Acoustic Research AR48S Restoration Part 1

We’ve all done it. Up late, browsing Facebook marketplace and you see something you just have to have. So much so, you take leave of your senses and pay well over the market value because the seller is a friendly chap, and “you can’t take it with you”. In this case, it was a pair of AR48S loudspeakers. The pictures showed a pair of speakers that had seen some action, clearly, but nothing stood out as being impossible to fix. What turned up was a pair of speakers that looked like they’d been ‘rode hard and put away wet’ as the saying goes. The feeling of “what the hell did I do” grows, but you’ve spent the money now and there’s no turning back, so you might as well dig in. That’s my excuse, anyway.

The AR48S is a three-way large bookshelf loudspeaker with a 10 inch acoustic suspension woofer, 4 inch midrange and 3/4 inch liquid-cooled fabric dome tweeter. It is a sealed box cabinet with a combined series parallel crossover, flimsy spring-clip speaker terminals and a brown cloth grille. It was part of a five-strong range that included the AR18S, 28S, 38S, 48S and 58S. Here’s the manual for those interested.

These were in a bad, bad way. They’d been stored in a damp environment – a conservatory I later discovered. Like most speakers of the time the cabinets are made entirely from chipboard. Chipboard does not respond well to moisture of any kind. There was swelling, cracked joints and pealing vinyl, and several areas where the chipboard had failed entirely. Great.

I knew the speakers needed re-foaming. It’s rare to find a pair of ARs that haven’t had the foam surrounds replaced yet, but they’re still out there. A seller by the name of ’Strong Hi-Fi’ “restores” ARs with incorrect rubber surrounds. My AR28LS spent some time in their hands before I purchased them from the previous owner, and the surrounds are badly adhered to the front of the cone (they should be on the back), with the cone improperly centred and old glue residue and bits of old foam left behind. I’ll sort them properly one day. Rubber surrounds on AR woofers will ruin the bass response. It’s also crucial to use the correct surrounds, and glue them to the back of the cone if they were originally, as cone height is important to maximise excursion. And lastly, sticking the new surround foams on the front of the cone might save you having to clean up the old adhesive, but it also affects the cone height and is just shoddy workmanship. Do it properly, or don’t sell them to unenlightened buyers for ridiculous sums of money.

More on re-foaming later. More worryingly there was mould on the cones and evidence of rust on the drivers. One of the midrange enclosures had fallen off and was floating around inside the cabinet, the other fell off during the restoration. The glue holding the terminal trays, crossovers and components, and some of the cabinet joints had failed. The cabinet sealing adhesive had also failed in many areas.

The grilles had tears in the cloth, and the glue holding the cloth to the frames had failed too. At least the badges were still in place. As it stood, the speakers were worth a tenth of what I paid, being generous. Better get on with the restoration.

Inside

I started by removing the internal damping material, bagging it separately for each speaker. I then glued the separated midrange enclosure back into the one cabinet using a PVA wood glue, following the original glue line to position it correctly. When the other fell off later in the restoration, I repeated the process.

Once they were dry I removed the remains of the original glue. It was interesting to note the inaccuracy in their positioning from the factory. Clearly AR didn’t consider the relation of the enclosure walls to the driver to be of great importance, as though both enclosures cover the drivers fully there is a difference of about 6 mm in their alignment on the front baffle. For the sake of originality I opted to follow the original glue line rather than positioning them to be concentric with the baffle cutout.

Now that the enclosures weren’t floating around in the cabinets, I removed the remaining glue from the terminal trays, cut the wires off the spring clips and removed the trays. I left the crossovers in place. I tested the capacitors with a capacitor tester and they appeared fine, and there is no evidence of physical degradation so I left them alone for now.

A bit more internal work was necessary. There was a lot of excess glue dotted around in the cabinets from the factory. I removed some of it to clean things up and scraped away any loose bits of adhesive from the joints. The cabinets were then sealed with Evo-Stik ‘Sticks Like Sh*t’ construction adhesive. This is designed as a one-stop adhesive for the construction trade and will bond to almost any material you can name.

It is a combination adhesive and sealant with incredible bond strength and forms a watertight, airtight seal. The cabinet panels were sealed, as were the midrange enclosures internally and externally. It was also used to shore up the crossovers and to glue in the terminal trays.

Stripping

Moving to the outside I began the long process of pealing off the original walnut vinyl. The damp storage conditions had kickstarted the job for me, and the vinyl came off a lot easier than it might otherwise have.

Stubborn areas were heated with an old hairdryer, and an old chisel was employed to lift and cut the vinyl and scrape it away.

Had the cabinets been well cared for in dry conditions, lifting the vinyl would have been almost impossible and would have wrecked the chipboard surface. Pealing the vinyl from these cabinets left a mostly clean surface with only a few nicks here and there, which shows how bad the cabinets really were.

Fortunately most of the serious chipboard damage was at the rear of the cabinets, mostly around the rebate into which the back panel is set. This is a problem, however. Any damage here can break the airtight seal of the cabinet which is absolutely crucial to the performance of an acoustic suspension design. There was so much chipboard damage at the rear, not just missing pieces but swelling and cracking too, that I decided to take a router with a flush trim bit and cut the rebated section off entirely. The damp smell that emanated from the dust as the router cut away the chipboard is something I wish I could convey to you.

I removed the black vinyl covering from the back panels, and made frames out of pine timber strips. I made the frame strips slightly wider than the originals to cover the gap around the back panel and the cabinet side rebate. The frame pieces were glued and pinned in place.

Once they’re filled, sanded and re-wrapped nobody will ever know any different, and the replacement timber was the only way to ensure a fully airtight cabinet and a lasting repair. It also gave me the opportunity to poor PVA into the gaps around the back of the cabinet, hopefully sealing the chipboard to prevent any further swelling or layer separation..

There were some edges where the chipboard had separated. The panels are mitred, so any separation here can also affect the performance of the cabinet. I filled these with PVA and clamped them up to rebind the joints.

The cabinets were then filled. I used a 2 part filler because of its strength and gap filling ability. This stuff drys harder than the chipboard and is extremely durable, so it’s ideal to reform square corners. It’s also completely stable so it won’t move or crack under the new vinyl wrap which would be most frustrating.

Quite a lot of filler was used. Every possible gap was over-filled, including at the front where the mitred joints of the cabinet meet, the edges, the crushed corners, the seams along pine framework and the pin holes, and the nicks on the sides and back surfaces.

I then hit the cabinets with an orbital sander with a 120-grit disc, before more filler, more sanding, and finally a fine surface filler to get the remaining imperfections, with yet more sanding to achieve a smooth finish. Lumps, bumps, cracks, nicks and holes in the surface will show through the vinyl, so it’s important to get this stage right.

The vinyl I used is Fablon Deep Walnut, which is about as close a match to the original American Dark Walnut colour as is currently available. I shopped around and found that Dunelm and Hobbycraft are the best places to buy genuine Fablon, at least in the UK. There are alternative vinyl products available from other manufacturers including DC Fix, but I haven’t tried them. Your mileage may vary.

The job is easier with two people, one to keep the vinyl pulled tight and another to rotate the cabinet and smooth the vinyl. Then it’s a matter of carefully cutting and wrapping the corners and frames for a seamless finish. My father, a skilled tradesman with a lot more patience than I will ever have, is better at this than I so I was only too happy to delegate the task. Cheers Dad.

Black ‘Chalkboard’ Fablon was used to cover the back and front panels as per the original cabinet. I managed to remove the original front stickers by heating them with a hairdryer and gently pealing them up, cleaning away the residue with WD40 electrical contact cleaner. They were then re-adhered with the same contact adhesive used in the re-foam.

Re-Foaming

Now to fix the drivers. Foam degradation is very common after several decades, even if the speakers haven’t seen much use. The job of the driver surround is to support the cone, keeping it centred in the magnet gap and also to keep its movements linear, and to prevent deformation in the cone surface as it moves. The type of surround you use is especially important in a sealed acoustic suspension design, as it has to be flexible enough to not resist the back pressure of the air behind the cone and to allow forward excursion when the air inside the sealed cabinet pushes the cone forward, yet stiff enough to prevent cone breakup.

Cone breakup is the deformation of the cone across its surface when it meets the resistance of air when it is moving at a given frequency. In theory the pressure of the air behind the cone in a sealed loudspeaker should be fairly consistent across its surface, but in practice the air pressure in the cabinet is not constant due to the obstacles in its way (the midrange enclosure and the back of the tweeter for example), and if the cabinet resonates with the vibration of the speaker its vibrational energy can affect the air pressure seen by the cone.

I purchased the foam surrounds initially from Simply Speakers in the US. Their custom foam formula seems to be the best available substitute for the original foam on the AR drivers. I also appreciated their guidance and informative demonstrations covering the re-foaming process, as surprisingly this is my first time re-foaming a speaker.

The kits included a rubber-based contact adhesive which I used. Their method recommends centring the cones by feel so that was the method I chose, as it’s easier than removing the dust cap and shimming the voicecoil. The woofer dust caps had to come off anyway as they had holes in them and were dented in a way that would affect performance not to mention aesthetics. I wanted to learn the centring method for future repairs that would hopefully have their caps intact, but I could also have shimmed the cones on these.

Given the choice I would not use Simply Speakers again. The re-foam kit for the AR midrange drivers didn’t fit. When I enquired about this I was told they were fitted from the front of the cone, but they didn’t fit that way either as they were too small to fit the driver frame, and stretching them as suggested deformed them significantly to the point where performance would have been affected. I received a refund from them for the midrange foam as they seemed unwilling or unable to supply a larger surround. Responses, though fast and generally helpful, were rather terse, and the flat rate shipping to the UK was generally overpriced.

After some recommendations from the folks over on the Audioshite Facebook group I ended up ordering replacement foams, and woofer dust caps, from Speaker Repair Shop in the Netherlands, which came with a far more reasonable shipping rate, a much more friendly response to enquiries and, most importantly of all, parts of the correct size. I should have used them to begin with and will order from them in future.

The old foam was first stripped away. It pealed easily from the frame, leaving behind some glue and small bits of foam that scraped away with a small, sharp craft knife and an old chisel. “You can use a new chisel, it’ll be an old one by the time you’ve finished with it.” – Richard Stilgoe.

The foam came away from the cone fairly easily too. It’s rear mounted in this case. Residue was cleaned away with the knife and careful application of 99.9% isopropyl alcohol and flux cleaner taking care not to get it all over the coated cone or in the voicecoil. I don’t think removing every trace of the adhesive is important or necessary, so long as most of it is removed and you don’t have any big lumps left behind. Make sure to allow any chemicals adequate time to dry so they don’t affect the application of new adhesive.

When the foam was pealed from the midrange drivers, a gasket came off with it. The gasket was disintegrating (more evidence of moisture) so had to be removed entirely from the frame. It is a 1/64 inch (approximately 0.4 mm) gasket and should be replaced to give the correct cone height. I used 1/64 inch automotive gasket paper to reproduced the 2 gaskets, with a 102 mm outer diameter and 93 mm internal diameter.

Working on the smaller midrange drivers was a lot more difficult as the gap between the frame and cone edge is much ampler, but you can carefully ease the cone forward to get behind it. The adhesive on the mid range was more stubborn so after watching a lot of American techs on Youtube I purchased a bottle of ‘Goo Gone’ which cleaned the stuff up right away.

The foam is then carefully fitted into place. It slips behind the cone and should be centred to allow the clip of the surround to glue to the cone edge, usually following the old glue line. You lift the cone up at the edges, apply a thin bead of adhesive all the way round, and then move around the cone with fingers on top and your thumb beneath to press the two together. There’s a lot of movement at first while the adhesive tacks off, and some residual squeeze which was mostly cleaned away but dries clear and doesn’t affect performance. Once the adhesive starts to tack, final positioning can be achieved to ensure that the surround is centred on the cone, fully seated and pressed firmly into place all the way round for a tight bond without any gaps.

After an hour or 2 to set, the foam is then glued to the outer frame. This is an easier process. Lift the surround, apply a bead of adhesive and start working around the edge with your fingers to cause the adhesive to spread while it tacks off. This is the crucial part where the cone must be checked repeatedly for free movement. I was lucky with the woofers at least as setting the surrounds to be even with the lip of the frame centred them perfectly.

That luck didn’t extend to the midrange drivers. When I got the speakers one of the mids was missing its foam. I can only conclude that a previous owner had tired to re-foam them but unsuccessfully, as on further inspection the cone is also misshapen. There were no traces of the old foam in either the driver itself or the speaker, so somebody had clearly been messing with these before I got them.

To test for free movement, place your hand on the cone with fingers surrounding the dust cap and lightly press down. The cone should move freely up and down without any rubbing which you will easily hear and feel. You should also press down and slightly to the side around the cone, getting a feel for the distance the cone must move laterally to cause a rub. Doing this you will be able to use muscle memory to centre the cone. When you’re done, and the surround has dried, you should be able to press down on the cone without any rubbing, especially when the driver is vertically oriented as it would be in the cabinet.

More rubbing and squeezing of the surround gets it firmly seated to the adhesive on the outer frame, before plenty of drying time to let the adhesive fully set. The process is the same for the midrange drivers, though more difficult due to their side and cramped frames with limited access from behind.

Next, the terminal trays. I hate spring clip terminals with a passion, and these were broken anyway. So having removed the springs the terminals were cut from the trays using a Dremel and a pair of sharp pincers. The terminals were glued in from the factory but there was no way of getting them out without risking further damage, so cutting them away from the top and then filing out a hole was the safest solution. I thought after the fact that a hole saw in the drill press would have been a better solution.

I filed a hole in the back of each tray in a rough rectangle, leaving enough room to clear a couple of new speaker terminals. I then made a plate from black 3 mm acrylic to perfectly fit the tray in width, rounded on both long edges to give a smoother transition between the plate and the surface of the tray. The plate was sized to seal the hole where the original terminals once were, clearing the AR logo that I wanted to retain. I installed a pair of gold-plated five-way binding posts that also accept spades and banana plugs.

Time to put everything together. One original driver screw snapped and several were stripped. I looked everywhere for black imperial pan head chipboard screws and eventually found the Blackjax range. These use 1 inch (25.4 mm) screws to mount the mid and tweeter and 1.1/4 inch (32 mm) screws to mount the woofers, both now 8 (4.2 mm) diameter. The thread isn’t as course as the originals but they give an extremely tight hold in chipboard and are a perfect fit. I’ve got tons left over which will provide a good excuse to fix more ARs in the future.

The original driver gaskets were damaged, and 2 were missing. I replaced with self-adhesive EPDM foam rubber strip stuck to the back of the drivers. This stuff is ideal for forming an airtight seal between the drivers and cabinets.

The tweeters aren’t perfect. The cones are dented and I suspect, due to some harshness at high volume, that the ferrofluid cooling them has dried up. I will probably get these re-coned at some point, or look for an original pair of AR tweeters in good condition.

I left the original crossover parts and wiring in place. The only modification I have made is the aforementioned changes to the terminal trays. I might revisit the crossovers and replace the old wiring at a later date, but likely for no sonic benefit so I’m leaving it original for now.

I re-used the original grille fixings, but lined them up properly. I thought about using magnets but as they weren’t broken I decided to keep them original. I’ll finish the grilles in the next instalment. Here is the finished product.

Sadly despite my best efforts, the misshapen midrange is no good. The voicecoil is partially melted so there is no saving this driver. The other works, sort of, but seems to be excessively loading the amplifier. One of the tweeters also sounds excessively harsh, and I’m sure the diaphragm dents are doing them no favours, though both do at least work of a fashion. For the time being, until I can find a set of replacement mids and tweeters, these are shelved and we end on a cliffhanger.

Some lessons were learned in this. Don’t buy things online without thorough enquiries first, asking for further pictures and details if required. Don’t pay over the odds for broken things, because these days it is highly likely you’ll end up buying someone else’s discarded project and not something untouched from the factory that needs its first round of TLC.

In the meantime Stay tuned for more to come, including the refurbishment of a classic Pioneer PL-12D turntable.