Restoring the AR-3a
As I wrote recently, I am lately the proud and lucky owner of a pair of vintage AR-3a loudspeakers. Acoustic Research manufactured many models of speakers from the 1950s on, but the 3a is considered their finest achievement, and one of the best loudspeakers ever made in America. But it was also very expensive (equal to about $3,200 in 2008 dollars), and that, coupled with some quirky technical issues (which I will describe), mean that the AR-3a is seldom encountered in the used market in especially good condition. Mine are virtually pristine.
Before I brought these speakers home last month, I knew practically nothing about them. But, a quick internet search took me to a wonderful website called The Classic Speaker Pages, and there I found the document that became my bible as I undertook the project of bringing my AR-3as back to life: “Restoring the AR-3a”. The Classic Speaker Pages also have a discussion forum devoted to Acoustic Research loudspeakers, and several members of the board have extensive experience with the AR-3a going back decades. They provided valuable assistance as I remedied the little technical issues that, understandably, affect a forty-year-old speaker.
What goes wrong with the AR-3a over time? A couple of things. The first is something that goes wrong with all old loudspeakers: woofer surrounds deteriorate. Most speakers use foam to connect the cones to the baskets, and over time that foam disintegrates and crumbles away. The early AR-3a used cloth surrounds, and while the cloth itself will stay perfect, the material used to keep that cloth acoustically sealed fails over time. Acoustic suspension speakers like the AR-3a rely on a practically airtight cabinet, and when the material that seals the cloth surrounds deteriorates, the cones move too freely. In the bass-rich AR-3a, too much movement in the woofer can destroy the voice coil. Replacements for AR-3a drivers are no longer made.
Meanwhile, inside the AR-3a cabinet, two issues almost always need addressing. First, the forty-year-old capacitors have drifted from their original values, and need replacing. That’s no big deal, since capacitors are inexpensive and easy to find. The second problem that plagues many Acoustic Research loudspeakers is corrosion in the potentiometers. The 3a has one potentiometer each for the mid-range driver and the tweeter. They are fairly simple contraptions, but invariably have become oxidized or corroded, and that means they often no longer make electrical contact.
So, bad capacitors and bad potentiometers mean that many old AR-3a loudspeakers no longer sound the way they should. And, when people play them loudly without making sure the cloth surrounds are still sealed, they risk serious damage. I was extremely lucky that my 3as had no damage whatsoever. Even the walnut cabinets are perfect, which is almost unheard of.
Following directions in the “Restoring the AR-3a” guide, I quite easily re-coated the cloth surrounds on the woofer with Permatex Hi-Tack Gasket Sealant that I readily found at the auto parts store. A very thin coat over the surrounds and dust caps–which are also cloth–sealed everything very well.
With the woofers out, I removed the fiberglass that fills every AR-3a cabinet. The crossover network inside seems very complicated at first glance, but after staring at the schematic it began to make more sense. I was fortunate that these cabinets had never been opened, and all the wiring matched the schematic perfectly.
Not wanting to get in over my head, I took the potentiometers out, cleaned them, and reinstalled them one at a time. One of the four was like new. The rest showed a good amount of oxidation. My Dremel tool took care of that. I re-soldered each potentiometer as I put it back in, and avoided needless confusion that way.
The original capacitors in the AR-3a look nothing like the capacitors made today. One large box inside the cabinet houses a 50- and 150 micro-farad capacitor. A separate small box contains the 6 micro-farad capacitor. I simply cut out the old capacitors and soldered in the new ones. It is amazing how much smaller modern capacitors are than the ones installed in older speakers.
Having cleaned the potentiometers and replaced the capacitors in each cabinet, I then replaced the grey putty that encircles the woofer opening, providing an acoustic seal. The original was still somewhat malleable, but I didn’t trust it to do the job, and new putty is cheap. I ordered it from an eBay seller that specialized in Acoustic Research restoration projects. I placed the heavy woofers back in each cabinet, tightened the screws, then re-installed the grilles. I was ready to test the speakers.
I carried them into the living room, hooked them up to my stereo, and played the music I always use on these sorts of occasions, Trevor Pinnock’s recording of the Brandenburg Concerto No. 2. It is complex music that isn’t too bottom-heavy, but has plenty of texture. I set the volume to a low level and pressed play. I put my ear to each driver on the right speaker, and was thrilled to hear sound from each. I had to tweak the potentiometers on the back of the cabinet: the AR-3a has plenty of bottom end; the mid-range driver is strong, too, and needs to be turned down a bit; and the tweeter is bright and clear, but needs to have its pot turned wide open. Satisfied, I put the grille back on and moved over to the left speaker. The woofer and tweeter seemed perfect, but the mid-range driver produced no sound. I was crushed.
My first thought, of course, was that the potentiometer was still a bit dirty, so I rotated it, hoping to find the sweet spot. No luck. So I carried the speaker to the back room and began the arduous process of opening it up again. The guts seemed just right. The right colored wires connected to the right components, just like the schematic. I inspected all my solder spots, and they seemed to be fine, too. But maybe one was a little loose, so I did it over, re-stuffed the cabinet, re-sealed the woofer, and carried it back out to the living room for another test. (It took me several days to find the time to do all this.) Still, the mid-range driver produced no sound. I began to worry.
Of all the things that can go wrong with the AR-3a, the worst, of course, is a bad driver. They don’t make accurate replacements for the originals, and when you can find authentic ones on eBay, they sell for hundreds of dollars. That seems like a lot of money for a one-and-a-half-inch dome speaker. If I did have to replace the driver it wouldn’t be the end of the world, since I paid nothing for the speakers, and even putting a few hundred dollars into them would still be well worth the expense. But I certainly hoped that it was almost anything besides the driver.
I didn’t have a multi-meter to test the driver, but the engineer at work loaned me a small tone generator used to test speakers. You press a button corresponding to the frequency you want to generate, hook the leads to the speaker and see if makes a sound. I was dreading the result, but to my great relief, the mid-range driver emitted a 15kHz tone, which told me the driver was fine, and the problem was somewhere else. I went to Home Depot the next day and bought a small multi-meter, then began the process of testing each component inside for connectivity.
The potentiometer was still the most likely source of the problem, but the meter showed that it functioned perfectly. The wire from the potentiometer to the speaker terminals on the front of the cabinet were also fine. So the problem lay somewhere before the pot. My multi-meter doesn’t work on capacitors, but I doubted that a brand new cap would be a dud. I was stuck.
I turned to the AR Forum on The Classic Speaker Pages, and posted a message describing my malady. A short time later somebody posted a list of potential problems, still convinced that it was a wiring error. One thing mentioned in passing, though, caught my attention, because it related to something I did inside the cabinet. When I removed the potentiometer for the tweeter, I un-soldered all the wires leading to it, then re-soldered them. But when I removed the pot for the mid-range, I cut the last quarter inch of wire, re-stripped, and soldered the pot back in. A long coil of seemingly bare copper wire leads to the mid-range pot. (You can see it in the center of the picture here.) I just soldered that where it was supposed to go. It never occurred to me that that copper wire had a thin but tough lacquer coating. Having cut the last bit of that copper wire, I had cut out the portion with the coating removed and re-soldered a coated portion. That was the cause of my trouble. I corrected my mistake, closed the cabinet, hooked up the speaker, and repeated my listening test with felicitous results.
The AR-3a sounds amazing. The bass is deep and solid, the mids are stunningly lifelike, and the treble sparkles. Listening to Bringing it All Back Home, I was stunned by “Mr. Tambourine Man”. The acoustic guitar seemed completely natural, so that if I closed my eyes, I couldn’t tell that there wasn’t some living person playing an acoustic guitar in my living room. Meanwhile, new details are being revealed in songs I have known for a long time.
I am indescribably lucky. Huzzah, AR-3a!
Filed under: Music, Technology on March 5th, 2010 | No Comments »