A local client brought me a radio which was in the family for many years, a cute little “fridge radio” made by Westinghouse from 1945-1947. The same year Westinghouse released this new model of home refrigerator, they also released this Little Jewel radio in the same style as a companion. Many examples of this radio found their way into homes and kitchens over the years and are a great little collectible from the post-war era.
This little radio is a 6-tube AC/DC radio with a series-string tube line-up, 12SK7 12SA7 12SF7 12SJ7 35L6 35Z5. It has an integral tuned loop antenna and an RF Amplifier for great performance.
They sure packed it into here, too. The radio is held in place by two chassis bolts, one in each side and a rod which stabilizes it in place.
The dial scale has badly warped with age, but the grille cloth is nicely intact.
This radio’s owner mentioned that it had been playing but recently cut out. No surprise why – the original parts under the chassis looked to be in pretty bad shape!
Melted, leaking and blown capacitors will cause all sorts of issues – no wonder it wasn’t running like it used to! This radio was pretty cramped underneath the chassis, but there was enough room to install both electrolytic filter capacitors and the cathode bypass capacitor on the output tube under the chassis instead of needing to re-stuff the can on top.
I had a reproduction dial scale made up to replace the warped original, and replaced some tubes which were testing weak.
Finally, it was time to test it out and tweak the alignment trimmers, then back together! It sounds a lot better than I’d expect with such a small radio – the case makes an interesting vented enclosure. On 880 KIXI, you can really hear the low notes, and the vent on the top seems to serve as a port for the speaker.
After being refurbished electrically, this radio is going to perform beautifully for many years. It’s got a rich tone and great audio clarity even more than you might think from such a small speaker.
I recently got to work on something a little bit different – a tube hi-fi stereo console! This particular one, a 1962 Zenith MK2670, was a very high-end unit back in its day. It’s a dual-chassis unit with 19 tubes total, using the 12K25 stereophonic FM tuner and 7K31 stereo amplifier chassis with EL84 outputs. All housed in a beautiful mid-century modern cabinet with built-in stereo speakers featuring 12″ woofers and mid- and high-end compression drivers and horns.
With 19 total tubes, this model features hi-fi integrated speakers with high-efficiency 12″ woofers and mid and tweeter horns, push-pull EL84 output tubes for each channel, and built-in FM Stereo Multiplex decoding for true stereo hi-fi reproduction.
This hi-fi’s owner reported it was working well for several years but started to go downhill shortly before she got in touch to have it fixed. It was taking longer and longer to warm up and sounding more and more distorted, no longer delivering the rich warm sound of a classic hi-fi console. It was good she got in touch – waiting any longer could have led to catastrophic consequences such as component failure or even a fire. I visited her home to test the tubes and pull the chassis, then it was back to my shop for repairs. And what a job it was!
This unit had been serviced a couple of times in its life – there were some ’70s era film capacitors installed, and some of the output tubes had been replaced. Most of the tubes were original Zenith fittings and tested strong, though, so very little needed to be replaced. Since it came into the shop in working condition, too, it made the diagnostic process much easier!
The amplifier circuit is a bit different than most I’ve seen. The negative phase of the output transformers was connected to chassis, and there were two positive phased taps each connected to half of the speakers. It’s an odd arrangement, certainly, which would have let Zenith use woofers of different impedance then the mid/tweeter network without an expensive and complicated impedance matching network. Underneath, though, it’s pretty easy to work on:
Right away there’s some visible damage. This molded ceramic capacitor blew a piece of the ceramic clean off from overheating. It’s unlikely it was doing much of its job at this point. Despite a nice ceramic body and epoxy sealed ends, it’s still an acid-paper/foil capacitor inside subject to failure, and fail it did.
Many of these early disc capacitors were failing as well.
I tested and replaced resistors as necessary, and capacitors. Here’s an in-progress shot showing partial replacement complete. Even the resistors I did remove were very close to spec – Zenith clearly used high quality resistors in this hi-fi. All were shiny with intact bodies and paint, and none had the woody, chalky appearance of a tired carbon resistor.
With the amplifier chassis sorted, it was time to move onto the tuner. It’s nicely shielded on the bottom.
And there’s a lot going on inside.
This radio was build very, very densely and in layers. It’s also a modern hi-fi instrument sensitive to wiring changes, so I had to work without disturbing most of the arrangements. Fortunately, Zenith specified high quality carbon film resistors and they were overwhelmingly within tolerance, so it was just a straightforward capacitor replacement.
Even film capacitors are much different these days. The ’70s 0.47uF 400V capacitor, top, was replaced with a 0.47 uF 630V capacitor about 1/4 the size. Working with a needle-nose pliers in each hand, I was able to thread capacitors into the proper location under component networks without disturbing the top layer.
Now it was time for a power-up test. This particular radio uses an odd multi-tapped output transformer arrangement with drivers hooked up to both taps in the cabinet, so I ended up hooking up 4 distinct speakers for testing.
It fired right up and sounded great! There were a few minor issues to resolve, though. For one, there was a bad volume-invariant hum on the AM band only. This was due to a small short which was corrected. The volume controls were behaving pretty erratically, though, which took some investigation.
The unexpected behavior turned out to be due to a control scheme I hadn’t run into before. The “balance” control isn’t an actual fader; instead, the volume controls for the left and right channel are ganged together with a friction clutch. Turning the outer ring turns both together – but turning the inner ring adjusts the friction clutch allowing one to be turned independently to achieve the left-right effect. At the new set point, then, the outer ring will turn both volume controls together to adjust the loudness equally after the fade is applied. It wasn’t a popular control scheme, being replaced after a couple of years, because many consumers found it to be annoying and counter-intuitive.
After understanding how it worked, however, it turns out that it wasn’t gummed up, it was actually working properly. So, on to the next steps! Adjusting the bias on the channels:
RF and IF Alignment:
Lots of parts came out of this one!
This radio is going to continue to serve faithfully for many years, pumping out a warm and rich hi-fi sound and be a beautiful family heirloom to pass on. They just don’t build them like they used to!
I recently had the chance to work on this beautiful Motorola 99-FM21 from 1948, currently owned by the granddaughter of its original owner. It’s been in the family since its purchase new in 1948 and was lovingly cared for the entire time, even after it was no longer in active operation. It really showed, too.
This is a post-WW2 radio when very few radios contained any shortwave bands. This Motorola has the AM Broadcast Band, and the newly developed FM Broadcast Band from 88-108 MHz. That’s right – you can pick up modern radio stations with this tuner. It has a pair of 6V6 tubes for the output (although its bass response is limited due to an undersized output transformer) and a treble roll-off tone control.
The radio looks like it had been serviced one or two times in the past. Most parts were original, but two capacitors looked to have been replaced in the ’50s.
Here’s that small output transformer. To give you an idea of size, it’s a little bigger than two stacked boxes of matches, like a restaurant might give out. A similar hi-fi amp, also driven by a similar pair of 6V6 output tubes, uses a transformer about the size my fist in the same circuit position. The amount of iron in a transformer is directly correlated with its frequency response, so Motorola seems to have cut a corner by fitting this particular transformer. It sounds good, but doesn’t have quite the bass response that’s possible.
This radio’s owner requested the ’50s capacitors be replaced with period units, and the can capacitors restuffed and the shells retained. Not many elect for that additional service as it’s quite time-consuming to perform, but in this case it’s a great choice with the radio being in the same family for so many years. I started with labeled clip-leads in place of each replaced unit to keep the circuit straight. With the schematic to double-check, this ensures error-free service.
Restuffing capacitors involved heating them to melt off the wax, then pulling on the leads until the body of the capacitor was removed. I’d then ream the cylinder, add a new capacitor, seal with polymer clay and dip in wax, then replace in the radio. I found two period-correct capacitors to replace the two ’50s units, as well.
Capacitor re-installation was pretty straightforward:
I also used similar sized electrolytic replacement capacitors which look similar to how a period repair would have been performed. It’s useful to have a bag of donor parts around to provide shells for restuffing.
At this point, I hooked up my phone as a test source and it played loud and clear through the phono input. AM reception was okay, and FM was quite poor, however. It was time to do some further testing.
This radio is one of the first to use integrated couplets, blocks of several components in a single package. In this case, the component in question is a 47K 1/2W resistor paired with two 50pF capacitors. The resistor is reading 73.9K Ohms, which is well outside of its tolerance. This one was replaced with a similar visual appearance carbon composition resistor from IRC.
I replaced quite a few more resistors, and attempted an alignment. First, I tried using the triggered oscilloscope and sweep generator, which even required building some of my own test leads:
Unfortunately, my sweep generator’s FM options are geared towards more modern equipment which deals in whole-number modulation percentages; the Motorola wanted deviations around 2.2% and 1.4% which I couldn’t provide. So I went with the FM Alignment with AM Generator option, using my laboratory signal generator the Leader 3216. The procedure was a rather lengthy one, involving de-tuning the discriminator coil secondary to cause it to respond to AM signals, then tweaking other adjustments for peaks.
After adjusting the IF transformers in order and the RF trimmer for dial tracking, I re-peaked the discriminator primary using an insulated driver to eliminate the AM noise. The FM discriminator is supposed to be only sensitive to frequency deviation; a no-deviation signal (like an AM signal) should produce no sound at all. It was a very touchy alignment, taking about a half an hour of the tiniest of adjustments to null the signal. The curve had a very slope and it was incredibly difficult to peak – similar to the discriminator in the GE F-135, but even more precise because of the much higher frequency (10.7 MHz vs. 455 KHz).
Finally, however, it did peak up and I attached a 300 Ohm twin-lead dipole. The FM antenna has a known impedance, while the AM is a flat-board loop with a terminal for an optional external longwire antenna.
After this alignment I was able to pick up stations across the dial loud and clear on AM and FM, and the phono input performed perfectly. At this point I let it play for several hours hooked to a bench speaker as a burn-in test, then arranged with the owner to return the radio and test it out. Sadly, her home is on the declining edge of a valley which offers exceptionally poor radio reception and there’s nearly nothing to be heard. Where I received nearly all AM stations and many FM stations on the same equipment, there was only a few faint broadcast stations to be heard. We ended up attaching an amplified TV/FM antenna with a 75 ohm output to a 75:300 balun and attaching that to the radio, which vastly improved the FM band – although on some of the strongest local FM broadcasts at that point, now we were driving the front-end of the radio into distortion. So clearly, this wasn’t a great location for radio. However it sounds phenomenal on period music and is going to serve for many years to come.
This is a beautiful cabinet in original condition and with reconditioned electrics is going to be a great conversation piece, keepsake and music player for many years to come.
I recently had the privilege of working on this beautiful 1936 Zenith 5-S-29 tabletop radio. It’s a beautiful 5-tube table radio with a 6-inch speaker and the iconic black Zenith dial and lightning bolt Z pointer. This one is special, too, because it has a swept second hand to enable fine tuning.
This radio came to me locally from its owner who had purchased it on eBay a short time ago in “restored” condition. Unfortunately due to a memory issue with my camera the first set of photos was lost. The radio had definitely had some work, but this is definitely a case of “buyer beware” on eBay: it looked like the previous technician got bored half-way through and left most of the original capacitors intact. About 3 had been replaced with film capacitors, and the electrolytic capacitors had been replaced, but otherwise it was all original. The dial had some coloration wiped off the back, too, and a dried out rubber band instead of a proper dial belt. The zipties were there as well, although they appear to be serving their purpose so I left them alone.
The curved glass dial is held in place by a metal clip ring around the outside. Inside between the dial face itself and the glass was a ring of cork as a spacer.
In order to replace the dial belt it’s necessary to remove the dial ring.
The dials, side by side:
New dial belt slipped over both pulleys, and dial holder replaced:
I carefully re-glued the cork spacer (with its original gap in the ring) to the perimeter of the new dial face.
Dial pointers reinstalled on the dial:
I performed an RF and IF alignment to peak up the signal and really bring out the rich tone. Afterwards, 880 KIXI is coming in nearly spot-on. Prior to the alignment, it was coming in about 910. It tuned very well through all 3 bands – even bringing in 3 shortwave stations on Band C with the shop antenna! It’s that kind of reception that made Zenith famous with their “Long Distance” radios. Even an entry level set like this one was capable of excellent performance.
Reinstalled in the cabinet! Fully serviced and aligned, this radio will play beautifully in its owner’s home with a wonderful rich tone. Just in time for Christmas, too!
A change of pace from the Bose equalizers and hi-fi I’ve been working on a lot of lately, I had the pleasure of working on a 1936 GE Model A-52 antique radio.
This is a nicely designed and straightforward table radio with 5 tubes, AM and one Shortwave band. Back in the ’30s, RCA and GE shared chassis and designs quite closely and it’s no surprise this one uses all RCA metal tubes, 6A8 6K7 6Q7 6F6 5Z4.
This radio had been serviced in the past but was due for another go-around. Most of the capacitors had been replaced in the ’70s or ’80s, although there were a few that still needed to be replaced. I swapped the 4 capacitors which were definitely in need of replacement, but the other units tested fine and are recent enough I’m not too worried about them.
The radio power switch, though, had been bypassed. The radio’s owner reported the switch was sparking in the back. I tracked one down after several weeks and was able to get it installed and it functioned perfectly after that.
The radio’s alignment was already spot-on so no adjustments needed there. I re-assembled the radio and let it play for several hours of burn-in testing before sending it back to it’s home where it will continue to play beautifully for years to come.
Once in a rare while I’ll run across an old Bose Active Equalizer that I buy for myself, but they never last too long – I get a lot of requests to purchase a complete Bose 901 Series I equalizer to go with a set of speakers which long since lost the matching controller. I did just recently have one in stock and it went quickly; once they’re purchased I repair them on demand before sending them along.
This one’s all original as far as I can see. It was reported to have a dead channel when I purchased it, and the resistors had certainly drifted out of their tolerances with age. The case was in decent shape for being 40+ years old, too, although the light doesn’t quite catch it all very well.
Component replacement was pretty straightforward.
Some Bose 901 Series I and Series II equalizers used BC239C-labeled transistors, others used 2N5088s. They’re nearly identical – indeed, the rest of the circuit is identical – but they have a slightly different gain spec. Practically, this just translates into a slight difference in the volume control on your receiver – the generated curve is the same and both are identically factory specification compliant. When I need to replace transistors, I use all 2N5088s – but in this case, all transistors were good, so no replacement necessary! The neon bulb was flickering, though, so I replaced it with a brand new NE-2A and current limiting resistor. Then, a good solid control cleaning so all the switches moved freely and made good contact.
This one went to its permanent home next week where it should perform for many years to come! With precision metal foil resistors and new electrolytic and film capacitors, not to mention the very light duty cycle experienced by the equalizer (which draws only 1.5W total power consumption), mean it will be a long time until this needs service again.
I can repair your Bose 901 Series I, Series II, Series III or Series IV Active Equalizer for a low flat-rate with some optional upgrades. Most every one of the Series I equalizers needs to be reconditioned at this point. The majority of Series II do as well, and even the later series are coming up with defective capacitors and op-amps more regularly.
I’m happy to announce that starting soon, I’ll be offering professionally restuffed vintage capacitors for historically accurate repair and restoration of your antique radio! In my rebuild process, the old capacitor is carefully stripped of the old wax coating and the innards carefully removed. The body of the capacitor is lightly cleaned and a new high voltage axial lead film capacitor is installed in the cylinder. The ends are then filled with a medium density clay filler to provide stability, followed by a fresh coat of bee’s wax. The end product is nearly indistinguishable from one in original condition, perfect for performing a historically accurate repair of a valuable antique radio.
If you’re interested, leave a comment! Pricing is expected to be $3-5 per capacitor depending on type and value.
I was lucky enough to get to work on a beautiful 1954 Philips 778-2 radio console. It had been moved around and played for a while but eventually ended up giving only buzzing instead of audio output. I’m surprised it lasted as long as it did on original components, especially with a few long periods where it sat in storage. The owner asked me to bring it back to full performance, and now it sounds fantastic.
This is a rare and very high end example of antique radio definitely worth repairing. It has early hi-fi circuitry with a powerful amplifier stage and an efficient speaker in a ported cabinet, and this particular one has been retrofitted with an aftermarket turntable which is a bit less original, but probably higher quality.
Very little information is available for it that I’ve been able to find other than a schematic which was published in the Radio College of Canada service manual series; even the Radio Museum doesn’t seem to have an entry for it, which I’ll have to correct in the near future. It looks great, too, in a massive cabinet weighing at least a hundred pounds.
Interestingly enough, this radio is AM only whereas you’d find a similar radio from the U.S. with FM from the same year. As I was told, FM radio hadn’t yet made it into Canada due to licensing issues; that lagged a few years behind the United States. If anyone has more information on that topic, I’d love to hear about it. There’s also a name plate on the top indicating this particular cabinet was custom-built for a certain wealthy Toronto family. As you pull the front panel down to reveal the tuner and record player, the top is attached to a linkage and slides back similar to how a piano might open.
This chassis in this cabinet was very expensive when it was new – I wouldn’t be surprised if it was over a thousand dollars – 1954 dollars.
With 14 tubes, including four 6V6s in parallel push-pull, this is a great performer. The transformer (a 25-Cycle model) is absolutely massive, too, and the entire chassis itself is thick stamped steel. This is built like a tank and very serviceable.
It used a very interesting linkage to control the position of the dial band indicator, there’s a push-pull wire through a cable housing that extends along a cable guide and rotates a drum which says the name of the band you’re looking at.
This radio was serviced once about 10-15 years ago, and a few times back in the ’50s. It still had most of the paper, and primitive ceramic disc capacitors which were still mostly wax coated and are generally suspect at this point in time. The shop which did the repairs left a lot of old components in place, only replacing a few. This wasn’t the best practice, but it was more common about a decade ago than it is today; radios repaired in the ’90s and before are often coming back in for service now as well. Needing an appliance serviced once in 20 years, though, is a pretty good service interval.
I replace all parts that degrade with new, precision components, so they should last quite a long time.
Another shot. Several different ages of components in this shot – before I’ve touched anything.
The resistors were all specified as precision types, and by some miracle, only a handful of resistors were outside their marked tolerance. The most drifted were the cathode bias resistors on the 6V6s which I replaced with precision metal film resistors.
There are a lot of large-value capacitors in this radio, which are fed by a pair of 6AX5s wired in parallel. Plenty of B+ current to go around. Based on my experience with this radio, I’ve added 68 uF and 100 uF to my stock, but during this repair I had to create those high values by adding 10s, 22s and 47s in parallel.
Four filter capacitors and the output cathode bypass capacitor in total: 5 very large capacitors in 3 cans. I’m leaving them there both for aesthetic reasons, and so there’s no open holes on the rear of the chassis with several hundred volts exposed.
This is an in-progress shot while mounting everything up. The negative tabs of the old can make great mounting points since in this radio all filter negatives connect directly to chassis, and so does everything else.
I’ve mounted a couple of terminal strips to hold replacement filter capacitors, soldered to a chassis shield piece.
I use red alligator clips to identify where to clip the wires to the right length and solder while relocating the capacitors to the new terminal strips. This was good for helping figure out lead dress of all the wires at once, without losing my place. It does look a bit chaotic.
The coils and tubes all checked out during earlier testing, so with component replacement complete, I replaced the power cord with a new polarized model switching the hot side, and substituted a bench speaker on the output transformer. For this radio, I made a house call to remove the chassis assembly from the cabinet, since otherwise it would have been impossible to repair.
Finally, it was time for the first power-up. No smoke! I run the first power-up without the rectifier to guard in case there’s a short in the transformer itself (detected in just a few minutes), and the second power-up with the radio fully energized worked perfectly and started playing!
With component replacement settled, it was time to reinstall the radio onto the chassis assembly and do some listening tests to make sure everything was operating normally. A shop in Toronto added a line input/output across the volume control; I hooked this up to my phone playing Pandora to test that function. The sound was very, very good when hooked to my test speaker. Very warm and rich tone, and the separate treble and bass tone controls provide a good range of adjustment. The low-end isn’t quite up to modern standards, but as this amplifier predates “true” hi-fi designs by just a little bit, there’s a little weakness on the low end. This is almost entirely due to the output transformer’s size, and one or two components around the audio tubes. You just need a lot of iron to have good low-frequency bass response, and that gets heavy and expensive quickly.
An Edcor transformer for the same power rating flat 20~20K Hz weighs 4.5 lbs. and costs nearly $60 by itself. That’s what I’d spec if the original transformer was bad, but overall this is a great audio amplifier section.
When mounting the chassis back to the board with the dial, it was important to get everything to line up so the mechanical tuning capacitor and the tuning indicator are in alignment.
There’s a lengthy series of instructions involving injecting test signals and adjusting trimmers to maximize and minimize various effects. I’m using my oscilloscope, period signal generator, and a test adapter on the first alignment step, which involves injecting a signal into the IF amplifier grid. This kind of alignment doesn’t benefit from dragging out the larger but precision-accuracy digital frequency generator, so I’m using something a repair shop would have used at the time.
The 455 kHz IF signal is coupled through a 0.05 uF capacitor, which runs very close to the IF transformer itself and so I insulated it with a sheet of paper.
I’m using my scope to watch the RF input (yellow, top) and demodulated audio at the speaker (blue, bottom). My EICO 324 signal generator is pretty unstable when measured with such precision, but it’s similar to what was used at the time and so is entirely suitable for this kind of work.
In this case, the generator’s internal modulation on this setting looks to be nominally 400 Hz. That’s reasonable. The top is the AM RF envelope; both are synchronized and it’s easy to see how the shape of the two waves corresponds.
Zooming in to verify the frequency of an RF alignment point and the level before switching back to watch the audio. That 400 Hz tone is encoded on the 570 kHz AM carrier.
The alignment was positive! Some components inevitably drift with this much age so it’s tough to get spot-on perfect (not to mention, rarely being that good when new anyway). This one is pretty accurate, though, with the dial tracking within one division of the scale (20ish KHz generally). The offset is slightly varied across the dial. This is often caused by permanent changes to coils – coil forms may change size and the coil’s inductance; temperature-compensated capacitors may be subject to drift. That sort of thing. It’s normal for a radio to have a bit of variability in it these days, although when new they were a little bit tighter. Modern radios use self-calibrating phase locked loops in place of L-C tank circuits.
With a 15′ foot wire antenna strung up, it has good tone on the loud music and talk stations. There’s just a few problems with hum that are resisting efforts to take them out, though. Below a certain volume, there’s a loud 120Hz hum and also a bit of buzzing. In a low-interference environment it’s not too bad (nearly normal-sounding, even) but in a more electrically noisy environment it turned out to be a major problem! Back to the shop for more investigations. One important lesson is that my bench speaker is much less efficient than the speaker this radio came with. The hum was much louder when installed with the original speaker.
By looking at the schematic, this is really a nice-but-pretty-well-settled-technology radio receiver coupled to a very high-end mono amplifier. There are 16 tubes total; 2 are the rectifier and 1 the tuning eye, leaving 13 working tubes. The radio receiving tubes (6SG7, 6SA7, 6SK7 and 6AL5) are the RF amplifier, converter, IF amplifier and detector. That leaves a full 9 for the audio amplifier: Five voltage amplifiers and driver tubes driving a full set of four 6V6 tubes. On a hunch, I started pulling AF tubes. At the time this was to check for issues with the shielding, but one stopped me: the schematic calls out five 6AT6 tubes, but I ended up pulling four 6AT6 tubes and one single 6AV6 in the first AF amplifier position.
They’re fairly similar tubes, with a key difference: the 6AT6 has a gain of 70, while the 6AV6 has a gain of 100. In practice, the 6AV6 is 30x more sensitive than the tube the circuit was designed for – and as a result, it was picking up interference the circuit as designed wasn’t sensitive to. This would have had follow-on effects, too: with the first position introducing the interference, every tube afterwards would amplify the bad signal with the good. Luckily enough I happened to have a single 6AT6 in stock to replace the incorrect tube and this radio began playing perfectly hum free as soon as it warmed up. Problem solved!
I’d speculate tube was replaced with an incorrect substitute last service, but we’ll never really know how that happened.
Now time to deliver it for real – reinstallation back in the cabinet:
Fully serviced, this radio will continue to play faithfully for many years to come! It’ll be able to keep up with the times, too, since it’s been retrofit with a standard audio connector – it would be perfect with a Roku or other Internet radio hooked up permanently! This was a great project. I love working on these top-of-the-line sets, seeing how they’ve been treated in the past, and how they’re being used in their homes – very few of these exist anymore, and I’m lucky to have had the opportunity to work on this one.
I got this Bose 901 Series II equalizer in for repair recently and have finally had a chance to send it back out! It’s going back to its home on the East Coast to be the centerpiece of a Bose 901 Series II system that’s being brought back to life.
This one arrived in what looks to be all original condition.
It’s showing its age, though. The filter capacitors have started to leak visibly. The seals on the ends of the electrolytic capacitors can wear out with age and let the electrolyte out and moisture in, which will lead to failure.
The smaller filters, and output capacitors, look like they’re starting to seep a little bit as well.
I set about replacing the capacitors as is standard on all of my repairs; I tested the pulls to see some of their health.
I also replaced the film capacitors with brand new units:
It’s time to test the transistors; I’d hate to get further down the line. The same multi-tester has mini-grabbers which are perfect for gripping the transistor legs in circuit.
This tester automatically identifies ECB and characteristics; I’m checking to make sure they’re all showing about the same.
Next up, replacing the resistors with precision, low-noise metal film models.
I tested some of the replaced resistors, too. I’ve seen as bad as +35% in other Bose 901 equalizers. The Series I seem to have the worst drift, but this Series II had some significant drifting as well. Drifted resistors can throw off the carefully designed equalizer curves and keep the 901 system from reaching its potential.
As it turned out, the neon indicator lamp had also burned out – the metal-glass interface where the leads exit the envelope frequently gets loose and the neon leaks out over time. Unfortunately, a camera mishap deleted the “pre” photos of the replacement process. The neon lamp and leads can introduce interference; at the time these were invented, LEDs weren’t available but a low power, low heat light source was required. To get around this, Bose used aluminum shielding tape over the light. I removed this tape, slid the housing off, removed and replaced the NE-2A neon bulb, coiled up the leads, reattached the bulb in its housing and placed it back on the panel.
This did, as expected, produce interference – so it was time for replacing the shielding. I opted for copper shielding tape; this is commonly available and is period-correct (in fact, the power transformer is already shielded with copper shielding tape from the factory.) I shielded the entire length of the leads as well, so there’s a bit larger. This tape is both solderable and has a conductive acrylic adhesive for a variety of shielding options.
All in all, it’s looking pretty good – and burn-in testing with my set of 901 speakers proved it sounds great, too!
This equalizer will return home for many more years of faithful service! One more back to its full potential, but there’s plenty more out there to fix.
I generally write about the 901 Series I equalizers which come into my shop for repair. That’s natural – the Series I is the oldest, and as such, its components are most likely to have failed. Series II, III and IV are all getting up to that age, though, with a tapering failure rate as you head towards newer technology. The Bose 901 Series I and 901 Series II Active Equalizers are substantially similar: in fact they can even be interchangeably used with either Series I or II speaker sets, as both models use the same curves. The main changes are some modifications to the power supply, and an improved equalizer network which eliminates the 22mH inductors in favor of some different resistors. With a few years advances in electrical engineering, Bose was able to slightly reduce the parts count.
Some differences should be immediately obvious. One of the 100 uF capacitors has been replaced with a 500 uF capacitor; there are two fewer signal capacitors, no inductors, a prominent ceramic cap across the AC input for additional noise elimination, shielding around the transformer and neon bulb, and the notable lack of a power switch. The Bose 901 Series II equalizer is designed to be connected to the switched outlet of your hi-fi stereo receiver or amplifier and have its power controlled from the single switch.
Otherwise, it uses similar carbon composition resistors which drift with age and conditions. In this case, about a quarter of the resistors I tested were exceeding their marked tolerance. Many repair services for the Bose 901 Series I and Series II equalizer only focus on capacitor replacement (or even worse, only on electrolytic capacitor replacement) and leave resistors outside their tolerances alone, which can change the curve applied by the equalizer and result in sub-optimal performance.
Replacement is quite straightforward. De-solder the old components to de-populate the board, replace the components, re-solder. In this case, I moved through the process in a couple of stages when I found time to work on the equalizer. There are around 80 parts to replace , it is fairly time-consuming.
I have an inexpensive but accurate digital multi-testser which can evaluate and auto-detect diode, FET and transistor characteristics as well as measure capacitance, inductance, ESR, and loss of various components. I applied it to some of the removed capacitors, which were very clearly dead:
This one is supposed to measure 5 uF – which is the equivalent of 5,000,000 pF. Its actual value less than 1/20,000 of its expected value. That’s not good for that output channel!
Another output capacitor didn’t fare much better.
And same for a power supply capacitor, measuring lossy with high ESR and registering capacitance of only 7.9 uF when it should read 100 uF. Bad power filtering in a hi-fi system is a guarantee of bad sound! Fortunately, replacing the tired parts will fix everything.
This Bose 901 Series II equalizer returned to its home and hi-fi, and after this complete overhaul, will continue to serve faithfully producing beautiful sound for many years to come.