Quad ESL-63 Repair (Complete)

My father and I have just finished the repair and refurbishment of a pair of Quad ESL-63 electrostatic loudspeakers.

(And no, Quad is NOT quadraphonic, it’s the name of the company that makes them.)

DSCN0429.JPG

A bit of Background

The ESL-63 first went into production in about 1982 or 1983, and ceased some time in the 1990’s. They are still regarded as one the best loudspeakers ever manufactured.

There is a vast amount of information on the net about the ESL-57 and its successor the ESL-63.

Why the name ESL-63? ESL just stands for Electro Static Loudspeaker. The 63 is because that is when the design work began. Peter Walker, the chief designer, was a genius who spent the best part of the next 15 years perfecting this design before it eventually entered production.

An electrostatic speaker works by moving a very thin, light, conductive diaphragm between two highly electrically charged plates. This presents a few obvious problems for manufacture:

- how do you generate the voltage (typically 5000V)?

- how do you couple the audio signal from your amplifier onto those charged plates?

- how do you get the sound out?

- how do you keep the highly charged plates from attracting dust and grot?

These problems were all solved with the first Quad ESL speaker back in about 1957.

The next generation, though, uses delay lines and other clever things to make the sound appear to be coming from a point source behind the speaker.

To quote from the Quad web site:

The ESLs work in a very different way to conventional loudspeakers. For those who don’t know, here’s a brief explanation of the principle. Whilst most speakers work with a cone type drive unit, the ESLs use a very thin membrane, one-tenth the thickness of a human hair. This MYLAR diaphragm has a special conductive coating and is stretched between two electrode plates. Relative to earth, both electrode plates have a high positive charge (+5000v), but there is a slight difference between the charge on each electrode. It is this difference that causes the negatively charged membrane to be more attracted to one than another, thus causing the movement of the diaphragm.

The diaphragm is mounted on an open frame and is thus free of any coloration from a cabinet. Because the diaphragm acts as a full range driver, there is no need for a crossover. There are no discontinuities in the frequency response curve, so you get the full music spectrum, as it was recorded.

The ESLs do much more than reproduce the full range – and this is where all other electrostatic speakers stop. Acousticians have recognised for a long time that an ideal loudspeaker should be a point source from which sounds waves ripple like a pebble in a lake. Ordinarily, an electrostatic would fall at this hurdle, but not the QUAD ESLs.

By using a series of concentric anodes, rather than just two plates, the Quads are able to produce a spherical sound pressure pattern. A series of electrode rings are fed with delay lines, so each ring responds to the change in current a split second after the previous ring, creating movement in the diaphragm identical to the ‘ripple in the pond’.

Failures, rebuilding and spare parts

Unfortunately, there are a few manufacturing quirks that come to bite an electrostatic speaker that’s 20 years old.

The major cause of failure is the glue used on the frame arrangement used to mount the stator plates and diaphragm. The high voltages and ionisation of air attacks the glue and over the years it becomes brittle and fails.

When the glue fails, the stator plates can move, and if they move too close to the conductive diaphragm (or to each other) it leads to arcs and sparks. They are not known to catch fire, but it doesn’t sound nice and it looks bad in a darkened room!

In Dad’s case we had one speaker that failed and needed to be repaired. Within days of that one being rebuilt, the other one went the same way.

Fortunately, Rob McKinlay at ER Audio in Western Australia has a wealth of repair information available, and sells kits that let anybody with a decent workshop do a detailed rebuild. When there is a major screw-up, spare parts can still be bought from Quad Germany or (in some case) the Quad factory in England.

Refurbishing

Here are some pictures and a bit of a commentary about the rebuild of one of the speakers. In this case we had a bass panel that was sparking (and it had to be the one at the bottom, requiring a complete dis-assembly, strip down and rebuild.

Disassembly

Taking the unit apart (at first) is fairly straightworward when you know where to start. The instructions from ER Audio were very helpful.

After removing the decorative piece of timber at the top, the black cloth is rolled down to the bottom out of the way. Then the front and rear grilles are removed, showing the transparent dust covers in front of the electro-static panels:

DSCN0439.JPG

Next, the dust covers are removed. These have a neat clip-in arrangement to the aluminium structure, so they just lever off. They are like a large sheet of glad-wrap on a frame, so they need to be handled carefully.

This leaves the electrostatic panels exposed:

DSCN0426.JPG

On this view you can see the concentric anodes of the treble/mid range panels. These are driven by delay lines, which makes the sound appear to come from a point source about 3 feet behind the unit. The green part is a (coated) copper layer over a printed-circuit-board-like substrate materal. This is the part that runs at 5000 volts. The holes let the sound out.

DSCN0428.JPG

Those big circuit-board like panels (the stators) are glued onto the plastic frame, and when the glue fails, the stators can flap around and move closer to each other. This is very common at each end.

Next step though, was to take apart the frame holding the panels in place and start to remove each panel, one at a time, working down from the top. In our case we had to repair the panel at the bottom, and inspect all the others for damage on the way. Each is held to the frame assembly with 4 screws – one in each corner:

DSCN0441.JPG

During the disassembly, the panel interconnect wiring needs to be desoldered and removed (and kept for later re-assembly):

DSCN0442.JPG

The bottom two panels have all of the signal wires that come up from the electronics underneath. These are soldered and heat-staked in place. Removing them is difficult – it is easier to cut them and solder up later:

DSCN0443.JPG

Disassembly and inspection of the diaphragms

Each panel needed to be taken apart to inspect the diaphragm for damage – burns, arc marks, holes, etc.

Taking the panel apart is easy – 3 screws and a few clips. Care needs to be taken to avoid damaging the Mylar film which is the diaphragm:

DSCN0444.JPG

Typical damage – a small cut or tear. The spotting seems to be normal.

DSCN0447.JPG

When a diaphragm is damaged, there is only one option. Remove it, clean up the frame, and make a new one.

Removal is easy – the Mylar is so thin it tears off easily. But it is glued onto the frame around the edges, and cleaning off the old glue and mylar is slow and laborious. Scraping with a not-too-sharp knife seems to work.

Making a new Diaphragm

Making a new diaphragm starts by marking the outline of the mounting frame on a sheet of glass, then laying a new sheet of Mylar over it. The Mylar comes on a roll, and is VERY thin (3.8 micron). As we found to our cost, cutting it with scissors doesn’t work! We should have read the instructions properly because then we did what it says: cutting with a hot soldering iron works a treat.

Once the piece of Mylar film is about the right size, masking tape is applied around the four sides to prevent tiny imperfections causing a tear during tensioning (this we worked out after 3 failed attempts, and a phone call to Rob McKinlay).

The Mylar is then tensioned to 2.4 kg using very strong fabric reinforced tape and a spring balance. Many pieces of tape are needed, the whole process of tensioning takes about an hour.

This shows just after we started:

DSCN0450.JPG

DSCN0451.JPG

After tensioning and priming, the cleaned up frame has a generous coating of Cyanoacrylate glue applied, and is pushed down onto the tensioned Mylar. Apply a LOT of weight (a big board, a carton of wine, some tins of paint and a few power tools). Leave for 24 hours.

Once the glue is cured the weights are all removed and the excess Mylar is cut away from all that black tape, again using a hot soldering iron. Then the edges are carefully sanded back to make it neat and trim.

Next, the special conductive coating is applied. This new modern stuff is clear, unlike the original grey. Making the holes for the screws is a nervous moment, but again the hot soldering iron works a treat. We don’t want anything going wrong at this stage, because the Mylar is under so much tension that the tiniest nick causes a big tear – almost like an explosion.

The newly coated diaphragm looks like this:

DSCN0449.JPG

Preventative Maintenance

Predictably enough, a vast amount of time had to be spent going over all of the panels with diaphragms that are OK, and re-glueing the stators. This is slow and painful, using a strong polyurethane glue and a very fine paint brush.

This shows part of the re-glued area on the right hand side:

DSCN0427.JPG

First testing

After all this fun and excitement – with a limited amount able to be done each weekend, it was time for preliminary re-assembly and testing.

The bottom two panels were put in, and the wiring soldered up. Instead of trying to re-do the heat staking, we used a modern silicon sealant/adhesive to tack the wires in place:

DSCN0452.JPG

A brief test at this stage showed no arcing, and everything works OK and sounds fantastic. Time to reassemble the remaining panels:

DSCN0453.JPG

Fabric Replacement

Along the way, the outer black fabric got damaged (it was a bit moth-eaten when we started), so next step was to remove the electronics at the bottom to get the fabric off.

Here is the inside of the electronics, with the delay line assemblies at the bottom:

DSCN0467.JPG

A quick burst by Mum on the sewing machine made up a new cover, and we could re-assemble the base, fit the dust covers, and test it again:

DSCN0468.JPG

All’s well, here is the finished result:

DSCN0469.JPG

DSCN0470.JPG

The complete rebuild and refurb took place over about 5 weekends (usually one 1/2 to 1 day per weekend).

Finished Result

Both ESL-63’s, refurbished (one with a bigger rebuild than the other), back at their home and in use again:

DSC00481.jpg

All ready now for another 20 years of service.

—————–

Late edit: 2 Jan 2009: There are other resources and information on the web.

One such is Mark Rehorst. He has a page in their dedicated to a diaphragm stretcher.

Powered by WordPress 2.8    Rendered in 21 queries and 0.561 seconds.    CleanBreeze Theme