Home‎ > ‎

Project History

On mature reflection

posted Sep 12, 2010, 5:40 AM by Chris Fox   [ updated Sep 12, 2010, 6:24 AM ]

Made up a second panel and set it up. I've been listening to these panels for a month or so now in what I think is a well sorted two channel setup. The crossover and EQ are done by the Behringer DCX2496 with X-over at 400 Hz, asymmetrically -- 24 dB/octave high cut for the woofer and 48 dB/octave for the ESL. Asymmetrical because there's a 6dB/octave low pass filter active for the ESL below 1kHz so I figured there's a rising character to the high pass section to start with so the x-over will need to be more savage. Not very scientific, but it sounds fine. A mate of mine is going to test them with his DEQX unit and calibrated mic so the setup should be a bit more objective after this.

Bass up to 400 Hz by the way is courtesy of Scan Speak Revelator 10" aluminium cone woofers in 63 litre vented boxes, mainly because I had them on hand. Not supposed to use vented boxes with ESLs because they're not 'fast' enough, but this combination sounds pretty good to me.

As to how they sound, I think they're just lovely, but then I guess their daddy will always love them. In particular I love the way the sound just hangs in space between the speakers. They have a breathy ethereal sound, but can produce startlingly loud and dynamic sound when wound up. Nice at any level, and very calm and easy speakers. The Lowthers (EX3s in TP1 horns that make up my other system) tend to get 'bitey' if the volume is wound up too high, or the recording is too dry and hard, whereas the ESLs remain sweet on almost any material . Voice is very natural. Piano sounds cohesive across the whole keyboard too, so frequency response can't be too bad. Transients are clean -- for example the Floyd "Time" track reproduces the alarm clocks very cleanly. 

Imaging is good without being as holographic as either the flat panel ESLs that these have replaced or the full range Lowthers. The music does float nicely between the speakers. Great soundstage.

No measurements so far, but a wide range of listeners have auditioned them and comment has ranged from impressed to startled. I'd call them a great success. There's two more sets to build now for the other shareholders in the group.

Noticed some minor wrinkling in the Mylar on both speakers. On one panel it's quite localised, on the other it's more general. The first panel was 'brushed' with the heat gun to remove any wrinkling evident after the panel was cut loose from the stretching table. This tensioned the Mylar across the panel a little more than I'd like, but it did remove any stray wrinkles very quickly. The second panel I left with the minor corrugations in the Mylar without using the heat gun. Listening tests didn't show any difference between the two. Maybe a bit of minor wrinkling doesn't matter? 

Panel spacing problem and fix

posted Jul 28, 2010, 12:17 AM by Chris Fox

After some listening I decided something was amiss. I had a hard look at the panel and noticed the DS tape horizontal spacers weren't touching from opposite sides as they should. Panel was as in A in the diagram. A few minutes of gentle bending and massaging the rear stator put more curvature on it, as in B. When clipped together the two panels then had good contact between the DS tape insulating spacers at all points.

There was a noticeable improvement in sound quality after this little remediation. Without being definite about it I reckon there was some distortion of some sort before. Don't know if it had been there all along and hadn't noticed it, or whether the panels moved around a bit. 

I'll set the next panels up with a definite bulge in the rear stator to make sure this doesn't happen again. Another way to deal with this would be to use the DS tape from both stators to bond both stators together through the membrane, but I'd like to keep the two sides separable for cleaning purposes.

Assembling the panel and first listen

posted Jul 25, 2010, 5:34 AM by Chris Fox   [ updated Jul 25, 2010, 9:07 PM ]

While the glue dried it was time to put the diaphragm contact onto its stator. This is the stator without the diaphragm glued to it. In this case it's the inside or rear stator. 

The contact is made using self-adhesive copper foil which makes an open loop around the outside of the stator. It's brought out and terminated at a screw terminal which is super glued to the rear stator using Loctite 438.



When the diaphragm contact was fitted it was time to free the stator that had the diaphragm fitted to it. This meant cutting the tape holding the stator to the table, then cutting the excess membrane off using a pair of scissors. I'm not sure how the really thin Mylar would cut with a sharp blade. I can imagine it ripping.

I was a bit horrified to see some wrinkles in the Mylar near the corners and at the ends of the panel. At first I thought the glue had let go, or possibly wasn't fully cured. Careful examination showed the glue was fine, it really was wrinkled Mylar. Remembering that Mylar could heat shrink I tried heating the Mylar gently using a paint stripping gun on low. The wrinkles immediately disappeared, so I was careful not to heat it too strongly.

If you look at reflections in the film you can see some bowing of the Mylar inwards, so there is clearly some tension across the panel. I didn't attempt to measure this bow, but it must be well below 0.5 mm.

Next the diaphragm was coated using ER Audio's diaphragm coating which is water based and transparent. I made up a spreader from a couple of icypole sticks and some scrap plastic foam. The coating is diluted 1:4 with distilled water and spread over the exposed side of the film. 

After the coating has dried the panel was assembled. The panel is clipped together with paper hanger extrusion. This is the plastic strip used to hold posters straight. You can buy this in news agencies or poster suppliers. I got mine from ER Audio. The extrusion was a  bit tight to slip on a whole length, so for the prototype in my impatience I cut the extrusion into shorter lengths and slipped these onto the sides and ends of the panel. To make it a bit easier to slide them on I warmed them up with the heat gun for a few seconds. Some of them got a bit too warm and became soft, but the end result was fine. The panel is a bit too wide for the extrusion really. I'll probably try to make a better clip, maybe out of a length of PCV tubing with a slot routed along its length. 

The panel was stood up on top of some Scan-speak woofers that used to be part of the old hybrid ESL system. Using the electronics from the old ESL system, connection was made to EHT power supply and transformer and we were away. Music. Crossover is at 400 Hz, so the panels are not working too hard. There's 6 dB/octave EQ from about 1 kHz down using some old analog bits with ordinary old TL072 op amps, and I'm using a 185 W RMS per side MOSFET amp. Nothing fancy or expensive. Next step is to put a Behringer DCX2496 into the system as crossover and EQ.

In the background you can see a pair of Lowther TP1 corner horns with EX4 drivers. They're the other system in the home. Totally different sort of sound but great listening too. see https://sites.google.com/site/lowtherhorns/ for the full story of these speakers.

The sound is pretty impressive I have to say. The sheer amount of output is surprising. These ESLs are really quite LOUD, much to my amazement. My old system always struggled a bit for output. These speakers are a bit more forward than my old ones too. I put this down to the different diaphragm material. The old ones used some third party Mylar equivalent. I noticed this same forwardness in a friend's flat panel system too. Same Mylar in his system.

Dispersion was pleasing. The sweet spot is wide enough to be able to move the head by half a metre from side to side when sitting about 2.5 metres away from a single speaker without losing the treble completely. None of that 'beaming' that is so annoying about flat panels.

There's a certain spareness about music played through ESLs. They really give you just the music. Voices are feathery and clear and the overall impression is of accuracy and smoothness. Treble is terrific -- cymbals are very natural and the whole thing is very easy. It's possible to listen to them for hours without any sense of fatigue. 

Diaphragm stretching and gluing

posted Jul 24, 2010, 7:23 PM by Chris Fox   [ updated Jul 24, 2010, 8:40 PM ]

A sheet of Mylar was cut to length on the stretching table, and a dozen or so 6 " lengths of clear packaging tape were cut and stuck to the edge of the table. 

Each corner of the Mylar was tacked down to the table with minimal tension applied. Left untethered, the Mylar drifts around uncontrollably. Even walking past it is enough to make it ripple and fold. 

The centre of one end was taped down. A 1/2 inch 'handle' was formed on the end of a piece of tape by folding it over, and the handle end stuck onto the paddle of the stretching gauge. (This stretching gauge is just a spring balance with a piece of aluminium plate attached) The other end of the tape is stuck gently down onto the Mylar, leaving plenty of tape between the two. Pulling on the gauge stretches the Mylar, and when the correct tension is reached the tape between the paddle and Mylar is slapped down onto the stretching table. Ours Mylar was tensioned to  2.5kg. 


The strategy was to work out from the middle, adding more strips of tape as we went. The hope was that the middle bit would be relatively wrinkle free without the need for any sideways tension. And that's pretty well how it worked out. If you look at the next photo you can see a zone of smoothness flanked by major ripples. As tapes are added at both ends, the zone of smoothness widens. Magic! 


When both ends were fully taped down, we decided to go back over the tapes again from the centre to even up the tensions. This turned out to be a mistake and just introduced further tight little wrinkles in the centre of the Mylar. 

That was take one. Ripped the Mylar off and started again. Second time round we just worked out from the middle and got a lovely smooth result. After some inspection there were some slight ripples in the part of the Mylar we wanted to use, so two pieces of tape were used to ever-so-gently pull these ripples out. Had we measured it, I'm guessing it would have been no more than a few tens of grammes of pull.  

Taking the outer stator now and using a sharp knife we lifted the backing tapes from the double sided tape spacers to expose the adhesive. A thin bead of polyurethane glue was run around the edge of the panel on the PVC spacer. Lifting the panel carefully we placed it down on the stretched Mylar film. 

The DS tapes were pushed firmly onto the film to make sure they stuck down, then we pressed the glued PVC spacers down firmly onto the Mylar. There was excess glue in places and it spread onto the membrane in the active part of the speaker. Next time I'd be a bit more abstemious with the stuff. Hopefully it won't cause problems.

Books were piled on top of the stator to weigh it down while the glue dried. This polyurethane glue takes 6 hours or so to grip the Mylar strongly enough to be able to cut the panel free. Full strength in something like 24 hours.

The whole exercise took about 1 1/2 hours, included the whole learning curve. I'm guessing that we could do the next panel in under an hour. I had help, but strictly speaking it could be done single handed.

The smooth portion of the Mylar was about 450 mm wide, so that's about the widest panel that can be made this way. The Mylar on the roll is about 650 mm wide.

Gap Filling

posted Jul 24, 2010, 3:35 AM by Chris Fox   [ updated Jul 24, 2010, 4:01 AM ]

Having sighted the photos of the stators that appear below, Rob Mackinlay kindly pointed out the gaps between the PVC strips that make up the insulating spacers. Pasting his comment here:

Just a quick comment on the spacer construction. You've mitred the corners of the air gap spacers to make a nice join. Make sure you fill the gaps (assuming there are some) between the spacers or tape over them as there is a potential for the audio signal to arc from the stator to the EHT rail through them or for the EHT to leak to the stator. This assumes you are using a full circle rail. If you are using 2 discreet rails that do not cross the corners, ignore this!!

To fill these gaps I used sodium bicarbonate powder with cyanoacrylate glue. This makes a strong filler that is presumably a good insulator. I'll find out later if it's not!
A reference for this use of sodium bicarb can be found here

Process was to spread a little sodium bicarb powder onto the join to be filled using a handy tool made from a cut down icypole stick. (I think that's Australian for popsicle stick)

level it with some gentle pressure from a fingertip,


then add Loctite 438 glue.

 After this the joint was filed smooth with a small fine file

When cleaned up the joints looked tight and sealed. By the way, if the photos look a bit blurred, that's because they're taken with my trusty mobile phone. Auto-focus is never all that accurate on macro.

Building stators

posted Jul 19, 2010, 6:21 AM by Chris Fox   [ updated Jul 19, 2010, 7:07 AM ]

Collected the prototype stator from the powder coater. Apart from a couple of perf holes where they'd poked wires through to string the stator up they'd covered it nicely. They even remembered to tape up the spade lugs so they didn't get painted too. 

Next task was to  flatten out the stators a bit. They'd been rolled to a smaller radius of curvature than intended, partly because it's hard not to, and partly because it's always easier to handle a more curved object in transit. A nearly flat bit of 1mm thick perf steel's pretty floppy. Flattening them took quite a bit of massaging and testing against the stretching table curvature. Using fingers and thumbs it's not too hard to chase down small deviations from cylindrical, but it does take time and patience.


Putting the insulating spacers on comes next. The spacers are soft UPVC, so they cut quite neatly. After trying a chisel and a sharp box cutter, I ended up using a sharp pair of wire cutters. These had a real knife edge and cut the plastic cleanly and with little deformation.

A thin bead of Loctite 438 was run down the middle of each precut insulator. To spread the glue a bit I rubbed the plastic section back and forth against the perf steel. Ordinary old clothes pegs worked fine as mini-clamps. The Loctite 438 is a toughened black water-thin glue. It sets in seconds and seems really strong. Being black it makes no mess and leaves no white fumes across the work, which is a problem with the standard stuff.


Curvature means the top and bottom spacers needed a bit of rounding off to remove high spots at the very corners.

Two mm thick double sided tape is used as the horizontal supporting insulators. This is the Biolink tape. Odd stuff, it has a squishy clear base. Not sure of its composition, but it's nearly a jelly. Pulling on the tape stretches it out like a jelly snake lolly. I have no idea what the acoustic quality of this really soft, flexible material will be, but I'm betting it'll be quite damping of resonances. Could be a good thing.

Circumferential insulator spacing was varied to broaden any bass resonance. An pseudo-random pattern was chosen with 12 sections over 1000 mm. This makes each section just 85 mm wide, a D/S ratio of just 40 or so. The MLs have an even lower D/S ratio of about 30 by my estimation. This is way lower than usual (80 - 100) but is probably necessary with the more tightly curved stators. The sections nearest the bass unit was made quite narrow to help it resist being blown into the stators by the woofer. ML do this too. This coupling can be a problem, so a thin section may be able to resist.

A 'cartoon' of the tape positions was drawn onto a large sheet of heavy paper using a 0.8 mm permanent marker. This was a good guide to follow for placing the first set of cross spacers. The second was placed by lining the two pieces of perf steel up and clamping them together. The tape for the second sheet was then just put directly on top of the tape in the first sheet. The two ended up agreeing nicely. The blue plastic cover on the second adhesive side is left in place for now until the two sides are assembled.

Clip to hold panels together

posted Jul 18, 2010, 5:36 AM by Chris Fox

From ER Audio.

30M at $2.00pm       $60.00 = GST     $66.00
Courier   (inc GST)                            $20.00
TOTAL:                                            $86.00

Cutting Insulating Spacers

posted Jul 18, 2010, 1:45 AM by Chris Fox

The insulating spacers are cut from plastic angle in extruded UPVC which is  19 x 19 x 2 mm in size. This has to be ripped down to an L shape 8 mm wide on the long side and about 3 mm on the short side. The short side of the angle will wrap around the edge and side of the perf steel to help prevent arcing between stray bits of the diaphragm and a stator.

Set up the (new) table saw for close-up work. To make it all work I had to remove the blade guard, something I wasn't keen to do. Unfortunately to get the rip fence close to the blade it wasn't possible to leave the guard in place. Plastic is best cut with a blade just skimming the work so you need a saw that has adjustable height.

To cut something as floppy as the thin angle is quite a challenge. The first 3.6 m length was about 50% usable, the second about 70%, while the third was all fine. Keeping the angle straight and level while ripping was the problem. The best solution in the end was a thin piece of wood holding the angle against the rip fence, and optionally another piece pushing the angle back against the fence. Takes quite a bit of finger fiddling to get everything lined up and to feed the angle smoothly through the blade.



posted Jul 14, 2010, 4:35 AM by Chris Fox

Parts ordered from ER Audio arrived yesterday. These include Mylar, conductive coating, wire, copper tape, and samples of polyurethane glue and plastic extrusions for angle and clip to hold the panels together.

Parts order

posted Jul 12, 2010, 1:40 AM by Chris Fox

Ordered parts from ER Audio. 

15 mtrs 3.5um Mylar C Film     $87.00
5 mtrs Red EHT cable                24.00 
5 mtrs Beige EHT cable              24.00
25 ml conductive coating            46.00
copper foil tape                          29.30
diaphragm adhesive                      5.00
Postage                                      18.00
Sub Total                                 233.30
GST                                          23.35
Total due $256.65

1-10 of 16