stators, 51% minimum open area is best because the powder coat will make the holes significantly smaller.

Links to perforated metal sources:

(best prices and my preferred source).
036 x 12 x 48 / .125 holes / 40% open: $20 each + shipping.


Trimming stators to size and working them flat:
Most suppliers will trim the metal to size; otherwise a band saw works well. The panels will need to be as flat as possible and some suppliers can roll them flat if you specify that flatness is a requirement.  None of the perf I've ordered arrived perfectly flat. The perf from McMaster Carr was the flattest I've gotten and even those pieces were slightly bowed.

After match-trimming and edge-smoothing the stators, you will likely lose about 1/8" of width so you will either need to order the metal 1/8" oversize in width, or make allowance for it in your speaker frames.  If you order other than 48" length, the ends will also have open-hole edges and require smoothing.   I advise building the stat panels first, then cutting the speaker frames to fit the panels.

The hole perforations are punched thru the metal (not drilled), and that process produces smooth hole edges on one face of the panel and sharp hole edges on the opposite face. The panel face with the smooth hole edges must face the diaphragm to minimize the chances of coronal arcing.

As already shown, any sharp points on the stator edges are prone to coronal arcing and should be ground off smooth.  

Place the mating stators together with their smooth hole edges facing the inside and their holes exactly aligned, then clamp them together and grind smooth all sharp points on the edges.

After edge finishing, inspect each panel for flatness. It's very likely that you will need to do some final straightening. I straighten my stators by hand on a flat work table covered with a 1/8" thick soft rubber "router-mat"purchased from Harbor Freight tools. The rubber mat has enough give to allow straightening the metal without over-bending it.

CAUTION: If you try to flatten out a dent in a stator with a hammer on a hard surface, chances are you will expand the metal; causing it flop in and out ("oil canning") and it would then be ruined.




Right: After trimming to size, the front and rear stators are aligned and clamped together, then sharp points along the edges are ground down and rounded-over to prevent coronal arcing.











Left: Straightening a bent stator by hand using a soft rubber mat under the stator to allow it to "give" without over-bending. For straightening sharp creases in the metal, a rubber or wooden mallet can be used. The rubber mat is a router mat purchased from Harbor Freight tools.








Cleaning and coating the stators:

For safety reasons alone, I wouldn't use uncoated stators. Even with coated stators, I would use speaker grills if there are children or pets in the home.
Before painting the stators, the wire leads must be soldered on. Mine have the power leads soldered to an outside corner. There are other coating options but my stators are spray-coated with a 2-part automotive polyurethane. First,the stator panels were thoroughly cleaned to remove the protective oil coating, then sprayed with Martin Senour Crossfire paint system with 1-part base-coat/2-part clear-coat from NAPA auto parts. I used 1-pint of thinned black base coat and a about 1.5 quarts (mixed volume) of the 2-part polyurethane clearcoat. If using a HVLP paint gun, you might get by with a single 1-quart kit of the 2-part clearcoat. I also used the fast dry catalyst and sprayed the clearcoat "just wet" so that the coats would set quickly and not thin-out over the sharp edges of the perforations. Here is the procedure:
  • Hang panels and attach weights to keep them steady while spraying.
  • Thoroughly spray-rinse panels with Naptha or laquer thinner to clean off the oil coating. Allow to dry.
  • Spray on the 1-partcolor coat as needed for coverage. Allow to dry at least 1 hour.
  • Spray on 10-14 mils of 2-part polyurethane clearcoat. Be sure to coat panel edges as well as faces and spray from different angles to cover edges of perforations. The first couple of coats should be "misted" on withat least 10 minutes tack time between coats. Spray remaining coats"just wet"; allowing a few minutes tack-time between coats.






Left: Stators hanging ready for solvent spray cleaning and coating.







Left: Closeup of a completed stator with polyurethane coating applied. Note that edges are rounded over to prevent coronal arcing.

Diaphragm/Stator Spacing & Span Between Support Spacers: 
Minimal diaphragm-to-stator spacing (d/s) is desirable for highest efficiency because the electrical field strength decreases with the square of the distance. Thus, doubling the d/s spacing would require four times the power input to produce the same volume. For hybrid speakers, as little as 1/16" (about 1.5 mm) d/s spacing can be used but 1/8" (3 mm) or more spacing is needed for full range panels to accommodate their longer bass excursion.

The Cookbook guidelines recommend placing diaphragm supports at spans equivalent to 70-100 times the d/s spacing. I prefer to go conservative and set my limit at 80x d/s for added protection against bass driving the diaphragms into the stators. With 1/16" d/s and 80x max spacing, the span between support strips should not exceed (.062 x 80)= 4.96". Since my 12" wide panels have a span between the edge spacers of 10.5" and I'm adding two 3/8" wide vertical support spacers, I end up with three sections, each 3.25" wide, which is well within the limits, so there is no problem.

Spacers:
Note:  The photos show my original panels, which used 3/4" wide urethane foam tape spacers.  I now recommend 1" wide tape only, with 1/4" of the tape wrapped over the stator edges as shown here: 


























Many builders use plastic spacers and glue but I prefer to use 3-M brand .063"x 1" double-coated urethane foam mounting tape because it's fast and easy. The tape secures the diaphragm to the stators instantly with minimal fuss and sets the diaphragm/stator spacing at 1/16"; which is ideal for hybrid panels. The 3-M tape isn't cheap but I would not use any other brand. Also, you want to buy the tape from a source that sells enough of it to keep their stock rotated and fresh. 3-M advises using the tape within 2 years of manufacture in order to retain its full tack and adhesive strength. 3M has said that the foam tape will yellow over time but should last for about 25 years indoors (Really?... I don't know... just reporting what 3M said).

Apply 1" wide foam tape along the periphery edges of both stators; allowing the tape to overhang the edges 1/4".  The 1/4" overhang will be wrapped over the stator edges after bonding the diaphragm. 

Also apply 3/8" wide tape support strips at the intervals needed to support the diaphragm. Note in the the photos that my 12" wide panels use two 3/8" wide diaphragm support-strips. The tape sets the diaphragm-to-stator spacing (ds) at 1/16", which is the preferred spacing for hybrid speakers. The 1" x .063 double sided foam tape is McMaster Carr part number 7626A115

Diaphragms:


Left: Diaphragm shown being tensioned on a pneumatic bike-tube tensioning jig.

Diaphragm material:
I used 6-micron Hostephan polyester purchased from The Audio Circuit (TAC) in the Netherlands. TAC ships immediately but it can take several weeks to clear Customs.
Later I found 6-micron Mylar on Ebay for a great price and it arrives in about 10 days. I recommend the 6-micron film but some builders use the heavier 12 micron film for bass panels because it can be tensioned much higher. If using a mechanical tensioning jig, you will need film wide enough to wrap around the jig. Several sources and sizes are listed below.

Links to sources for polyester film diaphragm material:

Current best deal:

Other sources:
McMaster-Carr #8567K104 - 12-micron 27" x 25' roll $12.25 + shipping
Diaphragm Tensioning:

Tensioning the diaphragm is a compromise between conflicting requirements: If you use a high polarizing voltage on the diaphragm for highest efficiency you will need to tension the diaphragm very high to prevent electrical forces from pulling the diaphragm into a stator. Also, higher tension reduces the chance of bass driving the diaphragm into a stator (the pressure waves from any woofers in the room will couple to and move the diaphragms). On the flip side, lower tension gives a lower diaphragm resonant frequency; which is also desirable, as that would allow a lower crossover point. The diaphragms resonance depends on it's tension and the span between diaphragm supports. Typically, the diaphragm resonance will fall somewhere between 50-120 hz. With fairly close support spacing, as used in my panels, the resonance is probably somewhere near 100 hz; although I have not measured it.

We don't want to play the panel at frequencies low enough to excite the diaphragm resonance, as output is savagely distorted at the resonance. The rule of thumb is to set the crossover at least two octaves above resonance if using a 24db slope or at least one octave above resonance if using a 48db slope.

The amount of tensioning needed and the methods of achieving it are subjects of heated debate among builders. Some recommend heat shrinking and others prefer mechanically stretching the film. I even read somewhere that Quad at one time used mechanical stretching followed by a proprietary heat treat process to "stabilize" the tension.

I use a pneumatic "bike-tube" tensioning jig made of 3/4 MDF, sized 1" larger than the stator on all sides and the edges are 2" thick. (see the photo) There is a hole drilled on one end for the bike tube's valve stem. The bike tube is a size 27 x 1.25 (700 x 32 metric) and has a Schrader type valve (don't use a Presta-valve tube because it's valve isn't spring loaded and will leak down). The general consensus for mechanical tensioning is to stretch the film to 1%-2% elongation, depending on the span between the support strips. For my panels, I use 6-micron diaphragms exclusively and I tension the film to 1.5% elongation.

With 12 micron film, much less elongation would be needed to reach the same tension as 6-micron film stretched to 1.5% elongation. Conversely, a thinner film would need more elongation to reach the same tension. I can only vouch for 1.5% elongation if using 6-micron film.

Mechanical pre-tensioning procedure for 6-Micron polyester film:
Before tensioning the film on the jig, have the stators ready (3M foam tape spacers and center support strips in place).

1) Prepare the jig: Apply double backed tape on the underside edges to hold the film in place.
Also adjust the top edge of the bike tube flush with the top surface of the jig.
2) Cover your work surface with a bath towel or cloth to protect the fragile film.
3) Cut a piece of film large enough to cover and wrap around the jig with a couple of inches to spare on all sides. Use very sharp scissors to cut the film cleanly, as any ragged edges are prone to tearing when handling the film.
4) Layout the film onto the prepared work surface and place the jig, face down, onto the film.
5) Wrap the film over the jig edges and secure it on the backside with double backed tape. Be
very careful here-- the film is quite strong in uniform tension but tears quite easily from the
edges. If you do happen to tear the film along an edge and the tear does not extend onto the face of the jig, you can still save it by patching the tear with tape.
6) After the film is secured to the back edges of the jig, you want to pull out any slack at the four corners of the jig-- use scotch tape but don't extend the tape over the bike tube.
7) Carefully turn the jig right side up. The end of the jig that has the tube valve can hang off the edge of the worktable as needed to attach the air pump.
8) Using a fine tip felt pen, place reference marks on the film exactly 6.000" apart, in the width direction. (If the film is wide enough, you could place marks 12" apart)

In case you were getting complacent -- Hold your breath and pray during the next step because you will be stretching the film almost to its breaking point.

9) Slowly inflate the bike-tube to tension the film. When the distance between the reference
marks reaches 6.090", the film is tensioned to 1.5% elongation. (If you used 12" marks, stretch the film until the marks are 12 3/16" apart for 1.5% elongation). You can breathe again now.
10) Immediately bond the stator to diaphragm while the diaphragm is under tension (see
below).
Bonding the stator to the tensioned diaphragm:
1) While the film is under tension, place the stator panel over the film, adhesive side down, and
press along the edges and center support strips to secure the film to the stator.
2) Release pressure and trim the excess film net to the edges of the foam tape.


As previously noted, the edgesof the stators are prone to coronal arcing, which could short-out the panel and even damage the transformers and power amp. Enlarge the photo below and note that the foam tape overhangs the edge of the stator 1/16" on all sides. Do this on both stators because it provides an extra measure of insulation that virtually guarantees that no arcing will occur between the stators.




Left: With the diaphragm fully pre-tensioned on the pneumatic jig, the stator is placed over the diaphragm and pressed down. 1/16" thick 3M double-sided foam tape secures the diaphragm to the stator.







Left: After bonding the diaphragm to the stator, the excess is trimmed away and the diaphragm is now ready for the conductive coating.



After bonding the diaphragm to the stator, I recommend
tape insulating the stator edges as shown in the link below:

Below:
Techspray conductive coating applied to the diaphragms (still wet):
In order to prevent the bias charge from leaking off the diaphragm, we want to leave the outermost edges of the diaphragm uncoated. The coating only needs to cover the inside areas and extend far enough to touch the copper foil charge ring on the opposite stator (see diagram at top of page). So, before coating the diaphragm, we want to mask off the periphery edges, about 1/8 to 1/4 inch, as shown in the photo below (blue masking tape).


Popular conductive coating options:
One light coat sprayed just wet enough to form a continuous film will dry to a clear permanent coating 1.5-2.0 microns thick with E7-E9 conductivity. This coating is tried and true and I use it exclusively.

2) Any liquid hand-dishwashing detergent containing sodium laurel sulfate: Apply a light coating, undiluted, in circular fashion, using a cotton ball or soft cloth. Some builders recommend this coating because it's easy and cheap and has ideal resistance reported to be E9-E11 range but I have no means to measure it. I did try it out on a test panel and it does work perfectly but I don't know how long it would last. Tape and adhesives won't bond to this coating so before applying it you would need to mask off the areas on the diaphragms that bond to the center support strips/ spacers.

3) Nylon in solvent (Elvamide); More on Hostaphan and Elvamide supplied by The Audio Circuit: Said to be the best coating around but it looks to be a pain in the ass to mix and apply.

4) DIY coatings posted on the DIY Audio Forum-- I haven't tried these:
http://www.diyaudio.com/forums/showt...17#post1837517
Copper charge ring:
Copper foil tape applied to the periphery of the mating stator makes the electrical connection from the high-voltage DC bias supply to the diaphragm as shown below.


Soldering the lead to the copper foil charge ring




Charge ring installed on the mating stator



The connecting wire must be soldered to the copper foil before it's applied to the stator panel. In order to reduce the risk of arcing and shorting, it's very important to keep the solder joint as thin and flat as possible so that the insulating tape is not excessively compressed when the panels are assembled. I carefully sanded away the excess solder to flatten and thin the connection before transferring the copper strip to the stator.


Below: Stators being positioned for assembling the panel
This is a tricky operation because the foam tape makes a permanent bond on contact so the stators must be flat and exactly aligned when contact is made. Since I didn't have any help, I found it easier to position the panels vertically using a board clamped to my table and a bank-stop at one end, as shown.



For mounting the panels in the speaker frame, I use a firm foam rubber weatherstripping between the ESL panels and frames. This insulates the panels from the frames, not only electrically but also acoustically, as the stators vibrate somewhat in operation.  Click the "Insulating the stator edges" link below to see how the panel edges are insulated and mounted in the wooden frame.