Capehart-Farnsworth 661-P Television (1948)

Charming, distinctive, and seldom seen, this 1948 Capehart-Farnsworth 661-P television is a favorite of mine. With a 10-inch picture tube and a tall, slim profile, it takes up less space than most vintage consoles, a bonus in our crowded home.

Finding the Capehart-Farnsworth 661-P

When I saw this Capehart-Farnsworth advertised 150 miles away, I made a beeline to the owner's house for a look. The asking price was $250, but after I pointed out some issues, including a damaged picture tube, they parted with it for a reasonable $120.

Although I had intended to remove the chassis for transport, everything seemed very secure when I checked it out. I removed the knobs, and then, with the seller's help, carefully laid it on quilts in the back of my SUV.

First Look

Here is the TV, immediately after I unloaded it at home.

The cabinet stands 42 inches high, 19 wide, and 22 deep. It's in decent cosmetic condition, with minor scuffs, many tiny dots of latex paint, and a couple of little scrapes in the grille cloth.

From the rear you can see the chassis above and a generous 12-inch speaker below.

The hinged metal cover has been swung open in this view. The power cord mounts on that cover with a safety interlock.

Here's a closer view of the speaker. Capehart's high-end radios were known for fine audio, so I'm curious to hear what this TV sounds like.

A peek at the chassis reveals why I wasn't willing to pay top dollar for this TV. The tubes are all missing, except the picture tube, whose base has been torn off. The CRT socket is lying in one of the empty rectifier tube sockets.

The small tubes are common types, so they won't be difficult to find, but replacing a couple of dozen will cost a bit. The picture tube is a bigger problem. Although it's possible to reattach a base if the wires are intact, that tube might be a dud for all I know, which could add another $100 to the parts bill . . . if I could find another good one.

Two notable features are seen in the previous photo. There are two power transformers and the one on the left has a rotary switch, likely to adjust for different line voltages. Mounted up and behind the left transformer is a small subchassis that houses horizontal sync circuitry.

Before bringing the TV indoors, I wiped the cabinet inside and out with mineral spirits. I'll clean it more later, but at least this will remove any mildew and spiderwebs.

Who Were Capehart and Farnsworth?

The Capehart-Farnsworth moniker links two notable figures in radio-television history, although they were not partners, and each of them eventually lost the company that bore his name.

Homer Capehart

Homer E. Capehart was a businessman and politician who grew wealthy manufacturing phonographs, radios, and jukeboxes, and served as a Republican Senator from Indiana from 1944-1962.

Starting out as a salesman, Homer Capehart founded the Automatic Phonograph Corporation in 1927, which became the Capehart Corporation in 1928, with headquarters in Fort Wayne, Indiana.

Capehart phonographs and radios from the late 1920s and early 1930s were aimed at the luxury market, housed in large, ornate cabinets and sporting lofty names like Amperion and Orchestrope. Nowadays, certain Capehart sets are eagerly sought by collectors, particularly for their audio components.

Here is a brochure for the 1932 Capehart 400 radio/phonograph in a Chippendale cabinet:

From early days, Homer understood the market potential of an improved record changer. Devised in 1928, the Capehart Turnover record changer could play both sides of a record and it could also play records of different sizes.

By 1932, Homer had been pushed out of his own company by investors. He promptly bought the rights to the Simplex Multi-Selector automatic record changer, which allowed patrons to choose a specific record rather than play whichever record came next in the stack.

In 1933, the Simplex changer was bought by Wurlitzer, which was then on the brink of bankruptcy. Homer Capehart also joined Wurlitzer, where, as vice president and sales manager, he achieved a dramatic turnaround in the company's fortunes. Before Capehart, Wurlitzer leased fewer than 300 jukeboxes, or roughly 1% of the US market. Three years later, under Capehart's aggressive leadership, the total jukebox market had tripled and Wurlitzer's share had skyrocketed from 1% to nearly 50%.

This video shows a 1936 Wurlitzer Model 412 with Simplex changer, playing a song.

While early Capehart products had been marketed to the elite, the jukebox, used in public places, was pitched as a democratizing influence. Speaking as a Wurlitzer executive, Homer Capehart declared that, "Poor people have the right to enjoy good music with their beer or sandwiches."

In 1938, six years after Homer's departure, the Capehart company was bought by Farnsworth for its brand recognition and manufacturing facilities.

By the 1940s, Homer had turned to politics. He was elected as a Republican Senator and served three terms, under Eisenhower and Kennedy, until his defeat in 1962. Capehart-named products, under different owners, were sold until 1965. Homer passed away in 1979.

Philo Farnsworth

Philo T. Farnsworth was a Utah farm boy who became interested in electronic television in his teens. Prior TVs were largely mechanical, relying on devices such as mirrors or spinning discs. In 1930, Farnsworth received a television patent based on his "image dissector."

Although Farnsworth fans sometimes call him the inventor of television, that's an exaggeration. Television, like the automobile, was not invented by a single person. Both were complex devices whose time had come, and both were the result of people developing similar ideas in parallel—and often in competition—over a period of decades.

Farnsworth was embroiled in patent litigation with RCA during the 1930s and he sought connections with the British company of John Logie Baird, a Scot who had created a successful mechanical TV system. He worked at Philco for a couple of years and his image dissector camera tube was one of two types adapted in Germany to televise the 1936 Olympics in Berlin. (The other type was the iconoscope, which later formed the basis of United States television broadcasting.)

By 1938, the Capehart company was fizzling and Farnsworth had left Philco. Having gained a $1 million licensing payoff from RCA for his 1927 patent, Farnsworth bought the Capehart company. He moved to Fort Wayne, home of the Capehart factory.

The new company was called the Farnsworth Television & Radio Corporation and it sold electronics, including my 661-P television, under the Capehart and Farnsworth names. Capehart supplied a factory and a bankable reputation, while Farnsworth supplied engineering drive and a desire to develop consumer products.

Both names were used in postwar advertising, with Farnsworth representing economy sets and Capehart the luxury line.

In this Life magazine ad from September, 1946, Farnsworth radios present a "blending of quality and economy" and they are priced from $25-$350. Capehart "instruments" are aimed at the well-heeled music lover and they are priced from $300 to a regal $1500—the cost of a new automobile.

The 1946 product line didn't include TV. The ad promised that television receivers would be "available soon in some cities, particularly where programs are being televised."

By 1947, television made its first appearance in a Farnsworth ad. Notice the little tabletop in the lower right corner:

Evidently, Farnsworth TVs weren't quite ready for market. No model number is given for that set, and while it resembles a model GV-260, the ad states, "AM radio also available." In fact, no Farnsworth tabletop of the 1940s included a radio, and given the already crowded cabinet, it would have taken a miracle to shoehorn a radio chassis in there, too.

The GV-260 did reach the market in 1947 and its streamlined cabinet makes it highly desirable to collectors. The Early Television Foundation owns a rare GV-260 prototype and their article provides interesting details about that model.

1948 brought new Farnsworth TVs, including my model 661-P. The 610-P and 651-P look exactly like the 661-P, except in a smaller tabletop cabinet. The mammoth 501-P was a TV/radio/phono in a breakfront cabinet, seen here in a 1948 ad:

(The 501-P looks imposing, but if I'd designed the cabinet, I would have put the TV screen in the middle, at a comfortable viewing distance, rather than on top, where seated viewers had to crane their necks to watch it.)

Farnsworth's company prospered from government contracts during World War II, but it wilted in the postwar years. In 1949, he was forced to sell the company to International Telephone & Telegraph. Philo spent his remaining years primarily doing research into other subjects.

Since the company resulted from acquisitions, you will find the names Farnsworth, Capehart, and Capehart-Farnsworth used somewhat interchangeably. My TV's serial number plate and service manual calls it "The Capehart," built by the Capehart division of the Farnsworth TV & Radio corporation. This 1949 IT & T ad describes the merger of IT & T and "Capehart-Farnsworth television," referring to "Capehart's advanced laboratory techniques" and reducing Philo to a mere hyphenate:

Farnsworth radios are much easier to find than televisions. My GT-051 and ET-064 Farnsworths were built in Fort Wayne in the late 1940s.

The company sold TVs in small numbers throughout the 1950s and it ceased production in 1965. These televisions were conventional for their time, reflecting technology commercialized by other companies, notably RCA. By the mid-1950s, the Farnsworth name had all but disappeared. IT & T marketed its TVs as Capeharts and fine print at the ends of ads identified the company as the Capehart-Farnsworth division of IT & T.

The Library of Congress has an interesting photo of women working the TV assembly line at Fort Wayne in the early 1950s. Then, as now, electronics were designed primarily by men and assembled primarily by women.

Although their names appeared on the labels, I doubt that Homer Capehart or Philo Farnsworth had any real involvement in the 1950s products. By then, Capehart was serving in the U.S. Senate and Farnsworth spent his time researching subjects like radar and nuclear fusion.

Philo spent his later years pretty quietly, one exception being a 1958 appearance on the TV show I've Got A Secret, where he successfully stumped the panel and declared that he "invented television at age 14." He passed away in 1971.

The old Farnsworth factory in Fort Wayne was demolished in 2010:

So, what was Farnsworth's true legacy? That will be debated for as long as people watch televisions. To some people, he personified the myth of the lonely inventor, sparked by inspiration and struggling against mighty forces to bring his ideas to life. Others see him as a stubborn hick, weak in business sense and delaying the adoption of television for years through his refusal to play nice with others. My personal view falls somewhere between those extremes.

You can learn much more about early television history at the Early Television Foundation website.

Description

The Farnsworth 661-P uses a ten-inch 10FP4 picture tube, equivalent to the 10BP4, which was the most popular late-1940s CRT. The 10BP4 was used in my 1946 RCA 630TS as well as my RCA 721TCS, Admiral 24C15, Hallicrafters T-67, RCA T-100, Westinghouse H-181, and many other sets by various companies.

A memorable feature of this TV is its tall, narrow console cabinet.

Most 10-inch TVs were either tabletops, such as my Hallicrafters T-67 and RCA 630TS, or larger "combos" that housed a TV, radio, and phonograph in one cabinet. Admiral made a few 10-inch TV-only consoles, such as my 24C15, and RCA made the notable 721TCS.

Here's a look at the restored chassis.

In the first photo, you can see the tuner at lower right, mounted at a slight angle from the main chassis. The third photo shows how the horizontal AFC (automatic frequency control) subchassis is suspended above the main chassis.

A two-page schematic is available from Sams, folder 95A-1. The factory service manual is much more detailed; click the following icon to download the 15-megabyte file. (Thanks to Chuck A for scanning and hosting this manual.)

Much of the 661-P's electronic design is similar to other TVs of the day. The horizontal AFC circuit is the RCA "Synchrolock" system, a carbon copy of which is used in my 630TS. The 661-P and 630TS both use a "split sound" audio system, and, consequently, both TVs use more tubes than later sets: 29 in the 661-P and 30 in the 630TS.

The power supply has two transformers and two low-voltage rectifier tubes. Each transformer supplies roughly half of the tubes in the TV. One transformer, labeled "Television" in the manual, is switchable for different line voltages.

The high-voltage section uses a voltage doubler with two 8016 HV rectifier tubes; the 8016 is equivalent to the more common 1B3GT.

Here are the types and functions of the Capehart's tubes.

The 661-P's deep, narrow chassis made for a compact cabinet, but it didn't simplify life for the serviceman. As the tube layout diagram shows, several of the 661-P's tubes are well hidden under the picture tube. Some are impossible to remove without pulling the chassis from the cabinet.

The inspection label and serial number plate are located on different parts of the cabinet, but I combined them in one photo, which identifies this as set 496799:

I would not take that number to suggest that Farnsworth—a small company—made nearly half a million of these TVs. Manufacturers used many different schemes for serial numbers, which don't necessarily refer to a single model or even start numbering at zero, for that matter.

Here is our first look beneath the unrestored chassis:

The three big black cylinders contain nine electrolytic capacitors for the power supply. At lower right is the high-voltage section and at upper right is the tuner.

The tuner is unusual. Unlike the conventional turret or wafer types, it uses ganged air variable capacitors, more often found in communications radios.

These capacitors tune continuously by their nature, but the tuner has mechanical detents to make it click into place for each station. A similar General Instruments tuner is found in my Scott 800/B radio/TV/phono console.

Here's a view of the tuner from the side, above the chassis. Although many tuners are fully shielded to reduce interference from external sources, this one is somewhat open to the breezes.

Electronic Restoration

I always work on the electronics before the cabinet and other cosmetics. If a TV can't be made to work, I don't want to invest a lot of money or time making it into a beautiful but useless "shelf queen."

Cleanup

The chassis was covered thickly with dust, but that was easily cleaned up and I saw no rust underneath.

Moving indoors, I took a close look at the picture tube base. Not encouraging! The 10BP4 CRT base uses five leads. Four of them are intact, although bent, but the fifth is broken off flush with the glass. Who knows whether this can be repaired?

In the previous photo, you can see this TV's unusual ion trap magnet, which is shaped as a wide, full collar rather than a curved magnet held by a narrow clamp. Removing the magnet tells us that the tube is a 10BP4, and now we can see that the broken lead (left arrow) connects to the CRT's second grid (right arrow). The 10BP4 data sheet tells us this is the lead for pin 10.

Repairing the Picture Tube Base

This looked grim, but I had read about reattaching loose CRT bases, and even repairing a broken lead by grinding away the glass and soldering a new wire to the stub.

Calling on the VideoKarma TV forum for advice, I got several suggestions, including tips on how to grind the glass with a tiny diamond burr or how to secure a new lead with silver-filled epoxy. For more details, I'll refer you to that discussion.

Before attempting the repair, I tested the CRT using the universal adapter on my Sencore CR70 tester. If the tube's a dud, why waste time trying to fix the base? The stub was too short to grip with a test lead, so I held the lead against it.

Woo-hoo! This CRT's emission is as strong as any I've seen on that tester. It would be a shame to discard it without chancing a repair, especially since the last CRT rebuilder in the United States closed up shop last year.

The next photo shows my first attempt at soldering a new lead to the stub. I had first scraped the stub shiny with a razor point and cleaned it with soldering flux to improve adhesion. (In hindsight, I should have used a thinner strand of wire, rather than one that approximated the gauge of the original leads; a thinner wire will bend under stress before it breaks the joint.)

So far, so good. In the following photo, I have attached guide wires to the other leads and begun to slide the base cap onto the tube end. At this stage, the new lead was electrically connected and I got the same strong results on my CRT tester.

Alas, the joint broke loose when I gently wiggled the cap into place. The surrounding glass makes a very effective heat sink, so perhaps this was a "cold" solder joint, sufficient for an electrical connection but lacking the adhesion needed for a lasting bond.

The experiment taught me two things: ordinary solder won't work here, and that heavy un-stranded wire is too stiff. Following the advice from VideoKarma folks, I ordered some silver-filled epoxy and found a roll of silver solder used in a previous project.

The conductive silver epoxy is designed expressly for electronic repair, such as fixing broken printed-circuit boards. It doesn't come cheap: these little tubes contain about 1/3 ounce of epoxy and they cost $31. Still, that's much less than it would cost me to buy a new, strong 10BP4 tube.

I removed the stiff guide wires from the original leads, re-cleaned the stub, and then carefully silver-soldered a double twisted strand to it. Without disturbing the wire, I applied silver-filled epoxy around the joint and let it cure overnight under low heat.

Before trying to attach the cap, I re-tested the tube to confirm that the repair was successful. As one forum member pointed out, the grid doesn't carry much current, so even a small electrical pathway should be sufficient.

I initially tried soldering new guide wires to the other leads, but this was tricky, since you need to make a butt joint rather than wrap the new wire around the lead, which would make it too thick to fit through the pin. Instead, I carefully straightened the old leads and adjusted their spacing to fit the cap precisely, holding the cap pins next to them as a guide. Then it was simple to slide the cap on.

In the next photo, I have installed the tube base over the leads. You can see the new stranded lead sticking out of pin 10.

To remove old solder from inside the pins, I simply heated each one and then rapped the cap on the workbench. After trying the fit, I partly removed the cap, squeezed some RTV "sensor safe" sealant inside, pressed it onto the CRT neck, and then let the RTV cure for a few hours.

The next photo shows the final repair, after the leads have been trimmed and the pins have been re-soldered.

Whew! I had been nervous about this repair, but it proved to be simple. It's still possible that the joint will fail under the stress of heating and cooling during normal use. That would be more worrisome if it were attached by nothing but a miniscule dab of solder. With the reinforcement of conductive epoxy, I'm hoping it will last a long time.

Gathering Parts

With CRT repair out of the way, I began collecting the parts needed for this project. In addition to the usual capacitor replacement, I needed to supply over two dozen tubes.

Digging deep into my tube stash, I came up with 21 of the 28 needed small tubes. The seven additional tubes in boxes were ordered by mail, as were new caps.

Now I could populate the chassis. The tube socket holes were probably dirty from sitting empty for years, so I wiped each tube's pins with DeOxit before insertion and then unplugged and plugged the tube several times. Now our little Capehart is looking more civilized!

The previous photo shows a slightly different view of the tuner than previous shots. It's mounted at an angle to avoid bumping the bell of the picture tube. Remember, this same chassis was used in a tabletop cabinet, so keeping it narrow made sense.

If you are ordering capacitors for one of these TVs, note that the service manual's parts list doesn't include every part in the chassis. It's more of a price list. While it lists every type of component used in the TV, it didn't list their quantities, as many service manuals do.

Thus, the parts list shows a Part Number 90, a 30/30/20-mfd multi-section electrolytic can. If you look at the schematic, however, you see that Part 90 is used at three different places on the chassis, so to replace those three cans, you need nine capacitors, not three.

Replacing Electrolytic Capacitors

Since capacitor replacement is a common procedure, I'll refer you to my recapping article for the basics. The next photo shows two of the Part 90 type electrolytics. To keep things straight, I named the three cans 90A, 90B, and 90C, marking each with that name on the schematic.

Here, I'm restuffing the original containers, installing the new caps on the old bases without unsoldering all of the leads. You can read more about that technique in my RCA CTC-7 article.

Capacitor 91 is another multi-section electrolytic in a can, comprising three 40-mfd capacitors rated at 25 volts. The schematic shows its positive leads connecting to the cathodes of the 6SN7GT vertical amplifier tube.

As often occurs with schematics, parts close together in the logical diagram may be far apart in the physical layout. The vertical amplifier tube is on the opposite end of the chassis from this can. Tracing Part 91's positive lead to that socket, I found an ugly sight. These leads were badly burned. One of them lost its insulation completely. Perhaps one or more of these capacitors suffered a short circuit.

I made a note to check the other components connected to that tube's cathodes: the vertical linearity potentiometer (Part 103) and a 1K resistor (Part 15).

The Part 91 can holds three 40-mfd capacitors wired in parallel, which equals a single 120-mfd cap. Using parts I happen to have on hand, I'll replace them with a 100-mfd cap and a 22-mfd cap wired in parallel (both are rated for 35 volts DC). The rightmost photo shows the restuffed can, wired into place and ready to go.

The leftmost photo shows the old capacitor innards pulled from the can. In a healthy capacitor, this material would be moist and fresh, completely filling the can, not a shrunken, dried out carcass. It's unlikely that the TV could have worked properly with that junk part in place.

You might wonder why Capehart used this can containing three capacitors, since they are paralleled to act as one cap. Economy is my guess. Perhaps they got a great deal on a carload of 40-40-40 cans, or the Capehart-Farnsworth factory used the same can in more than one radio or TV model, either as a single 120-mfd cap, an 80-40 can, or a 40-40-40 can. In any event, my replacement will work just like the original.

That concludes the quick 'n easy portion of our program. When I saw how the next two cans were hidden under the horizontal AFC subchassis, my first reaction was, "You've got to be kidding!"

Fortunately, the subchassis is connected by flexible cables. In the next photo, I have loosened its wing bolts, swung it aside, and then removed its support bracket from the side of the chassis. (You can also detach the subchassis completely by unplugging its cable from the main chassis.)

Sawing the cans takes less time than taping and masking the area.

Installing the replacements will be easier if I mount the chassis on the stand that I built to hold my DuMont RA-113 on the workbench. Sturdy bolts run through the chassis' own mounting holes, securing it to the stand.

Each of these cans contains three electrolytics. Part 89 has three 10-mfd/450v capacitors, while Part 93 has two 500-mfd/10v caps and one 40mfd/10v cap.

After installing the new caps, I took photos from underneath to show the can bases all wired up. The can for Part 89 was too small to hold all three new caps, so I wired one separately under the chassis. I also needed to swing aside the big cardboard-coated can capacitor and temporarily remove the white and black "doorknob" capacitors from the high-voltage doubler board.

That part of the chassis is so crowded that I couldn't have fit all of the new electrolytics directly underneath. Simply reaching the cans' terminals to solder the new connections was rather challenging.

Much later in the project, after hours of playing on the workbench, I was satisfied that the new capacitors worked reliably. I reattached their cans with two-part epoxy, temporarily holding them in place with a few licks of tape.

Replacing Small Capacitors

Replacing the small non-electrolytic capacitors was straightforward, with a couple of exceptions. If the underside of the main chassis seemed crowded, it couldn't hold a candle to the horizontal AFC subchassis.

In the first photo, I have flipped over the subchassis and removed its bottom cover. This small space contains twelve (!) paper capacitors. The second photo shows the subchassis after initial recapping, which opens up more elbow room.

Three hidden capacitors are easy to overlook. The first is above the chassis, concealed by a shield over the volume control. I removed the shield and set it aside for this photo:

The second hidden cap is in the same area, but underneath the chassis. We saw this spot earlier when restuffing an electrolytic can (first photo). The .05-mfd molded paper cap can't be seen until you remove the audio output transformer (second photo).

There's no need to unsolder anything from the audio transformer. Simply remove its mounting screws and swing it out of the way while replacing that capacitor.

Hidden in a different way is this paper capacitor on the high voltage board. Painted with black insulation, it's easy to miss unless you're carefully reading the schematic and checking things off. I happen to have a 1000-volt rated cap on hand that will be more than adequate to replace it.

Replacing the Ruined Width Control

I had earlier noticed signs of overheating, including burned wires and a couple of components whose leads had either burned or broken off. While inspecting for other damage, I noticed that the Width potentiometer had wires connected to only one terminal—the middle, or wiper, terminal. In the next photo, I'm pointing to the empty terminal where something obviously was connected before.

When I turned the control, it made a scraping sound inside. I opened it to learn that its wirewound resistive element was ruined, clearly the victim of a meltdown:

No chance of repairing something like this. I ordered a replacement and continued checking controls and resistors.

A glance at the schematic (yellow oval) explained why the repairman had moved one of the leads from the end terminal of the potentiometer to the middle. This bypassed the potentiometer, taking it out of the circuit.

With that modification, the width was not adjustable, but at least the television would operate.

Charred remains like this send up a red warning flag. Parts don't burn up for no reason! As I continue this project, I'll look for what caused this catastrophic failure.

Resistors and Other Bad Stuff

This set uses lots of power resistors, and a few of them were bad. This photo shows a wirewound resistor on top of the horizontal AFC chassis. The old one was open; this replacement will stand vertically as soon as I put in its mounting screw.

Mounted on the phenolic board for the high voltage doubler are two chains of resistors. My DuMont RA-113 uses a similar doubler circuit. In this TV, the 680K resistors in one chain had drifted far beyond the specified value, so I replaced them:

Everything on the doubler board, including these resistors, was coated with insulating paint to reduce corona from high voltage. Another corona-reduction strategy, which you should follow when replacing parts here, is to make all solder joints smooth, with no spiky protuberances. I might recoat the board with corona dope after I'm satisfied that the HV circuit works correctly.

I generally try to avoid "shotgunning" (unnecessarily replacing) resistors, which are more reliable than capacitors. In this set, however, the closer I looked, the more bad ones I found.

I paid special attention to the horizontal AFC subchassis, whose circuits demand precision. I had heaved a sigh of relief after replacing a dozen capacitors in that cramped space. It would have been wiser to check all of its resistors, then, too. By the time I finished this second pass, I had replaced about a dozen resistors, leaving only a few original parts in the subchassis other than the transformer and tube sockets:

Before powering up, I did another round of testing, relying on the resistance chart in the factory service manual. This uncovered more sketchy resistors.

First Power-Up

At last, I was ready to try the set, using a metered variac to monitor how much current it drew. The picture tube displayed a raster, suggesting that my CRT repair had succeeded.

This is not a coherent screen image, but a few things are working. The low-voltage power supply is providing heater and B+ voltage, and the high-voltage power supply is making enough juice to light the screen.

In addition, with a signal source connected, I could hear faint audio from the speaker. This told me that the tuner and audio sections worked, at least marginally.

This trial lasted only a couple of minutes, long enough to observe the screen and make some basic voltage checks.

I noted with dismay that the TV drew more than 360 watts, well over the specified 300 watts. One common cause for this would be a short circuit in a power transformer (bad news, if true). The screen was also very dim, suggesting inadequate high voltage, which in turn might mean a defective flyback transformer (worse news, if true).

To add to my worries, when I turned the TV back on for more checking, the raster and audio disappeared.

Get Busy, Sherlock!

Musing over ugly possibilities, I pulled the plug and put my detective hat back on. The evidence of a past meltdown, combined with excessive current consumption, made me loath to reapply power until I had thoroughly examined the TV.

First, I checked my dozens of capacitor and resistor replacements, confirming that the right values were installed in the right places, and ticking off each one on the schematic to ensure that nothing was missed.

Finding no mistakes, I went back to the manual's resistance chart, repeating measurements for each pin of the TV's 28 small tubes.

In the course of these two passes, I replaced even more marginal resistors.

Checking the Horizontal Output and AFC Circuits

This TV's horizontal output circuit is a little different. It uses a 6L6 tube where my other 1948 sets use a 6BG6. It also has no high-voltage damper tube and no adjusters for horizontal drive or linearity.

I paid close attention to every component around the 6L6 tube and I traced continuity around and through the flyback transformer, confirming that its windings were not open (although they still might harbor small internal short circuits).

The horizontal AFC subchassis holds the rest of horizontal sweep circuitry, so I went through the same drill there.

Testing the Power Transformers

In a conventional design, you could remove the horizontal output tube and power up the set, at least briefly, for voltage checks. Since this setup was unfamiliar, I asked the VideoKarma TV forum whether this seemed safe.

Bill (aka oldcoot88) endorsed pulling that tube and he also suggested a basic test for the power transformers: monitoring the current draw and heater voltage after pulling the 6L6 tube, then pulling the two 5U4G low-voltage rectifier tubes one by one, and finally measuring the draw with all tubes removed. Here are the results:

Drawing only 17 watts with all of the tubes removed suggested that the power transformers had no shorts. Good news!

Executive summary: after hours of testing, I hadn't found any big problems, although I had replaced a few more components as a precaution. Of course, it's reassuring to know when things work correctly. However, the TV's heater voltage was still too high, causing excessive current consumption.

Will The Bad Tubes Please Raise Their Hands?

While all of the tubes were out, I decided to re-test them. They had been tested when I put them in, but as long as I had spent hours eliminating other potential trouble sources, why not eliminate these, too?

Surprisingly, three tubes tested bad: one in the tuner, one of the video amps, and one of the 8016 high-voltage rectifiers. Or, perhaps not so surprising, I grumbled, since I'd scrounged pretty hard to find 21 tubes in my workshop, even borrowing a few from other sets. Those three might have been marginal in the first place, and just happened to give up the ghost.

No wonder my previous trial ended with a blank screen and no audio. A dead HV rectifier means no high voltage, hence, no light on the screen. A dead tuner tube means no signal, hence, no audio (or video, for that matter).

Our First Real Image!

I was running short of tubes at this stage, but I located three replacements, finding fresh 1B3GT and 6AC7 tubes in my stash and borrowing a 6J6 from my DuMont RA-103.

After installing all of the tubes, I tried another of Bill's suggestions: using a variac to reduce the line voltage until the TV drew the specified 300 watts, and then measuring the heater voltage. He guessed that the correct heater voltage would be achieved with a line voltage around 105-110 volts AC.

Bingo!

As predicted, 110 volts on the line input gave a heater voltage of 6.3 volts. Although the horizontal frequency was initially far off, after adjusting that and a couple of other controls, my 661-P made a real picture for the first time in decades.

The image was stable, but problems remained in the vertical circuits: notice the bright foldover band at the bottom, for instance. The height and vertical linearity controls were overly interactive, too, making it nearly impossible to get sensible linearity without distorting one part of the image or another.

Back to the workbench, where I finished replacing every last resistor related to the vertical tubes and controls.

A Better Image

After those replacements, and further adjustments of the picture controls, the image looked much better.

Now we have good contrast and brightness, with excellent focus. The vertical foldover is cured and the horizontal and vertical stability are great.

Although the audio was initially weak and fuzzy, I improved that by carefully tweaking the audio IF adjusters while listening to a weak signal, aiming at maximum volume and clarity. This "bareback" adjustment isn't as precise as following the factory procedure with a signal generator and oscilloscope. However, the TV now has clear sound and plenty of volume, so I'm content for the meantime.

For the previous photo, I tried to adjust the height and width to fit the small mask in the Capehart's cabinet. The vertical is correct, but I couldn't reduce the width enough. Excessive width can be a symptom of weak high voltage, so that remains on my to-do list.

Reducing the Line Voltage

It's clear that I can get good performance from this TV at reduced line voltage. The question is how to manage that. The line voltage at our house generally runs around 120-121 volts AC. With that input, the heater voltage (hence, the entire TV) runs too "hot," drawing too much current.

The old width control was burned up and I had found other toasty remnants here and there. While playing the TV earlier, I could feel warmth on the chassis by the width control. I certainly don't want the control to overheat and melt down again.

Speaking of the width control, the Capehart manual has this entry in a list of problems and possible causes:

The TV was now playing happily and I had previously inspected the 6L6 wiring, so that eliminated possible causes A and C, leaving cause B: breakdown of filter capacitor and short of B+ voltage to ground.

Of all the ills that old TVs and radios can suffer, bad electrolytic filter capacitors are just about the most common. I'll bet that someone fried that width control by turning on the TV after it had sat unused for years and the power-supply filter caps had dried out.

There are various ways to reduce line voltage. I could use my variac, of course, but I need it for everyday use, so that's not ideal. I could also build a humbucking transformer and plug the set into that, but I didn't have the necessary parts on hand.

While thinking about this, I remembered that one of the power transformers (labeled Television in the manual) has a voltage switch on top.

The manual calls this the voltage selector and it specifies the line input as a range from 105-120 volts AC. Clearly, the intent is to let you adjust for different line voltages. This schematic of the low-voltage power supply shows the voltage selector on the Television transformer (blue oval):

Looking at my voltage selector, I surmised that it was on the lowest setting, for 105 volts. If I switched it to 120 volts input, that should reduce the transformer's output and solve the problem!

Unfortunately, I couldn't get the switch to budge, after removing its setscrew. It couldn't be turned, pulled upward, or moved at all. Perhaps it had never been moved since 1948 and now it's stuck for good.

Then I remembered another item from the workshop: an old RCA "TV Isotap" that I bought at a swap meet years ago.

Built for TV service, the Isotap lets you change your incoming line voltage to different AC voltages and it provides both isolated and direct outputs. In the photo, I had marked our house's line voltage at the time, 121.2 volts, and the voltages available at the Isotap's plugs when its switch was set for an incoming level of 125 volts.

Problem solved! I almost never use the Isotap (a variac is more flexible), so I can leave the 661-P plugged into it indefinitely, using an output that approximates 110 volts.

Mid-Course Status Check

At this stage, I'm satisfied with most of the electronics, but The horizontal linearity isn't perfect. The test pattern shows how the left half of the image is stretched relative to the right half.

This degree of non-linearity isn't very noticeable in everyday viewing, and, given that the TV has no horizontal linearity adjuster, I may have to live with it.

Since the horizontal AFC system was a little unfamiliar, I was curious to see what sort of waveform it produced. This photo shows the sawtooth waveform that is fed to the 6L6 output tube:

It's a very stable waveform and the horizontal lock is rock solid. The designers were so confident in this system that they didn't provide a horizontal hold control on the front of the cabinet, and in fact, none is needed in everyday viewing.

The only front horizontal control is the Framing control, which lets you adjust the horizontal centering. This is in addition to the centering control on the rear apron, so I'm a little curious why the designers provided it in front. Centering is not something you need to adjust every day, and this TV has outstanding horizontal stability.

Tuner Mechanism

As noted earlier, this television's tuner uses air variable capacitors with mechanical detents. These photos show the tuner from below and from one side.

The following photo shows the detent mechanism. The lower yellow arrow points to a spring arm with a ball that clicks into little slots on the disc edge, securing the tuner at the right spot for each station.

The upper yellow arrow points to a hinged lever connected to a "clutch" that's mostly hidden inside the tuner. This switches between the low VHF stations (2-6) and the high ones (7-13). As you may know, the VHF television band has a gap between stations 6 and 7, which is partly occupied by the FM radio band from 88-108 megahertz. When you move the tuner between channels 6 and 7, the clutch automatically switches between the low-channel and high-channel ranges, skipping the FM radio band entirely.

It's interesting to compare this tuner to those in my DuMont RA-103 and DuMont RA-113. DuMont made a virtue of the fact that their continuous tuners covered the FM radio band, providing radio reception as a "free bonus" integral with the television.

Experimenting with a Width Sleeve

Earlier in the project, I had been unable to reduce the width enough to fit the CRT mask. Bill suggested trying a width sleeve, which was a new concept to me.

I found an explanation in an old TV service book:

I cut a strip of copper foil, rolled it up, and taped on a cardboard arm, so that I could slide it in and out of the yoke after reinstalling the ion trap magnet and CRT socket:

Just as advertised, the sleeve was effective in reducing the sweep field. So effective, in fact, that I could block the sweep completely by sliding it in only a short way.

Clearly, that sleeve was wider than necessary, and there was so little clearance between the yoke and the CRT neck that sliding it was extremely tricky. I cut the sleeve down:

This narrower sleeve's electronic effect was less drastic, but it was even more difficult to slide. The soft foil bent easily and got stuck at the slightest obstruction. No doubt a stiffer metal would work better, but I had nothing suitable on hand, so I tabled the experiment for the time being.

Later in the project, when I prepared to reinstall the chassis in the cabinet, I made the final round of tweaks, including adjusting the yoke. Now, I was able to set the width without any extra tomfoolery. I'll remember the width sleeve trick in case I need it in a future project, however.

The biggest difficulty in this case resulted from the type of ion trap magnet used by Capehart. It's a wide collar that has to be pretty close to the yoke for optimum brightness. This scheme would work better with a CRT that uses a different style magnet, or no magnet at all.

Cabinet Restoration

After puttering with the electronics for weeks, I finally turned my attention to the cabinet.

Safety Glass

First order of business: removing the CRT safety glass and cleaning the mask. Here are before and after photos. In the first photo, notice how much sooty dust I had disturbed with one light finger wipe. The soot came off easily, using clear water and a damp towel.

From casual inspection, I had assumed that the mask was a piece of butter-colored plastic behind the safety glass. In fact, it is amber paint on the wooden cabinet, whose cutout opening has gently rounded edges.

After cleaning the glass, I reinstalled it. For the first time, I can see the TV working in its cabinet with knobs in place.

Before replacing the glass, I made a high-resolution scan of it. The painted lettering and pinstripe lines are in perfect condition, but it has little cracks in two corners of the glass. With a scan in hand, it would be possible (although difficult) to make a replica if I can't live with the cracks.

Calling in the Pros

After taking a closer look at the cabinet, I decided to turn it over to a professional. I have done worse cabinets than this one, but I don't enjoy refinishing as much as the electronic work, and when I'm not having fun, I tend to get impatient and cut corners.

I removed the back cover, safety glass, and speaker board, and then brought the cabinet to Michael Mueller in Seattle for a facelift. A short time later, the cabinet was back, looking like new. Thanks, Michael!

Final Assembly

Now for the most enjoyable part of every project—putting the restored chassis back into the refinished cabinet. Here's the chassis, ready to go.

Incidentally, after taking these photos, I substituted a metal 6L6 for the glass 6L6 horizontal output tube I had been using. The manual specifies a metal tube, but I found no perfomance difference between glass and metal, although swapping tubes (not surprisingly) required readjusting the horizontal frequency.

Here are some photos of the finished project.

The audio quality from the 12-inch Capehart speaker is very nice, and I appreciate having a tone control, a rather unusual feature in such an early television.

Here's a brief video clip showing the TV in action. The sound quality from my video camera leaves much to be desired, but at least you can tell that the whole thing works.

Final Thoughts

This was an enjoyable project. Repairing the CRT was simpler than I expected. The cramped chassis was not a joy to work on, and I ended up replacing more components than usual, but the final result was very gratifying, and there's some intriguing history behind this television.

I would like to replace the grille cloth, which has some scrapes and a stain that I wasn't able to wash out. Here's a close view of the pattern, which has 1/2-inch squares.

I haven't found a decent match with any of the usual suppliers, so if anyone knows where to find a suitable replacement, send me an email. The original color is a darker maroon, as I could tell by removing the speaker board and inspecting the edges that were protected from UV fading.