Greetings all from the Jazzman,
I'm here to share my passion for jazz and audio and especially electrostatic loudspeakers (ESL's). The hybrid speakers shown here use elaborate "beam-splitter" transmission line cabinets but a simple wooden frame for the electrostat and an MDF woofer box would do just fine and you needn't spend a fortune for ultra-high fidelity. I'm still amazed that you/me/anyone can take some perforated metal and a sheet of plastic and some foam tape and build a speaker on the kitchen table to rival any $10k high-end speaker on the market. I am most happy to share this; my first electrostatic loudspeaker project.
Enjoy!, Charlie (the Jazzman)
What & Why
My fascination with ESL's started in 2006 when I auditioned a pair of Martin Logan's finest. Their seduction was total but those high-end lovelies were light years beyond my humble 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 not a lot of bucks. Encouraged by those brave souls, I bought Roger Sanders'„¢ 'Electrostatic Loudspeaker Design Cookbook'¯ and dived in. After some weeks of study, I devised a hybrid design using a 10'¯ woofer in a transmission line (TL) box and a 12"x 48" flat-panel electrostat positioned above the woofer.
"Beam-Splitter" Transmission Line:
My initial concepts for a compact TL small enough to share the same enclosure with the stat-panel did poorly in simulation-modeling by GM at the DIY Audio Forum so my choices narrowed to accepting a smaller, less capable sealed or ported box or a larger TL box. The challenge for the TL option was cramming the larger box volume into a reasonable footprint while also placing the woofer at the best location to blend with the stat panel. My solution was extending the box volume upward, behind the stat panel. Of course, a flat/parallel box surface behind the stat would bounce the backwave right back to the diaphragm to bad effect, so I formed the woofer box into a 'V'¯ shaped 'beam-splitter'¯, which splits and diverts the backwave out the open sides of the speaker.
I chose the 10" Aurum Cantus woofer for it'„¢s nice construction and wide frequency response (20hz-1khz) because the woofer's output extends into the lower midrange in this design. I opted to cross the woofer over to the stat panel at [a rather high] 340 hz to mitigate the low frequency phase-cancellation or 'suckout'¯ that's characteristic of dipole speakers in the lower octaves. The woofer is canted upward 6 degrees and mounted at the front of the 4ft3, 9ft folded/tapered transmission line. The bass would be uncompromised in this design.
The speakers are vertically bi-amped using a DBX model 223 active crossover with 24 db slopes feeding a pair of Carver TFM 25, 225 watt/channel power amps. In lieu of a dedicated shelving circuit, the stats' remaining dipole "suckout" is boost-compensated using an Audio Control C101 RTA/EQ. Before testing the speakers with music the woofer/stat balance and frequency response were tuned using the calibrated mic and reference pink-noise generator in the RTA/EQ.
The Moment of Truth
The stats are simply breathtaking and the bass is full and tight with less coloration than any speaker I've owned. In the second tune (Diana singing 'Peel Me a Grape'¯), the bassist does this wonderful riff with a long first note sliding upscale followed by two quick notes stepping down and down again and the Aurum Cantus woofers were right there all the way-- it was just yummy and affirmed my choice for transmission line bass.
The flat-panel stats are incredibly fast and detailed. Even at the lowest volume levels every nuance is heard and they can also produce ear-splitting volume 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.
In the sweet spot, though, the 3D imaging is magical indeed with exquisite detail and stunning speed. A friend described them as 'remote headphones'¯. A woman's voice thru these speakers takes my breath away and Diana's puts me on my knees.
Ironically, what may be perceived as these speakers' great flaw (beaming) is also one of their greatest strengths. That is; flat-panel stats have such phenomenal imaging precisely because they beam sound to the listener directly, with a coherence unmatched by conventional wide-dispersion speakers which smear imaging with multiple high-level reflections off room surfaces. So whether their beaming is considered a fault or a virtue depends on one's perspective.
Interestingly, the stats' sound doesn't seem to decrease much with distance¯ it sounds almost the same from 1 foot or 20 feet. However, the balance between the point-source woofer and line-source stat panel does change with distance, so precise balance only exists at the focal sweet spot. Moving out of the sweet spot in the horizontal plane, the highs roll off progressively and dramatically. Moving to an adjacent room, the sound decreases as expected but seems perfectly balanced and carries quite well.
Inside their sweet spot, no conventional speakers I've heard can match these for realistic imaging, clarity and speed. Moving outside of their sweet spot, they are still clean and listenable but progressively unbalanced and less impressive.
I live alone so I don't have to fight anyone for the best listening position and with the speakers precisely aimed at my usual seat on the sofa, I'm always in the sweet spot and any shortcomings outside of that zone are not an issue for me ¯and I do LOVE their sound. Now, "everything else is just two Dixie cups and a string".
Building the Stat Panels
Perf-metal stators hanging & drying after painting:
Stators are 40% open perforated steel, .048" x 12" x 48" with .125" round holes on 3/16" centers. Mating panels are match-trimmed to precisely align their holes for best performance. The hole perforations have smooth edges on one side of the panel and sharp edges on the opposite side. In order to prevent arcing, the smooth hole-edges must face the diaphragm and any sharp trim edges must be sanded smooth. The stators must also be insulated with paint or powder coat (best) for safe operation. My panels were sprayed with polyurethane primer to build up thickness, followed by a coat of black for color, then a couple of clear topcoats for cosmetics. The total paint film thickness is about 12 mils on both faces and the edges. Latex house primer is also reported to work well. If I had it to do again, I would go with powder coat because it's more durable and wouldn't come off if I ever have to dismantle a panel to replace the diaphragm. Another thing I would do next time is wrap a narrow strip of polyester tape over the stator edges for added protection against arcing. My 40% open panels have 60% of the area carrying electrical charge so they are very efficient and they do sound wonderful. In retrospect, next time I would give up a bit of efficiency for a more open sound using 51% open stators. Also, I would like to experiment with very small holes @ 51% open area, which would require using very thin stators and dampening them to prevent ringing. The sources shown below are for 18 gauge 51% open stators, which is what most builders prefer.
- Perforated metal sources:
- McNichols #1696311841
- McMaster-Carr # 9255T681
Diaphragm on tensioning jig:
The diaphragm material used my project is 6-micron Hostephan polyester purchased from a source in the Netherlands and it took couple of months to clear Customs. 12 micron polyester is also a popular choice and you can get it in a few days from McMaster-Carr.
PET Film Sources:
- 6 micron: More on Hostaphan and Elvamide supplied by The Audio Circuit
- 12 micron (.0005): McMaster-Carr # 8567K104
The diaphragm can be heat-tensioned with a heat gun but I chose to mechanically pre-tension the stator using the pneumatic MDF/bike-tube jig shown above. Jig-tensioning procedure:
- 1) Wrap film over jig & secure with tape on backside.
- 2) Use felt pen to place reference marks 6.000" apart on the film.
- 3) Inflate bike-tube with hand pump to tension the film. When the reference marks measure 6.090" apart, the film is tensioned to the desired 1.5% elongation.
- Caution: At 1.5% elongation the film is approaching tensile failure so DO NOT EXCEED 1.5% !
- 4) Place the stator panel over the film with adhesive side down and press to stick the film to the stator.
- 5) Trim film at edges and remove from jig.
Diaphragm secured to stator with 1/16 thick 3M foam tape:
My stators have 3/4" width tape at the edges with two 3/8" wide inner support-strips. In retrospect, 1" edge tape would have been a better choice to hold the diaphragm tension. The preferred 1" x .063 tape is McMaster Carr part number 7626A115
Jonas Karud at the DIY Audio Forum assembles his stat panels without tape, using magnetic spacers: Jonas uses a separate frame for the diaphragm with the stators held in place on either side using non-conductive flexible-magnet spacers-- this marvelous innovation allows easy and quick disassembly for diaphragm replacement down the road. I like Jonas' setup so much, I'm drawing up plans to retrofit my speakers with magnets. Here's a link to Jonas' speaker project: http://picasaweb.google.com/jkarud/DropBox#.
Minimal diaphragm-to-stator spacing (d/s) is needed for high 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 output. For hybrid speakers, as little as 1/16" d/s spacing can be used but 1/8" or more spacing is needed for full range panels to accommodate their [longer] bass excursion.
The Cookbook guidelines say to support the diaphragm at intervals equal to 70-100 times the d/s spacing but I prefer to set my limit at 70x d/s for added protection against driving the diaphragm into the stators when playing loud. With 1/16" d/s, the span between support strips should not exceed (.062 x 70)= 4.34". I used two vertical tape strips equally spaced across the width of the panel to support the diaphragm.
Diaphragm with the high-resistance conductive coating applied:
The coating shown above was applied with a solution purchased online from Just Real Music but there are a couple of inexpensive DIY recipes in the links below. It's important to mask off the periphery edges before applying the coating, as the coating only needs to extend far enough to contact the copper foil charge ring on the mating stator. If any of the coating were to get on the edges of the tape, it would create a conduction path and short-out the panel. I applied a thin coat using a cotton ball, then let it dry and it was done.
- More on Hostaphan and Elvamide supplied by The Audio Circuit
Mating stator shown with copper foil charge ring:
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. The copper foil tape is part # 7655A711 at McMaster-Carr.
Completed stator panel:
Assemble the mating stators to complete the stat panel, ensuring that the edges and hole perforations are exactly aligned -- piece of cake. The green wire connects the diaghragm to the DC bias supply. The orange and black wires connect the front and back stators to the step-up transformer outputs.
Update: New toroidal step-up transformers!
A Tandem pair of inexpensive Farnell 50VA 230v/2x6v toroidal power transformers step-up the amplifier's voltage to the levels needed to power the stators. These are wired backwards with their amp-side windings in parallel and stat-side windings in series. Theoretically, the voltage step-up ratio is 76:1 but, in praxis, about 68:1, accounting for losses.
Much thanks to Calvin at the DIY Audio Forum for sharing his experience using 230v/6v toroids in ESL's. These sound wonderful and their cost is significantly less than those high-dollar ESL toroids from Aplimo and Plitron.
The single-primary toroids I used are no longer available and I have not found any US sources for similar toroids. However, other builders have reported good results using tandem 2x115v/2x6v toroids with the 6v amp-side windings in parallel and 115v stat-side windings in series.
Be advised that any untried standard toroidal power transformer with dual primary windings carries some risk of shorting between the windings (and damaging an amp) because the strength of the insulation between the windings is not known. Others have reported good results with various toroids having dual 2x115 volt primaries but I have not tried any of them myself so I will only say: Use them at your own risk. That said, these Antek transformers would give the right step-up ratio: http://www.antekinc.com/details.php?p=27
Components on underside of speaker:
Diaphragm's Bias Power Supply:
The simple high-voltage DC power supply shown below is tailored to my speakers' 1/16"diaphragm/stator (d/s) spacing. This circuit could easily be modified for higher d/s spacing simply by adding one or more diode/cap voltage-multiplier stages.
Beam-Splitter Transmission Line Cabinet
Stuffing the transmission line:
The woofer cabinet is a "quarter-wave" transmission line tuned to 30hz. That is; the line is 1/4 the length of a 30 hz sound wave. The line is stuffed with polyfil (pillow stuffing) to a starting density of 0.5 lbs/ft3 to absorb resonances and slow the speed of sound propagation thru the line (further lower the tuning frequency). The shorter waves are absorbed as heat by the stuffing while the longer waves are slowed to exit the port in-phase and support the low-bass output. The tonal quality can be further tuned by adding stuffing to tamp down any "booming" or removing stuffing to liven up the tone if it sounds muffled. The line is 9 ft long, folded and tapered, using the classic "Cookbook" sectional-area formula: Area at the front of the line is 125% of the woofer's driven area and tapers to 100% of the woofers' driven area at the port end. A properly designed transmission line bass has superb transient response and goes 1/2 octave lower and with less resonance than any other type of enclosure. These characteristics make TL's a good choice for hybrid-bass ESL's. Their large volume, complexity and expense is why you don't see more commercial TL speakers.
Assembling the stat frame to the woofer cabinet:
Cabinet ready for the oak trim:
The cabinets are made from 3/4" MDF with 5mm red oak plywood sheathing and trimmed with solid red oak. These cabinets were very time-consuming to build but the results are equally pleasing.
Completed beam-splitter cabinets ready for the stat panels:
15" wide, 67.5" tall, 21.5" depth