RCA Model CT-100 Color Television (1954)

The RCA CT-100 is a Holy Grail for TV collectors. Introduced in 1954, it was the first color television sold in significant numbers, helping to launch a revolution in TV broadcasting. Very few CT-100s were sold during its short production run and only two to three dozen working examples can be found in the world today.

Meet the RCA CT-100 Color Television

The early 1950s were an exciting time in television. In 1946, RCA had scored an early success with its Model 630TS and established itself as an industry leader. During the next few years, customers lined up to buy TVs and manufacturers offered newer sets with bigger screens, lower prices, and better performance.

Television companies raced to supply this booming market with color and in 1954 that dream became reality. Although Westinghouse beat them to market by a few weeks, RCA's CT-100 was the world's first color TV produced in meaningful numbers.

With a 15-inch color picture tube, the CT-100 had a regal list price of $1000. For comparison, the base price for a new 1954 Chevrolet Bel-Air automobile was $1095.

Here is my CT-100 partway through restoration, before the cabinet had been refinished.

Color television was a huge novelty in 1954. Owners of CT-100s often invited friends and neighbors to goggle at this new invention. For many people, the CT-100 offered their very first color TV experience, even if it was only to peek at a demo in a dealer's showroom.

RCA built approximately 4,400 CT-100s. There are only about 160 known survivors, and of those, for reasons we'll explore later, only a handful of sets actually work. Some CT-100 owners call themselves "custodians," underlining the importance of preserving these rare TVs. I certainly feel lucky to have one.

The CT-100 was advertised by many local department stores around the country, but as far as I know, RCA published only one national ad for the set. This advertisement ran in the Seattle Times on Sunday, August 15, 1954:

Expensive and cantankerous, the CT-100 was soon superseded by TVs with bigger screens and improved electronics. RCA's second color television came in 1955, model 21-CT-55 with a 21-inch screen.

Still, the CT-100 marked an unforgettable "first." There are other working examples in addition to mine, but the only ones that I've seen personally are in museums, in Washington and Ohio.

CT-100 Electronic Design

There's much to be said about the electronics of the CT-100, so much that I put it in a separate article, RCA CT-100 Color Television Design. That piece covers design highlights; discusses CT-100 color demodulation and the NTSC color standard; and provides a library of reference material including a complete RCA service manual and the NTSC specification.

Here are the operating instructions from the CT-100 Owner Manual. These pages describe the user controls and the correct sequence of operation.

This diagram identifies more of the CT-100's controls:

A number of these controls are reserved to the serviceman. The upper row of controls is behind the hinged pencil box cover. Those in the lower row are accessed by removing the pencil box cover and wooden panel. The Focus and Convergence controls are in a small recess in the cabinet's side.

The next diagram gives a birds-eye view of the chassis and identifies all of the tubes and controls, including some service adjusters not shown in the previous diagram.

We'll refer to both of these diagrams later. Note: CT-100 schematics were published by Sams and by RCA. In this article, I'll refer to parts by their Sams numbers, since those are widely known and both Len and I used them in our restoration notes and charts.

CT-100 Restoration

This rest of this article relates some highlights of my experience with a CT-100. I was aided along the way by fellow collectors Pete Deksnis and John Folsom, two gurus of CT-100 restoration.

Finding an RCA CT-100

In May, 2009, I was contacted by a collector named Len Dole, who owned a CT-100. He had begun restoring it, but health issues prevented him from finishing the project and he wanted to pass it on to someone who would complete the work. He sent these photos:

We struck a deal and, with my son Peter, I flew to California and rented a moving van to bring the television back to Washington state.

In addition to the TV chassis and cabinet, we packed up two 15GP22 picture tubes, an oscilloscope and other test equipment, and several boxes of CT-100 parts and electronic components. In the third photo, after two very long days of driving, I have unloaded enough stuff to uncover the real goodies in the back.

Taking Inventory

Len had gotten this CT-100 from a neighbor, the original owner. Len had never tackled a TV, but he was an engineer and had restored several old radios. He began a total restoration, recognizing that it would be a major project.

Len made a lot of progress and exchanged email with Pete and John several times. He replaced all of the original paper and electrolytic capacitors and many resistors that were out of tolerance. He found replacements for a couple of other components: a new vertical dynamic convergence transformer purchased from John and a horizontal centering control donated by Pete.

My first task after unpacking was to inventory what I had. It looked daunting at first.

In addition to the chassis, cabinet, and two picture tubes, there were several boxes of parts, new and old, and assorted restoration supplies. Some things were obvious, others were not. I found two hinged "pencil box" covers for the front panel controls. One of them presumably belonged this cabinet. I found a box of knobs, some broken, others intact, and a few that looked like they belonged to other devices.

Len also left four notebooks with printouts of CT-100 articles, annotated copies of the schematic, and charts showing measurements of resistance and voltage. I'm sure it all made sense to him, but flipping through them, I couldn't tell exactly where he left off. Lots of work had been done, but how much remained?

CT-100 Provenance

Three days after we arrived home with the TV, Len passed away. This meant, sadly, that he wouldn't be around to see his project completed. It also left me on my own.

Here is what Len's widow told me about the TV's history.

The CT-100 was bought new by a local man who was an engineer at the original business here. Many of the old-timers who are still around remember going to his house to watch it with him. It was their only TV for many years until they eventually replaced it (I don't know when). Being an electronics engineer, he kept it in his garage, along with the spare tube that he had acquired along the way. After his wife died, he was cleaning the garage and decided to give it to Leonard, because of his interest in old radios. We moved here in October 2003 and the TV joined us shortly thereafter, and Leonard had a new hobby.

It's fun to know where the television came from. The original owner kept it for fifty years, and Len had it for six, working sporadically on restoration during that time. As third owner, my mission is to restore it to original working condition and preserve it in that state.

Serial, Model, and Chassis Numbers

My CT-100 has chassis serial number B8003603 and cabinet serial number 429. The chassis number is stamped below the flat resistor on the back panel.

CT-100 chassis serial numbers are believed to begin at B8000100. The earliest known number is B8000158 and the newest is B8004019. With chassis B8003603, my set falls near the end of the production run, among the newest dozen CT-100s that have been identified.

The CT in RCA model numbers and chassis numbers means Color Television. CT-100s have chassis type CTC-2 and you can see that number stamped in black ink on the rear panel of mine. Subsequent color RCA chassis were named CTC-4, CTC-9, and so forth. My CTC-11 was made in 1961.

You can read more about cabinet numbers in the cabinet restoration section below.

Electronic Restoration

You have to start somewhere. I decided to install the new vertical convergence transformer and reconnect everything else in the disassembled high-voltage cage. In the meantime, I'd continue sifting through boxes to determine if I had enough parts to make a working TV.

Restoring the High Voltage Supply

Here is the new convergence transformer, with the connection diagram supplied by John Folsom. Next to it is the case that contained the original transformer.

A bird's-eye view of the high voltage cage reveals several points of interest. Suspended from a tower in the center is the 3A3 HV rectifier tube. It will supply about 20,000 volts to the picture tube.

Two tube sockets are visible on the cage floor. The black socket on a bracket holds the 6BD4 high voltage regulator tube. To its right, screwed into the top of a round "doorknob" capacitor, is the socket for the 1X2B focus rectifier tube.

Mounted on a white plate on the other side is a big round brown potentiometer: the vertical dynamic convergence control (Sams part number R124). Two striped yellow 50-megohm resistors are wired to the control. One will be wired to the red lead hanging loose nearby. The other will be pushed into the little metal cup projecting to the left of the 3A3 tower.

Len had unwired and removed the 1X2B socket for some unknown reason. Also missing was a 500-mmf doorknob capacitor. There were various other loose leads to connect, not to mention the convergence transformer.

Wiring inside a high voltage cage leaves zero room for error, so I made a diagram showing the connections I needed to restore or check.

The dotted rectangles represent the two white plastic boards at the back and side of the cage. Of course, the complete circuits include things not shown here.

This hookup differs from the factory wiring in one detail. For John's replacement transformer, you add an 82K resistor to lead 2, which connects to the junction of C118 and the lead to CRT pin 6. Since C118 is a doorknob cap, I wired the resistor to a screw connector and insulated it with shrink wrap. The second photo shows that connector in place on the doorknob cap, along with the lead to CRT pin 6.

In a box of little parts, I found the original 1XB2 tube socket and a new one of the same size. The old socket was not damaged, and the new one would have required some adaptation, so I decided to reinstall the original. The next photo shows the socket just before I wired it up. The two black leads go to the filament coil, which is a simple loop of wire going through the base of the flyback transformer. The white lead comes from R23, the focus potentiometer (see diagram above).

The next photo shows the HV cage from the left side. The transformer is in place and everything else is hooked up. The 6DB4 tube stands in its bracket to the left. To its right is the little 1XB2 focus rectifier. One of the striped 50-megohm resistors stretches down from the convergence pot to plug into its socket on the 3A3 tube tower.

Note the little label: H. DYNAMIC CONVERGENCE. Len labeled many of the CT-100's controls, and there a lot of them. The CT-100 has over 40 potentiometers and who knows how many coils and transformers. The front panel alone has 24 adjusters; others are scattered around the chassis.

Len made a minor change in the wiring for the 6BD4 tube socket. The 6DB4 does not use pin 8, and the factory wiring employs that pin as a tie point for two unrelated components. Len installed a new connector near the tube socket and moved the pin 8 component connections there.

After this change, it's possible to use a 6BK4 tube as well as a 6BD4 in that socket. I don't know why Len considered this important, since both tubes are readily available for a few dollars. In any case, the change does no harm. In a box of miscellaneous parts, I found a weak 6BD4 tube and a strong 6BK4. I installed the 6BK4.

Time Out to Double-Check

In the course of all this, I exchanged some email with Pete and John, whose advice was invaluable. Pete also referred me to two pages on his website with detailed notes about CT-100 power supplies.

I then spent a few hours double-checking my work and closely inspecting all of the under-chassis work that Len had done, looking for any other changes from the schematic or possible wiring mistakes. Finally, it was time to apply power and find out how much of this TV functioned.

First Power-Up

Len had previously powered up the chassis for voltage checks, but never using this new transformer, so I was curious and apprehensive about the trial. As recommended by Pete, I connected everything except the picture tube and field neutralization coil. That meant plugging in the yoke, speaker, and purity coil.

Here, the yoke is mounted on a temporary stand. It's the big black and white doughnut shape in the middle of the chassis. When installed in the cabinet, the yoke hangs from a bracket on a thick wooden crossbar. The purity coil resembles a steel coffee can. In the cabinet, it will be located on the CRT neck. Here, it rests on a little magazine atop the low voltage power cage. The speaker is in a cardboard box to the side.

You can't see it in this photo, but I placed the CRT socket in a glass jar. That socket has three high voltage leads and the glass insulator prevents high voltage arcing to other things (including my body!). Although it's not strictly necessary, I also installed the HV cage cover.

Two other devices are visible on the bench. The clear box with dials is my metered variac. I used this to gradually increase the line voltage, keeping a close eye on the ammeter for signs of excessive current draw. Next to the variac is a DVD player. I wouldn't see a picture without a picture tube, but the rest of the TV—including its audio section—might work.

Crossing my fingers, I cued up a DVD of the Wizard of Oz, selected Channel 3 on the CT-100, and brought up the voltage.

The tubes began to glow. After a few seconds I heard the beginnings of audio, followed by the high ring of sweep circuits spooling up. Before you could say Professor Marvel, the movie soundtrack came through loud and clear! Quick checks showed reasonable B+ voltages and—even better—healthy high voltage, as well.

I babysat the television for a while and made a few more voltage tests. Finding nothing amiss, I turned off the power and scored this maiden flight a success.

Reality Check

Energizing the chassis without burning it up is a positive step, but of course the TV could still be far from functional. Len had replaced a few potentiometers, among other things, and some adjustments can't be made without connecting a picture tube, a stage he had never reached. Of those dozens of controls, many might be set to completely random positions.

Len also was fond of replacing wires. He completely replaced the cables for the field neutralization coil, purity coil, yoke, and picture tube. He also replaced a number of wires under the chassis.

It's unlikely that all of those wires could have been ruined in normal operation, and there was no sign of mouse chewing anywhere on the chassis. I suppose that Len was just being thorough. Although the power-up trial hadn't shown any problems, replacing a lot of wires offers many opportunities for error. Until I installed a picture tube and put all of the TV's circuits under load, I couldn't know what would happen.

Testing the 15GP22 Color Picture Tubes

The biggest question mark was the condition of my two 15GP22 picture tubes. These rare tubes are expensive and notorious for losing their vacuum (see the 15GP22 section of my design article for more technical information).

The thought of testing these tubes made me chew my fingernails. I had no definite history for them. Len's widow believed that one was original to the set and the second was a spare. A couple of old emails from Len indicated that he applied new aquadag coating (see below) to one and tested its filament. Possibly this was the original owner's spare, but that's only a guess.

Lacking a CRT tester, Len had first checked the tube's filament for continuity and then applied 50% of the rated filament voltage for one week. Had the tube leaked air, that trial would have burned up the filament. It survived the test, so it still held vacuum—at that time, anyway. This was two years later, and the tube had traveled over 1,000 miles in the back of a moving van, among other things.

I began by testing the "good" CRT's filament for continuity. That looked normal, so I connected it to my Sencore CR70 tester, using the universal adapter. Applying filament voltage revealed a welcome glow, so I proceeded with the test. This photo shows the results for the blue gun.

Wahoo! Not only the blue gun, but red and green showed strong emission, giving almost exactly the same numbers on the tester.

I had little hope when testing the "original" tube, but to my delight it also showed life. Its green gun was weaker than the other two, but all of them had emission. This tube should be ideal for putting in the chassis while I complete the restoration. Its colors may not be perfect, but that may not matter, depending on what issue I'm working on, and meanwhile my best tube can rest safely in its box.

It's common for a long-disused tube's emission to increase after running for a few hours at normal filament voltage, so I gave each tube a leisurely spa treatment on the tester.

Because the 15GP22 is leak-prone, some collectors apply Vacseal vacuum sealant to the flange problem areas. I'll illustrate that process later on.

Installing the 15GP22 CRT

Here you see my second-best 15GP22 picture tube and related parts, ready to install.

Notice the metal flange encircling the tube face. We saw that flange's cross section earlier in a diagram. The white lead clipped to the flange will deliver high-voltage current to the tube's anode. A thick black plastic gasket will encircle the flange to insulate it, with the anode lead peeking out between the overlapping gasket ends.

To the left of the picture tube are two metal parts. In RCA literature, the heavy steel cone is called a mu-metal shield. Mu-metal is a nickel-iron alloy with high magnetic permeability.

The shield slips onto the bell after you have installed the gasket. The circular brace around the shield has four holes for strong metal stays that will hold the tube in the cabinet.

As I was preparing to install the tube, Pete Deksnis mentioned that the tube's blue gun must be located at the top. The anode lead must be clipped at the right place on the flange perimeter—roughly 3 o'clock, viewed from behind—or it won't reach this short female lead at the front of the high voltage cage.

The next photo shows where the tube will go. Some of the needed parts are lying in the cabinet. The black wooden crossbar holds a metal bracket from which the yoke will hang. Four metal retaining rods with brass nuts are lying in front of the crossbar, along with one of the gray plastic braces that press the tube's rim from the top. The second gray brace is hanging from the cabinet's top frame.

The red lead connects to the field neutralization coil, a large circular coil that encircles the tube's face. This lead will plug into the back of the low voltage cage.

At the lower right of the CRT support, notice the thin copper ground strap. When the chassis is installed, this must be clipped to the chassis, providing a ground connection for the CRT, yoke, and purity coil.

With help from my son, Isaac, we put the cabinet onto a dolly, wheeled it into the workspace and brought in the CRT. For the first time in months, all the major parts of my CT-100 were together in the same room. With any luck, this will soon resemble a TV rather than a "basket case" of scattered parts.

Before lifting the tube, we donned jackets and safety goggles. There's a risk of implosion when handling any CRT, and color tubes are heavy!

You install the CRT with the cabinet lying face-down. Full instructions are in the RCA service manual provided in my design article.

When mounting the 15GP22, make sure that its internal mask is squared off with the mask in front of the cabinet. In the next photo, Isaac stands behind the cabinet, our installation complete.

See how the top panel lifts off, simplifying this and other service chores. The retaining rods reach from four cabinet corners to hold the CRT in the stout front bracket. The gray triangular braces have also been snugged into place and tightened against the top of the tube's bell.

Installing the Chassis

We're almost ready to put the chassis in the cabinet. At this moment, the yoke was still mounted on its temporary stand. The opening in the cabinet's right side gives access to two controls: focus and vertical dynamic convergence. You can see those control shafts on the right side of the big black high-voltage cage on the chassis.

This photo shows the yoke on its mounting frame and wooden crossbar. To install it, we'll flip it over, slide the yoke forward along the CRT neck (white gasket facing the front), then secure the crossbar to the cabinet with four bolts.

Notice the metal strap with a round connector, leading from the left of the yoke. This will plug into the metal case of the purity coil, thus connecting via the thin springs to the CRT's mu-metal bell housing.

Together at last! The chassis and CRT are in the cabinet. The low voltage power supply cage is on the left, just in front of the power transformer. The black cylinder is a ballast. Some people call this a "ballast tube," but it's simply a few power resistors in a ventilated steel can with a tube-style base. The high voltage cage is on the right. The steel box projecting from its back contains our new vertical dynamic convergence transformer.

Suspended above the chassis is the yoke. It bolts into a steel bracket attached to the black wooden crossbar. Behind the yoke is the coffee-can-like cylinder containing the purity coil. Its two-wire lead plugs into the right of the low voltage compartment. Up and to the left of that plug is the purity control, which Len had labeled Cross Purity. We'll see more of that control later on.

Three screws project like spokes from the purity coil case. Each has a knurled adjuster nut. The screws (not the nuts) are the convergence magnets, one for each color: red, green, and blue. In the convergence process, these are used to merge three colored dots into a single white dot on the screen.

Most of the cables for the picture tube socket come from below the chassis. Two white leads come out of the high voltage cage. They carry 4 kilovolts and 10 kilovolts, respectively, to pins 6 and 13 of the CRT.

One thing you won't see in other CT-100 chassis photos is a fuse panel. Len installed a panel with four fuses on the back of the low-voltage cage, between the transformer and the round ballast. The fuses protect components on the TV's four B+ voltage lines, an excellent safety addition.

First Picture (Cross Your Fingers!)

With everything hooked up, it only remained to power up the television.

This is always a nervous moment. The CRT tester said that the picture tube was functional, and I had spent considerable time testing the chassis under power on the workbench. That didn't guarantee success, however. Many adjustments can't be made when the chassis is naked on a workbench, and the CT-100 is known for needing lengthy setup in the best circumstances.

In the best case, I would get a picture that required hours of tweaking before it looked right. In the middle case, I'd get some kind of display, but with gross defect(s) from causes not discovered in bench tests. Worst of all would be a meltdown: one or more parts failing under real-world conditions and possibly causing other damage.

I took these snapshots during the first minutes of operation. Here's the Wizard of Oz.

I guess I'd call this the middle case: some kind of display, but with gross defects.

The colors are bad, but that's to be expected when you haven't even started routine setup. A bigger problem is evident, seen more clearly in the first photo. Although the image is complete, it's squished into the upper half of the screen. The electron beams aren't hitting the lower half of the picture tube.

The image was also tilted horizontally (first photo), but that was easy to fix. You simply loosen two wing nuts on the side of the yoke and rotate it in the bracket. In the second photo, I have leveled the image.

Half a Screen is Better Than None

Solving the vertical problem took a while. With advice from John Folsom, I confirmed that the new vertical convergence transformer was installed correctly. Among other things, he supplied this photo showing what waveform you'd expect to find on the convergence lead going to pin 13 of the CRT. In his words, "The fizz is the horizontal convergence waveform riding on the vertical convergence waveform."

The image on my scope wasn't as pretty, but at least I knew I hadn't reversed the transformer leads. In the course of further discussion, I emailed these images to John, remarking that the picture wasn't all that bad if you discounted the vertical problem (first photo). I also noted a blue "half moon" effect in the lower half of the image, clearly visible when the TV was switched to a channel with no signal (second photo).

John's reply was prescient:

I am not sure what your rising moon artifact is trying to tell you. Have you
tried sliding the yoke back and forth on the neck of the tube? Try also adjusting
the purity control, this can cause odd effects if turned up too high. If neither
of these makes it much better, probably time to put it back on the bench to see
what is up with the vertical.

Moving the yoke and purity coil had no effect on the vertical problem, but turning the purity control did.

When I turned that control all the way counterclockwise (near zero ohms), the picture suddenly jumped down into normal territory, and I was quickly able to adjust its height and linearity to fill the whole screen.

Disregarding color for a moment, a large number of things have to work correctly for a TV to display a picture like this. High voltage is somewhere in the right ballpark. We have good screen geometry, good focus, decent brightness and contrast. The picture is stable, with no rolling or jitter.

Most exciting, the precious 15GP22 picture is alive and well. Despite one gun looking weak on the tester, this "second best" tube should be good enough to use while trouble-shooting, at least.

Replacing the Purity Control

It didn't take long to confirm that the purity control potentiometer was defective. This is a wire-wound type pot and its winding had opened near the zero-ohm end of its range, making it work more like an on/off switch. Through most of its range, the resistance was infinite. This produced the half-picture seen earlier. Only when turned near zero did it conduct a little current and energize the purity coil enough to fill the screen.

The next photos show the interior of the old control, with its circle of resistive windings, and the new replacement installed on the back panel of the low voltage cage. It's the pot at the upper left.

Replacing the purity control made a clear improvement.

Now I could move the purity control throughout its range without losing vertical height.

Excessive Horizontal Drive

We're getting closer to a decent picture. I noticed something new, however. If you look carefully at the previous photo, there's a faint lighter vertical band running down near the spout of the Tin Man's hat. It was much more evident with different brightness and contrast settings.

When I switched from a DVD to antenna reception from my in-house transmitter, the problem leaped out.

It's impossible to miss the white vertical line, now. This is a familiar symptom, caused by excessive horizontal drive. Turning the drive control all the way down minimized the defect, although a very faint line remained in some images. (This picture is snowy because my transmitter's signal is weak in that part of the house.) The horizontal drive adjuster is located on top of the chassis, as you can see in the chassis layout diagram near the beginning of this article.

I made a note to revisit all of the horizontal settings after dealing with more basic issues. Meanwhile, I was pleased to find that the horizontal lock was very stable. Once things were basically set up, I could turn the horizontal hold control all the way from one side to the other without losing sync.

Rough Convergence Adjustment

Playing movies is fun, but real setup must be done with a pattern generator, following the manufacturer's instructions. After hooking up the generator, the first thing I looked at was the set's convergence using the generator's dot pattern. After making some rough adjustments, it didn't look too bad. The first photo shows the entire screen. In the second, I zoomed in and reduced the contrast so you can see all three colored dots.

In a color TV, every dot on the screen is formed from three dots, colored red, green, and blue. When the TV is perfectly converged, the electron beams for all three color guns converge so that the three dots merge into a single white dot.

In the second photo, you can see that the dots aren't quite converged. The red dot is a bit left and above the center. A bit of blue can barely be seen peeking out "behind." (Actually, since red and blue are mostly converged, that dot is mostly magenta.) The green dot is a bit to the right and below the center. Its color is purer because it doesn't overlap the others much, if at all.

As bad as it looks in close-up, this degree of misconvergence can actually create a passable picture after other settings are corrected. You would notice the defect most clearly in a high contrast scene. If a man is wearing a dark suit with a white collar, for instance, the upper left edge of the collar may have a thin fringe of magenta or red, and the lower right may have a faint greenish shadow.

As I learned when restoring my two CTC-11 roundies (CTC-11 and CTC-11A), convergence should be good for hours of fun later on. It would be silly to spend a lot of time messing with convergence at this stage, when the set's purity is still badly misadjusted.

My main goal in displaying the dot pattern was to check the action of the vertical dynamic convergence control. Because the transformer had been replaced, I needed to confirm that the control—and hence, the transformer—were working right. The control moved the dots around exactly as described in the setup instructions, so I crossed that item off my to-do list, breathing a silent thanks to John Folsom for building this vital component.

Rough Purity Adjustment

Purity refers to the TV's ability to create a uniform, or "pure," field of red, green, or blue, across the entire screen. Electronically, this means that the red electron beam lands only on red phosphor dots, the green beam lands only on green dots, and blue only on blue.

Looking at this photo again, notice how the picture is pretty good in a black and white sense—nice detail, contrast, and so on—but the color is weak everywhere, and hues are wrong. The Scarecrow's face should be straw colored, not magenta on one side and greenish on the other. You could watch this picture, but nobody would enjoy it!

Another early photo showed more clearly that purity was off. I took this when the TV was switched to a channel with no signal. When you loosen the yoke and the purity coil and move them around, you can make various odd multicolored images similar to this. It didn't occur to me at the time, since I was concentrating on the vertical problem, but I should have remembered this from setting the purity on my other color TVs.

To adjust purity on the CT-100, you first loosen the yoke and slide it back. Then you grasp the front edge of the coil form with one hand and hold the can with the other. This diagram shows the yoke and purity can.

Turning the front edge of the coil form rotates the coil inside the can. You don't want to turn the whole can, since its magnets need to remain as shown earlier in the rear chassis photo, with the blue spoke pointing straight up over the blue gun. You alternate between small adjustments of the coil and small adjustments of the purity control until the color is uniform.

Purity is set one color at a time, usually starting with red because when red is correct, the other colors are often very close. To set red purity on the CT-100, you turn the Blue Screen and Green Screen controls all the way down and turn the Red Screen control all the way up. You also set your pattern generator to display a pure "raster," or blank gray screen.

Your goal is to produce an even red screen, green screen, and blue screen. Here's what I got on the initial go-round. (Ignore the moiré patterns, which come from my camera.)

Those are hasty snapshots, but you can see pretty uniform color in each one.

Rough Grayscale Adjustment

The next step is to set the TV's grayscale. The human eye perceives the three RGB colors as having different brightness levels, green being the brightest by far. We need to balance these colors' relative brightness so that none of them predominates in a gray or white display.

On the CT-100, I turned the three Screen controls back to around their midranges, then adjusted them individually so that the generator's raster signal appeared as a uniform light gray, not tinged with any other color.

There's more to setting grayscale for the full range of dark to bright, as we'll see later. For now, I just wanted an approximation.

Our First True Color Picture!

Okay, test screens are boring. Did purity and grayscale adjustments make a difference that we can see?

The answer is a resounding Yes!

Yeah, we still have issues, but this is vastly better than where we began, with a dark discolored image squished into the top of the screen.

There isn't much blue in these scenes of Dorothy and Glinda, but we know that blue is present because flesh tones—a complex mixture of all three RGB colors—look pretty good.

These photos also illustrate the misconvergence mentioned earlier. In the close-up of Dorothy, note the faint green shadow on the lower right of her chin and ear. In the convergence close-up seen earlier, the green dot was below and to the right of the other dots. Convergence defects are most apparent in areas like this, with a strong transition from light to dark.

Purity is a textbook example of something you can't meaningfully adjust while a chassis is on your workbench without a picture tube. You can check voltages and view waveforms until you're blue in the face, but no oscilloscope can tell you which way to rotate the purity coil for best color.

Are We Home Yet, Dad?

We have reached two big milestones: stable operation and a full color image. The 15GP22 picture tube looks strong, knock on wood. The CT-100 project is far from over, however.

The initial results are promising, but I haven't performed a full-dress setup. That will begin soon.

The TV has some definite tics, too. When I change video sources, switching from a DVD player to my home broadcaster, for instance, it seems overly sensitive to signal strength variations. The AGC (automatic gain control) may need attention.

I have also noticed a strange fluttering in picture brightness at some combinations of signal and brightness/contrast settings. John remarked that this might indicate a problem with high voltage regulation.

The cabinet will be a project unto itself. It needs cosmetic work and I need to install the speaker board, "pencil box" control cover, and various other bits and pieces.

Last but not least, I need to learn more about the CT-100's electronics. This design is quite different than sets such as my CTC-11s, which RCA made only a few years later.

Second Convergence

In the next session, I decided to give convergence another shot. This time, in addition to using the vertical dynamic convergence control and static convergence magnets, I adjusted two front panel controls: vertical convergence amplitude and horizontal convergence amplitude.

The next photo shows the result, with a crosshatch pattern on the screen. (Ignore the slight blue tint; my camera makes all black and white TV images look blue.)

Center convergence is acceptable: the center dot and the lines in the middle zone look white, as they should. Using the front panel controls, I was able to improve the edge convergence somewhat, but the farther out you move, the farther some colored lines diverge from horizontal or vertical.

Again with the boring test patterns. Will this monkey-business make any difference in a real picture? As you recall, I snapped this hourglass photo a few days earlier, after replacing the purity control but before doing any purity or convergence work. The colors were not only wrong, but just plain weak.

Partway through this convergence session, I photographed the same scene for comparison. There's no shortage of color, and the colors are moving into the right places. (For normal viewing, you wouldn't turn up the brightness and contrast this high, of course.)

We haven't reached the end of the convergence instructions. Using an oscilloscope, I should view the waveforms for the horizontal dynamic convergence circuit, and I'll get to that, eventually.

The front panel has four additional convergence controls: vertical convergence shape and convergence amplitude, and horizontal convergence phase and convergence amplitude. Using these with the other controls—and they are all interactive to some extent—I hope to improve convergence at the screen edges. We'll use more front panel controls when adjusting grayscale. Again, here are the controls:

Setting High Voltage

While reviewing service instructions, I recalled that I hadn't checked the high voltage since very early in the project. I tested the anode voltage and found it high, 21.5 kilovolts to be exact. It took only a minute to turn it down to the specified 19.5 kilovolts.

Adjusting Low-Medium-High Grayscale

I then returned to grayscale. Although I had made a rough setting, you ideally want to see a uniform light gray screen throughout all brightness and contrast settings.

This definitely needed attention. Sometimes, turning brightness and contrast up or down would shift an entire scene too much toward one color or another. Moreover, in some cases, darker tones changed color differently than lighter ones.

Setting the grayscale for the low-medium-high brightness range involves ten controls. All of these appear in the previous diagram, except for Brightness, which is located with the Power/Volume knob.

The instructions show what these controls do:

1. Tune in a black and white program or test pattern.
2. Turn the color saturation and contrast controls fully clockwise. Turn the brightness control to its near maximum clockwise position. Alternately adjust the red, blue, and green screen controls for a light gray or low brightness white picture.
3. Turn the contrast control to its approximate mid-range position. Adjust the green and blue video gain controls for a satisfactory white picture. Turn the brightness control counterclockwise until the picture just becomes visible. Adjust the green and blue background controls for a gray picture.
4. Repeat steps 2 and 3 until satisfactory gray to white pictures are seen through all positions of the brightness controls.

The three Screen controls are clearly assigned to bright pictures, since you adjust them with brightness and contrast turned up full blast. The Video Gain controls are for medium brightness, and the Background controls are for the lowest levels.

Here's a snapshot from the first grayscale go-round. I took it in a well lit room with the brightness turned up pretty far. Lots of my previous screen shots were taken in a dim room because that was the only way to get a decent image. Simply being able to take this kind of photo is another measure of progress.

Not bad, but I bet we can do better. It takes practice to get a feel for what all of these controls do and how they interact. Here is our patient after another round of adjustment.

Better, indeed. I won't squander your bandwidth with a video to prove it, but now I can turn brightness or contrast up and down without strange effects such as making Dorothy blush red like a lobster.

Color Relapse?!?

After doing a little cabinet work (below), I decided to take some photos with everything in place before removing the chassis to put Vacseal on the CRT and make various checks & adjustments on the workbench. Things looked good at first:

After the set had been playing for about half an hour, the TV was stable, but I needed to adjust the hue and saturation controls.

Then the colors began changing, slowly at first, then more rapidly. Eventually, I had horizontal bands of blue and green moving down the screen, washing out the red. See how Dorothy's lips turn from red to green.

If I left the television on long enough, red vanished completely. Here are before-and-after photos, with red and without.

I remembered a bit in my RCA Pict-O-Guide that might apply. In a troubleshooting section, the book notes that a bandpass amplifier tube with a heater-to-cathode short can cause colored "hum bars" in the display. Examples showed screens with horizontal color bands, similar to what I had seen.

I pulled out tube V26, the 6AN8 chroma bandpass amplifier. Sure enough, the tester revealed shorts! A treasure hunt through my boxes of tubes turned up a spare 6AN8.

This seemed to work, but the problem—or something like it—resurfaced after the TV had played for about 15 minutes. Which suggested that the problem was heat-related and that it was not solely due to a bad tube (although replacing that was a good idea, anyway).

This episode was a sharp reminder. In my haste to stick in the CRT and see if the TV could make a picture, I had gotten ahead of my usual initial procedures, such as cleaning all of the tube pins and sockets.

There are plenty of other basics that Len might have left undone, such as cleaning potentiometers with DeOxit. Next time I pull the chassis, I'll attend to those, making no assumptions about what was previously done.

Before breaking for the day, I cleaned and re-tested all of the color tubes.

Wait, Doctor—the Patient's Coming Around . . .

Returning the next day, I started the set from cold and and waited for the problem to reappear. Thirty minutes later, I needed to adjust the hue control a bit, but red hadn't vanished.

Tired of staring at color bars, I put in a DVD and found something else to do. I got through the first half of Gone With the Wind and the problem with disappearing red and color waves never came back. In this shot, Scarlett O'Hara's lips look as scarlet as ever. Fiddle-dee-dee!

Who knows, maybe the cause was as simple as a funky tube pin. After I complete my cleaning routines, at least that type of issue should be gone for good.

Waveform Photo Album

Now the chassis is out and I can access the underside. I took this opportunity to snap photos of several waveforms displayed on the oscilloscope. These can be compared to model waveforms in the Sams service manual (W38, and so on) to get an idea how the TV is performing right now. And, if something like the vanishing-red problem surfaces, I can compare "before" and "after" photos. Many waveforms can be checked without the picture tube connected. Here's the setup.

The test bench is an old packing crate. I slipped a few magazines under the tuner cage to lift it enough to avoid crushing the signal cable. The diagonal wood brace doesn't really support anything. It's there to prevent tipping the chassis over if I knock it with an elbow. The yoke, speaker, and purity coil are connected. The CRT socket is out of view, cupped inside a little glass jar that's taped to the chassis.

In the next photo, I'm measuring a vertical output waveform, which looks decent. One more for the album!

Chassis Cleaning

With the chassis back on the workbench, I am reminded how grubby it is, coated with the usual decades-worth of grime. You can see where Len rubbed away some dirt to stick a label by the I-Gain adjuster.

It looks like Len made a stab at cleaning the tuner case, but didn't get far. A bit of plating is scraped away, revealing copper underneath. Whatever he used around the edges of the case—possibly an acid—just made the plating darker.

Here are some cleaning materials: isopropyl alcohol for grime, naval jelly for oxidized spots, and metal polish for a little shine. I started with isopropyl alcohol, using a toothbrush and a rag. For tight spots, I use folded pieces non-scratch cleaning pad held in a tool.

Here's a little test patch near the back of the chassis. Notice the tube's reflection.

We still have surface oxidation, hundreds of dark dots. I'll reduce them with naval jelly, which is basically phosphoric acid. This stuff won't turn you into a mutant, but rubber gloves are a good idea.

I have started spreading the jelly, taking care not to slop it onto components or into little holes in the chassis. I'll add a bit more to make an even layer.

Naval Jelly works slowly. I leave it on for half an hour or so, moving it around a little if needed to prevent drying too fast. Damp paper towels will remove it. Here's that area after I removed the jelly and used a lick of metal polish.

Progress! The dots are visible if you look hard, but the overall appearance is better. Just about any sort of metal polish will work. I used some Mother's polish that happened to be lying around. Simichrome is good stuff, too.

Cleaning the whole chassis will take hours. I usually wouldn't bother, since the TV works the same shiny or dull, but this set is special.

Some people go to extremes with chassis, stripping off all components, sandblasting, and either replating or painting. This chassis doesn't need such extreme measures, which might do more harm than good. To preserve the original stamps and labels, I'll simply clean around them.

Here's the chassis after more cleaning. Not finished, but it looks more civilized.

Voltage and Resistance Tests

The chassis hadn't been this accessible in a while, so I began testing voltage and resistance values, starting with the color tubes V10 and V27-V35.

Most readings looked right on the money, which is nice to see. I did find somewhat low voltages on the grids of the chroma sync phase detectors, V27 and V28. The schematic calls for balanced voltages around +20vdc and -20vdc. I measured about 12 volts where 20 is expected, not a huge discrepancy, but worth investigating.

I try not to get too obsessive about such measurements. For one thing, the RCA and Sams schematics give slightly different voltages just about everywhere. If the measured value is within a small percentage of either specified value, I count it good. If it's exactly on target, that's even better, of course. The ultimate question is how the TV performs. Some circuits have more latitude than others, and a set can look and sound great even if not every measurement matches the textbook.

A bigger anomaly popped up on V20, the horizontal AFC/oscillator. The schematic calls for about +4vdc on the cathode, but I measured about -11vdc. All of the resistances for that tube are correct (in fact, Len replaced all of the resistors), so the cause is not obvious.

Despite the odd voltage reading, the horizontal hold on this set has always been very stable, and the waveform at the grid of the horizontal output tube looks peachy. Go figure.

I'll investigate this further, nevertheless.

Vertical Dynamic Convergence Alignment

During this phase, I also checked a number of settings mentioned in the Sams manual. One that I hadn't checked earlier is the vertical dynamic convergence alignment. The procedure is pretty simple. Here is the waveform at pin 1 of V29 before and after adjusting the vertical convergence amplitude control and the vertical convergence shape control. The waveform isn't quite as symmetrical as the model, but now it's a clean trace instead of a bunch of fuzz.

Next, I checked the waveform seen on pin 6 of the CRT. This is an output of the vertical dynamic convergence transformer, which I replaced at the beginning of this project. This showed the same waveform modulated by the horizontal convergence waveform—basically what's wanted, in short.

Horizontal Dynamic Convergence Alignment

Next, I adjusted the horizontal dynamic convergence. For this, you simply tape the oscilloscope probe along the lead for pin 13 of the CRT. The loose coupling is sufficient to pick up the signal through the high-voltage insulation.

The adjustment involves two controls on the front panel, as well as the little transformer that we noticed earlier when reassembling things in the high voltage cage. When you're finished, you want the phase adjusted so that the sync pulse—the little jagged part—is centered at the top of the waveform. At this stage, we're pretty close.

Applying Vacseal to the 15GP22 Picture Tube

In between other tasks, I'm applying Vacseal vacuum sealant to my 15GP22 picture tubes, starting with the second-best one that we've been using.

The little 3-ounce bottle is dwarfed by the special packing for hazardous material. The cap of the bottle was sealed with super-sticky tape. The bottle was also sealed in a small plastic bag and then placed in a large bag of absorbent padding material. In turn, that was held in the inner cardboard container.

The recipe of Vacseal is a trade secret, but from the materials safety data sheet, it looks like a blend of polymers in some very volatile liquids. Primary customers for this product seem to be dudes using fancy lab equipment or building gizmos for outer space.

I'll begin with the tube face down, applying liquid Vacseal to the glass-to-metal joint between the glass bell and the steel ultor ring. This seam is known to be the main source of leaks. Another potential source is the ring's perimeter weld, so I'll coat the entire flange. Here's the 15GP22 diagram again, for reference.

There's no point being stingy with this potion. The bottle probably holds enough to treat two dozen 15GP22s, and I suspect its shelf life is not very long, once opened. Vacseal initially dries shiny with a slightly tacky feel. I'm working my way slowly around the perimeter. You can see a couple of spots that I haven't covered yet.

I allowed that application to dry overnight and then turned the tube on its side to seal the seam and flange facing the front.

The Vacseal website mentions curing at temperatures as high as 500° Fahrenheit, noting that room-temperature curing could take days or weeks. These tubes are too big for a kitchen oven, so I set up heat lamps and turned the tube regularly. After about a week, turning the tube every several hours, I sealed the second tube and set it up to cook.

If you're ambitious and impatient, I imagine you could build a "bake box" and heat it with forced air. The product literature gives no details about curing, so perhaps it's not that critical and the stuff would cure eventually, no matter what you do. After a week of baking and a couple more weeks sitting around, the Vacseal on both tubes feels cured, not at all tacky to the touch.

To Vacseal Or Not?

It's hard to know whether Vacseal will do any good. If my tubes retain vacuum indefinitely, who knows, perhaps they would have done so, anyway. If one develops a leak, the question would be exactly where that occurs, which can only be answered by expensive, specialized equipment.

Other possible points of entry are at the neck end, where the pin wires exit the glass stem and there is a copper pinch-off at the stem terminus. I'll put Vacseal on the pinch-off, which is accessible through a little hole in the base. The wires are out of reach unless I unsolder the pins and remove the tube base, which might do more harm than good.

Given the scarcity of 15GP22s, I think Vacseal is worth trying. The current price for a proven 15GP22 starts around $2,000, if you can find one at all. Investing $160 for my pair of tubes seems reasonable.

Renewing the Conductive CRT Coating

As in many picture tubes, the bell of the 15GP22 has a conductive coating that serves as a filter capacitor in the high voltage circuit. Notice the area marked "external conductive coating" in the data sheet:

The conductive layer makes contact with the heavy mu-metal shield surrounding the bell, which is grounded in turn through a flat spring rising up from the chassis. This coating is often called by the trade name Aquadag. Aerodag is another trade name. The 'dag on this tube is a little worn, so I'll refresh it.

Len had already recoated the other 15GP22 and he included a mostly-full aerosol can of Aerodag G in the stuff he sold me. In contrast to the exotic Vacseal, this stuff is simple: graphite powder in a binding agent.

You don't want to coat the entire bell since, as the diagram shows, there is a band of insulating material between the aquadag and the conductive ultor ring. Spraying the whole bell might short-circuit the 20,000-volt anode to the chassis—a Very Bad Thing!

Here's the tube all masked off, and a second shot after I'm done spraying. As with many painting jobs, preparation takes far longer than application. Aerodag G dries in seconds.

Now that both 15GP22s have fresh Aquadag and Vacseal, they're as fit as I can make them. Since I already know how Number Two behaves, I'll try Number One next.

Replacing a Missing IF Shield

While cleaning the chassis, I noticed that I'm missing a shield that covers the fifth IF amplifier tube and transformers. I'll make a replacement from thin aluminum.

The shield will need an angled top and openings for two control shafts. Like many cardboard boxes, it can be made of a single sheet cut and folded. By a lucky coincidence, the aluminum piece from the hobby store was exactly the right width. I ended up trimming a couple of edges, but this is the basic pattern. The little flat plate underneath was easier to make.

The shield encloses a tube, which will get hot. Perhaps I'll add a couple of little holes for ventilation.

That brings us to early July, 2010. I'll resume the story of the electronics after describing work on the cabinet.

Cabinet Restoration

RCA named this cabinet style the Merrill. It's an understated design, similar to other RCA cabinets of the mid-1950s.

In all of my projects, I work on the electronics first. If you can't get a radio or television working properly, there's no point in making its cabinet look beautiful. I value working radios and TVs highly, but there's no place in this house for dead "shelf queens," no matter how pretty.

With the TV functioning well, I turned my attention to the cabinet, starting with the front parts that were missing in the previous photo.

Installing the Pencil Box Control Cover

In the parts boxes that I inherited, I found two pencil box covers for the front controls. They look identical when closed, but only one of them belongs to a CT-100. The second photo unmasks the lower one as an impostor.

Glance back at the previous control diagram, and you'll see that the holes and captions in the upper box match the CT-100 controls, but those in the lower box don't. The impostor must be from another RCA of about the same vintage. I don't know why Len got it, since both boxes are in mint condition apart from a little dust, and this was his only vintage TV.

The cover had extra holes for two screws, but nothing on the face of the chassis matched those holes, so I wondered how it could be attached. Pete Deksnis sent these photos, revealing the answer.

The cover does take two screws, which go into vertical metal braces attached to the inside of the cabinet. My cabinet is missing those braces, but maybe they're floating around in one of the parts bins. Time for another treasure hunt!

I found the braces in a box of screws and nuts. When I placed them behind the pencil box cover and the wooden cover panel, another little mystery was solved. On the back of the wooden panel are two round spring connectors that snap into holes in the braces. These braces hold both covers to the front of the TV. The pencil box lip overlaps the wooden cover, so neither can be removed until you remove the pencil box screws.

Simple enough. I'll need to remove the chassis to install the braces, but that will be happening soon, anyway.

If you look back at the second photo from Pete, you'll see a little spring that I haven't found yet. Its straight portion is clipped on a tab on the chassis and makes contact with two (or is it three?) control shafts. One end curls into a cone and projects forward to about the depth of one control knob.

Pete believes the spring's purpose is to make a ground connection to those control shafts and the pencil box cover. This may be a bit of over-engineering on RCA's part, but presumably someone considered it important. If I can't find the spring, I'll make a substitute from thin wire.

Cabinet Serial Numbers

In the previous photo, you might have noticed the number 429 stamped into the wooden cover piece. All three cabinet pieces—main cabinet, top lid, and front cover—have a serial number.

Just as with cars, the serial numbers show that the parts are an original matched set.

You can still read the number on the top lid even though Len sanded the back edge bare, presumably in preparation for refinishing. I don't know why he did this. Sanding to bare wood is not necessary unless a piece is severely damaged. I'll use gentler methods to touch up the rest of the cabinet and then finish this surface to match.

There's no particular relationship between chassis serial numbers and cabinet serial numbers. Chassis and cabinets were made on separate assembly lines, and chassis number such-and-such would be put into whatever cabinet came next off that line.

Making Two Grille Pieces

The wooden grille in front of the speaker board is missing a couple of little vertical parts. They were present in the photo that Len originally sent me, but disappeared somewhere along the line. No worries, I'll make new parts and finish them to match.

I don't normally do woodworking on top of the Merrill cabinet, but the panther lamp provides a nice gentle heat to make sure the oil stain has fully dried.

I gave the parts a couple of lacquer coats and tried out the fit.

I'll lightly glue the grille back in place. The colors can be touched up, if needed, when I tackle the main cabinet.

Speaker Board

The speaker is in perfect condition. The grille cloth, like most, has faded where exposed to light, but that's no big deal. The darker parts will be hidden when it's reinstalled.

It looks like the grille cloth came loose in the center and sagged a little, possibly when the speaker board was removed. It will be easy to straighten it and secure the edge with a few staples.

The speaker board attaches to the cabinet with twelve wood screws. This job would be easier if I waited until the cabinet is empty and lying on its face, but I'm eager to put the cabinet together, at least temporarily.

Channel Selector Shield

Pete also identified another mystery part that I found in the bins. I combined front and back views in this image.

It's a light shield for the channel indicator, yet another part that goes inside the front of the cabinet. A pilot lamp shines through its small round hole to illuminate the selected channel. We'll look at the channel indicator in a moment.

Knobs

I found knobs in a couple of different boxes, including a few duplicates.

The small set of knobs is for pencil box controls. The two big dark ones in the middle are for side controls: focus and dynamic convergence. The big ones on the right are for channel selection and fine tuning. On the left are brightness and the brass power/volume knob.

Some of these are duplicates. There's one good channel selector and another with a broken shaft; one good fine tuner and another with a broken collar that was awkwardly repaired with super-glue and white plastic tubing; one good brightness knob and another that was cracked and dabbed with some kind of glue; and one good power/volume knob and another missing the D-shaped metal bit.

Not sure what to make of all this. Possibly some original knobs were broken and Len found good replacements somewhere. In any case, now I have a full set of knobs in fine shape.

The channel indicator has two pieces that snap together. Between them is placed the paper tab set for whatever channels your CT-100 is set up to receive. The buyer of a new CT-100 got a sheet with numbered tabs for all of the UHF stations (14-83). If you didn't insert any new UHF tabs, your indicator would say "UHF" in four positions.

When setting up the TV, the dealer would pull off tabs for local UHF stations and insert them in the indicator. My set had tabs for VHF channels 2-13, plus UHF channels 14, 16, and 18, so presumably this TV's first home town had those three UHF stations

See the section VHF-UHF Tuner in my design article for more information about this tuner.

Cabinet Complete At Last!

After a lot of fussing, I managed to reattach the speaker board from underneath. With the correct knobs and front covers in place, and a color picture on the screen, this CT-100 looks like it should—for the first time in many years!

This is the first time I've been able to judge the CT-100's audio quality. As you might expect, it's greatly improved with the speaker in the cabinet rather than lying bare next to the chassis. The sound is clear and strong, with no shortage of bass.

It wasn't really necessary to put everything together, but this television hadn't worked in decades and had spent the last several years in pieces. It was important to me to see it working, at least for a couple of days, before taking it all apart for further restoration. After stressing over this project for months, I was able to relax and watch a couple of movies on it!

Now the chassis goes back onto the workbench for checking and adjustments; the CRT comes out for Vacseal and Aquadag; and the cabinet will be brought to a refinisher. When I put everything back together, I won't forget goodies like the pilot lamp shield that turned up in my treasure hunts.

Cabinet Disassembly

Today, I empty the cabinet and bring it to Michael for refinishing. After removing the chassis, yoke, lid, knobs, and front covers, I gently lowered the cabinet face-down on the carpet. I built this dolly specifically to fit this cabinet. While the cabinet's away, I'll give it some finish, too.

The speaker board has been removed and I'm ready to tackle the CRT. The four retaining rods slip off after you loosen the knurled brass nuts.

Then you lift off the circular retaining ring and mu-metal shield, loosen the two plastic retaining pads, and lift the CRT up and out, grasping it from the bottom.

Returning to this photo, you can see the hardware for removing the CRT faceplate and safety glass. Eight clips slide out to release the four metal bezel parts from the front. Twelve screws secure the faceplate to the cabinet. You can leave the field neutralization coil attached to the faceplate by its four clip nuts.

Notice the position of the neutralization coil lead (upper left) and the little copper ground braid (lower right). You'll want to put these in the right place when reassembling. The RCA service manual in my design article gives complete CRT installation directions.

Needless to say, you want to keep track of all this stuff!

Here are views of the faceplate and safety glass assembly from back and front. The CRT support has a rubber pad along the bottom and a gasket to keep the tube's face from hitting the glass. These shots also show the slots where the bezel tabs go through the faceplate.

The steel bottom panel in the chassis compartment is secured by four screws in the back and two in the front. After I removed that panel, here are the three cabinet parts—body, lid, and front control cover—ready for transport.

Replacing Decals

The original legends by the front knobs were too worn to be preserved during refinishing. Michael took photos of them next to a ruler to show the scale and then I sent the artwork to a decal maker.

A week later, I got a sheet with 24 sets of reproduction decals, done with gold metallic ink. I don't need two dozen sets, but ink is cheap, so the decal maker fills up the sheet. If we mess up a decal or two during application, no problem, there are plenty of spares.

Back Cover With Some Local History

My CT-100's back cover disappeared somehow over the years. While waiting for the cabinet to be finished, I happened across a replacement. Here's the back as found:

Yes, it's scruffy and the cord is shot. The seller had two CT-100 covers. The first was in much better condition, but I chose this one. Here's why:

This back reportedly comes from a CT-100 used at KING television in Seattle. A Video/Air switch was added in back to flip between signals, which is consistent with use as a broadcast monitor or possibly a display set in the lobby. I have no proof for that, but I like the idea of having some local history. When I spiff up the back, I'll preserve these labels.

John Folsom noted that many CT-100s were converted for such use at TV stations and he provided this 1956 article from RCA Broadcast News describing the conversion. The CT-100 that Pete Deksnis demonstrated at the 2010 ETF convention used this video preamplifier circuit.

Cabinet Refinished!

Today I picked up the refinished cabinet, which looks gorgeous. The work was done by Michael Mueller, who also restored my DuMont RA-103 and RCA CTC-11 cabinets.

The color exactly matches the unfaded area of original finish that was uncovered by removing the faceplate and bezel. It's slightly red, but not garish. The finish is deep and vibrant. The gold trim paint was carefully matched to the brass bezel color, and the decals look perfect.

There's nothing significant left to do on the chassis. Soon, I should be ready to put everything back together.

Reassembling the CT-100

It's mid-July, 2010, and now I have a complete CT-100 television, ready for assembly. Phil's TV factory is open for business!

This time, I'll include parts like this pilot lamp shield, which turned up in a parts box. The plate goes in front and the spring holds it from behind. This creates a little spotlight to illuminate only the selected channel number. If you leave off the shield, more than one number lights up.

The cover plate for the Convergence and Focus controls goes behind a recess in the cabinet's left side. It's there for safety as well as cosmetics. Those control shafts protrude from the high voltage cage and we want fingers kept well out of that area.

Next, I installed the bottom plate for the chassis, as well as the two brackets that hold the front wooden control cover from the inside—two more little parts that appeared during a treasure hunt. Now the wooden cover will be held in securely.

The big front faceplate holds the picture tube and provides the safety glass and decorative mask. Notice the slots around the edges for the brass bezel.

The bezel is made of four interlocking parts. The slotted tabs go through slots in the faceplate and are held from behind with clips. You'll earn bonus points if you can identify the TV on which I placed these parts. It's fun to watch one vintage TV while working on another.

The clip with the copper tail will connect to the chassis after that's in place. It grounds the metal bezel, preventing any shocks.

After installing the bezel, I moved to a carpeted room and laid the cabinet on its face. The speaker board is in place, ready to be screwed down.

Now I'll try out my "best" 15GP22 picture tube. It looked slightly stronger than the other tube on my CRT tester. It's also the one that Len had prepared to use, although he never got as far as installing it.

The tube's serial number is smudgy and hard to read: LB 3825 or possibly 3625. The pins are somewhat grungy. I'll clean them before installing the tube.

Treading familiar ground, I snapped the anode lead onto the ultor ring, put on the heavy black gasket, lowered the CRT into its support, and stationed the corner braces. The two braces in my hand will steady the top of the bell. It would have been easier to loosely position them beforehand, but I can still slip them in.

Next, we see the CRT in the cabinet. The installation directions, referenced earlier, show you how to set the tube so that its blue gun will point straight up. They also show where to clip the anode lead to the ultor ring. After initially positioning the CRT, I tilted the cabinet up far enough to peek at the face, and saw that the tube's mask wasn't quite aligned with the cabinet's mask. After a quick adjustment, I thought I was done with this step.

Close, but no cigar! When I later installed the chassis, I found that I hadn't fastened the anode lead correctly. The tip of the male lead reached the connector from the high voltage cage, but it couldn't seat fully.

When I powered up the TV, I was treated to a glorious light show: 20,000 volts arcing across the gap between the lead ends inside the female's insulator. I took this photo after seeing the arcs and discovering that I couldn't stretch the male lead far enough.

Grumbling, I removed the chassis, lowered the cabinet back onto its face, loosened the CRT, repositioned the anode lead, and put everything back.

First Picture With New Picture Tube

I powered up the television with questions running through my head. Is this picture tube really good? The tester showed that it had emission, but it's another thing to apply high voltage. Assuming the tube is OK, how will the picture compare to the last go-round? It is about two months since I played this TV with a CRT. While I hadn't made major changes, I had tweaked and adjusted various things during that time.

At first glance, the TV had a great monochrome image, but little color, almost as if showing the luminance portion of the video signal minus the chrominance.

After some quick adjustments, I had lots of color—in the wrong places! Bands of red, green, and blue washed down over the screen, a pattern I had seen before (see Color Relapse). With the contrast turned up full, the Wizard looked psychedelic.

Color Sync Problem Recurs

I guessed that the problem involved color synchronization. Color decoding demands precise timing, and, as with horizontal and vertical scanning, the TV generates an internal frequency that must sync with the external frequency from the video signal. When the internal and external frequencies are not locked together, you may see such strange patterns as these rolling color waves.

I began by swapping the chroma reference oscillator tube (V29) with another 6AN8 tube. I have learned from black and white TVs that two tubes may test with about the same emission, yet work differently as oscillators.

After I swapped the tubes, the color bands rolled faster, suggesting that the chroma oscillator's frequency had changed. Not all tubes are equal when it comes to oscillation. To further test my theory, I adjusted the chroma reference oscillator transformer. The bands quit rolling and the colors locked into place.

Setting this oscillator "by eyeball" is definitely not the recommended method, but it allowed me to identify the problem. Now I can break out the equipment and approach it more systematically.

On the bright side, this picture tube looks strong and the TV is basically operational. To address the color sync problem, I'll take out the chassis and work through the color AFC alignment procedure.

What's our Video Input?

Thus began a round of measuring and adjusting, in hopes of licking the color sync problem. I exchanged some email with John Folsom, who asked a number of helpful questions, including how the input to my TV's video amplifier compared to the output from the pattern generator. This would tell me what sort of signal the tuner and IF stages of the TV were delivering to the color section.

It's possible to do this with the chassis in the cabinet. Here, I have connected one oscilloscope probe to the grid pin of V9, the 6CL6 1st video amplifier tube. The blue tube extender exposes the tube pins and lets you play the TV normally. The scope's other probe is connected directly to the generator's video output.

The bottom trace shows the video output from the generator, which is set to a standard eight-bar color pattern. The upper trace shows my TV's video signal at the video amp input, after the signal has traveled through the tuner, RF, and IF stages.

My TV's signal looks messier than the pure generator output, but let's not be too harsh. This set is 56 years old, and for all I know, it has never been aligned since leaving the factory. This signal should be good enough for the TV to render reasonable color bars if the color section works correctly.

Color AFC and "Killing the Color Killer"

I took out the chassis and worked through the Sams manual's procedure for color AFC alignment. If successful, this should lock the signals.

John also mentioned a harmless method for temporarily disabling the color killer. My pattern generator's signal seems stronger than the signal from a DVD player. If the color burst signal is too weak, the color killer could mistake this for a black and white broadcast and turn off the color. By temporarily killing the killer, I can eliminate that possibility from the equation.

After completing those two tasks, I reinstalled the chassis and took another peek. The color bars looked decent:

When I switched to a "real" signal, such as a DVD or a broadcast from my in-home transmitter, we were back to square one. The picture was almost monochrome, although I could see faint bands of mostly green and blue racing down the screen, a familiar and frustrating pattern.

I took a few more measurements while the set was still in one piece. The color AFC alignment instructions tell you to check your burst signal at pins 1/2 of the chroma sync phase detector (V27 in Sams, V130 in the RCA schematic). Here's what I saw:

It's tantalizing to be close, yet not have a solution. The colors came in loud and clear when I used the pattern generator but largely went away when I used any other source. Here's another DVD image from when I had adjusted things "by eyeball."

The snapshot doesn't look bad, but I could only reach that state with difficulty and the TV was very unstable. The picture quickly broke down if I touched fine tuning or various other controls.

I'm still convinced this that is a color sync problem, possibly related to signal strength. Tweaking obvious things such as the AGC control doesn't help. I also tried a variety of sources, so I know the problem's not limited to using one particular DVD player.

Advice From VideoKarma

I mentioned this problem in the VideoKarma color TV forum and some folks advised substituting a new color oscillator crystal in case the old one had changed frequency with age. Here's the plug-in crystal from my CT-100, bearing a date stamp for April, 1954.

Forum members advised that a type NTE358 crystal would work, so I ordered one.

Pete Deksnis mentioned that more than one CT-100 had been known to suffer this problem due to IF (intermediate-frequency) problems, which were sometimes cured by aggressive cleaning of the IF tube pins and sockets. I added "IF cleaning" to my list of future tasks.

Sending Bars Via the Agile Modulator

Meanwhile, since my generator's output seemed stronger than other sources, I tried a quick experiment, hooking up the pattern generator to the agile modulator that I use for home broadcasting. The first photo shows a modern TV in my office receiving the color bar signal through a rabbit ear antenna. The second shows the CT-100 in another room receiving the same signal through rabbit ears.

The result was like the signals from all sources apart from the generator. The picture was highly unstable, full of waves, and prone to break up horizontally and in other ways.

Switching the modulator to a live program from our satellite receiver produced no surprises. Most of the time, the picture looked similar to the first photo, only with rapidly changing (and wrong) colors. With great patience, I could get the colors to lock and display a reasonable color picture, as in the second and third shots. That's not new information. I already knew the TV could make nice pictures. The challenge is to achieve color lock and a stable picture.

What to conclude? The sync problem may be related to signal strength. Color bars from this generator are strong and easy to tune when the signal goes straight from generator to TV. The bars are weak and almost impossible to tune when received at a lower level via the agile modulator.

The strength of the signal received at the TV's video amplifier depends on various things, including front-end alignment. Perhaps Pete's suggestion was correct. I'll do more tests while waiting for a new crystal to arrive.

Making a Test Pattern DVD

John had suggested making a test pattern DVD, and at this time I found a website with downloads for making a free one. Due to inconsistencies among DVD formats, burners, and players, there's no guarantee that the resulting DVD will work on a given player. The one that I created on my laptop will play on only three of the five players in our house. But one working DVD/player combination is all that I need. (If this free solution doesn't work for you, various companies sell test pattern DVDs on the Internet.)

By using the DVD, I can eliminate my pattern generator as an issue. At this stage, I was able to display good looking bars with the generator, but lousy (and unstable!) images from other sources. If I can align the color AFC circuit or otherwise solve the sync problem using the signal from my DVD player, presumably things will look normal when I play movies, too. My goal is to watch movies and TV programs, after all, not generator patterns!

Substituting a Newer Crystal

While searching for something else in my workshop, I ran across not one, but two old color crystals. One had a date code from the 1960s and the other was dated 1977. Choosing the newer crystal, I beefed up its legs and bent one to match the socket for the original crystal.

This was my first opportunity to use the test pattern DVD. Here's one of the menu screens showing color bars among other patterns.

The black and white portions look reasonable, but the color bars show the telltale Star Spangled Banner waves indicating bad color sync. Simply popping in a different crystal wasn't a miracle cure, although I could tell from the speed of the waves that the frequency had indeed changed a bit, just as it had done when I swapped tubes in the chroma reference oscillator socket.

Industrial Strength Socket Cleaning

When the color sync problem originally came up, it seemed to go away, at least for a while, when I cleaned and reseated the color tubes. Although I had cleaned and tested tubes when beginning this project, I hadn't paid special attention to the IF tubes, as Pete suggested. His method of cleaning involves a little vial of DeOxit D100L and a tiny stainless steel brush, which I ordered directly from CAIG Labs. Here we are partway through the process.

While I had each tube out, I cleaned its pins as well as every hole in the socket, and I also put it on my tube tester. This was done for every tube, not only those in the IF chain. The CT-100 has 36 tubes not counting the CRT. Most tubes have eight or nine pins, so for each tube that means cleaning 16 or 18 little things, times 36 equals . . . well, you do the math. I checked off each tube on the chassis diagram when it had been cleaned and tested.

After each tube had been cleaned and tested, I set it upside down on the chassis as a reminder that it was finished. I also checked off each tube on the chassis diagram.

To the right of the CAIG bottle in the previous photo, you can also see the new NTE358 color crystal, which had just arrived in the mail. I had high hopes that the new crystal would solve the sync problem.

At this time, I also cleaned every contact in the CT-100's turret tuner. Although the tuner didn't look especially dirty, I had learned in other TV projects that a funky tuner can significantly degrade the incoming signal.

Substituting a New NTE358 Crystal

Like the circa 1977 crystal, the new NTE358 crystal's legs needed slight adaptation to fit the old socket. I bent the legs to fit, wrapped very fine wire around them, and then dressed them with solder.

After plugging in the new crystal, I took another stab at the color AFC alignment procedure. In this first step, I'm using the scope to make a calibrated voltage measurement.

Reinstalling the chassis, I saw mixed results. No matter how I adjusted the controls, I saw the same horizontal bands in the color bar pattern. In this photo, I'm using the pattern generator.

This photo is a textbook example of poor color synchronization, and one VideoKarma member mentioned that the number of bands gives a rough indication of how far off frequency the oscillator is. The entire screen is scanned 60 times a second (at 60Hz). Ten bands are visible on the screen and perhaps two more are off the screen during the vertical retrace interval. Multiplying 60 times either 10 or 12, we can say that the oscillator is roughly 600-720Hz off the ideal frequency

On the positive side, the picture seemed stronger and more stable than before. Perhaps all of that cleaning helped! Previously, the horizontal bands raced down the screen. Now, at least I was able to get them to lock in place.

The same defect appeared when playing a DVD movie. In this photo, I turned the brightness and contrast way down to show how stable and well defined the bands were.

It's generally a bad idea to make alignment adjustments "by eyeball," but since the color AFC alignment was obviously bogus, there was no risk in trying an adjustment or two while looking at the screen. By turning the reactance transformer adjuster (A40 in the Sams instructions), I could decrease the number of bands but not eliminate them. Here, I am still using the pattern generator. The number of bands has decreased from ten to five.

If the oscillator frequency were within range, I should be able to continue turning that adjuster until the number of bands decreased to zero. At that point, the TV would lock on the right frequency and colors would look normal. Before I reached that point, however, the adjuster hit the end of its travel. Some problem was keeping us too far out of range.

To eliminate the generator (and its signal strength) from the equation, I switched to the test pattern DVD, which gave basically the same result.

The number of bands is slightly different, but we're still out of range. Curses!

Replacing the Original Crystal

This was frustrating. I had installed a new crystal and repeated the color AFC alignment procedure. Perhaps I had made a mistake during alignment or there was an unknown problem in another circuit that affects the oscillator output.

Before quitting for the day, I decided to pop the original crystal back in the TV, mainly out of curiosity. The effect was dramatic. After making some quick adjustments, I got a stable color lock on both the color bar pattern and a DVD movie.

To make sure that the improvement wasn't transitory, I put in a DVD of Gone With the Wind and played the movie to its conclusion. The colors remained stable throughout!

Since I'm back to using the original crystal, where did the cure come from? Perhaps Pete was correct, and bad connections in the front end induced IF tilt and degraded the signal (including the burst signal) enough to mess up the sync. Perhaps I did a better job of color AFC alignment the second time. Perhaps the new crystal was defective, or it was damaged by the heat of soldering when I beefed up its legs. Whatever the cause, let's hope that the improvement lasts.

Adding a Video Adapter

In late 2010, I built a video adapter as described in the 1956 RCA Broadcast News article mentioned earlier. It worked, but I became diverted by other projects for a long time.

In 2013, I corresponded with Pete Deksnis, who shared information about an improved adapter that he had devised a few years earlier. I updated my adapter with those basic improvements and wrote an article describing my adapter and an even more elegant one devised by Pete.

Here is that adapter and a snapshot showing its output.

You can read all about it, along with Pete's presentation to an Early Television Foundation conference describing his design, in the article Video Adapter for RCA CT-100 Television.

To Be Continued . . .

Is a CT-100 restoration ever finished? When I tried out my newly modified adapter, I noticed that my set's edge convergence had become sloppy and the colors didn't track well when adjusting the brightness and contrast.

No doubt these are leftovers from all of the twiddling while I solved the color sync problem way back when. Whatever the cause, this TV is due for a full-dress setup, including purity, convergence, and all the rest. I recently moved the CT-100 into my office, where it sits a few feet from my desk, so I'm motivated to bring it back into good fighting trim, ready for frequent use.