Philco Model H3412L Predicta Television (1959)

This tabletop Philco Predicta TV was known as the Siesta. With a compact cabinet and signature stand-alone screen, it includes a clock timer that can turn the television on and off. Here's my Siesta after restoration:

Description

In 1959, Philco introduced three 17-inch tabletop Predictas, called the Siesta, Princess, and Debutante. The Siesta, with a clock timer, was the top model. The Princess was the same TV minus the clock. The Debutante, as the name suggests, was an entry-level set; it had a cloth grille and its screen lacked the big gold supporting arms found on other Predictas.

The Siesta was offered in five colors: gold, charcoal brown, vermilion red, beige, and mahogany. Mine is a gold model H3412L, with the 10L43 chassis that was shared with 21-inch Predicta tabletops. A magazine ad from December, 1959 shows it with the Consolette stand:

Like many manufacturers, Philco made up flashy names for their features. In this ad, their printed circuit boards were Perma-Circuits and the pivoting rod antenna—a commonplace item—was dubbed the Pivot-Tenna.

In a footnote, the ad mentions Philco's sponsorship of an upcoming Miss America pageant. My Philco Miss America article has more information about that long-running ad campaign.

The next ad, from January, 1960, touts Philco's new "cool chassis" design, claiming that it provided "43% longer TV life" than older designs.

The coolness of the Philco cool chassis was debatable, but we'll come to that later.

Philco took out two patents for the Predicta tabletop design. The first drawing, for patent 186629, depicts the famous screen. The second, for patent 183779, shows a "Predicta that never was." While the general profile was repeated in the 17-inch and 21-inch tabletops, the knob arrangement was changed before Predictas made it to market.

The Predicta's exterior is flashy, but its electronic design is conventional and rather spartan. It lacks DC restoration and various controls for focus, centering, horizontal drive, and so on, which were found on better quality televisions. Adjusting the focus involves moving a lead from one terminal to another on the printed circuit board—not a user-friendly arrangement.

One welcome sign of quality is a power transformer. Some other Predictas, such as my pedestal 4654 console, use a cheaper series-string power supply. Solid-state rectifiers are used in the low voltage power supply (early versions of the 10L43 chassis use a rectifier tube) and silicon diodes are employed in the horizontal phase comparator.

Like many TVs and radios of the day, the Siesta has several "couplates" (aka networks), an early form of integrated circuit with several capacitors and resistors in a flat package about the size of a matchbook. These will also be examined later.

Tube	Type	Function
S1T	6X8	Oscillator / Mixer
S2T	6X8	RF amplifier
S1	6AM8A	3rd video IF amp/Video det.
S2	6DE6	2nd video IF amplifier
S3	6DE6	1st video IF amplifier
S4	6BQ5	Audio output
S5	6DR7	Vert. oscillator / Vert. output
S6	6CS6	Audio discriminator
S7	6EA8	Sound IF amp. / Noise inv.
S8	6CG7	Horizontal multivibrator
S9	6AW8A	Video output / Sync separator
S10	6DA4	Damper
S11	6DQ6A	Horizontal output
S12	1B3GT	High voltage rectifier
	17DRP4	Picture tube

This Predicta's service manual is found in Sams Set 466, Folder 1. Midway through this project, I also got an original Philco service manual. Below are two pages containing the Philco schematic. To save the files on your computer, right-click on the icon and then choose Save Picture As:

If you restore a Predicta, I recommend using both the Sams and Philco manuals. This article refers to components by their Philco part numbers, which are printed on the circuit boards.

Finding a Predicta Siesta

Although I bought this Predicta in 2012, I first saw it 1998, when reporting on a tour of a fellow collector's premises. That visit was recorded in A Visit to Radio Heaven and the Predicta was seen on a shelf in this photo:

I wasn't particularly interested in TVs at the time, but I never forgot the sight of that one distinctive television amidst hundreds of radios.

Nearly 14 years later, I learned that the collector had passed away and his collection was being sold. I visited his son to see what was available. Off to the side, in a room full of radios, sat the same Predicta, looking just as I remembered.

First Look

We made a bargain and the Predicta followed me home. Here are the first photos of the set in my workshop:

As you can see, the television is in nice cosmetic shape. The grille and gold paintwork are fine, the plastic parts are intact, and the back cover and knobs are present.

After making a few checks, I slowly powered up the TV using my metered variac and was rewarded with a picture!

That's Jodie Foster in Taxi Driver, which happened to be playing on cable at the time. The TV is receiving the signal on its built-in rod antenna, broadcast from my home TV transmitter on the other side of the house.

The height was insufficient and the yoke slightly tilted, but the Predicta was basically functional. I powered it down and set the TV aside for a few months until I finished other projects.

Replacement Picture Tube

The Predicta design requires a picture tube with a very short neck. It's common for picture tubes to be replaced, and when that occurred with Predictas, sometimes a tube with a slightly longer neck was used.

That's what was done with my Siesta. In the previous rear view, you can see a small brown cap on the rear cover. The serviceman cut a little hole in the cover to accommodate the longer neck and added a cap to protect it.

The original 17DRP4 CRT has a 2.68-volt filament. Some of the compatible replacement tubes take a 6.3-volt filament, and when those were installed, a simple modification was made to supply the higher voltage.

Before you test a Predicta picture tube, make sure you know which type of CRT it has. You don't want to damage a 2.68-volt tube by powering it at 6.3 volts.

Restoration

This is my second Predicta television. The first was a 21-inch model 4654 in the pedestal console cabinet, which I restored about a dozen years ago:

My first Predicta restoration was quite a learning experience, which you can read about in that article.

Disassembly

The first step was to remove the chassis and find out what I had. Removal is simple, requiring only a screwdriver, nut driver, and pliers.

There are two approaches to disassembling a tabletop Predicta. If you have a jumbo workbench, you can remove everything from the cabinet, lay the picture tube on a cushion, and then reconnect all of the parts whenever you need to power up the TV.

I chose to remove only the chassis, leaving the tuner, clock, and picture tube in the cabinet. This saves space and reduces the risk of damaging the heavy and awkward picture tube, but you need to reinstall the chassis to test the TV under power.

In the next photo, after removing the back cover, I have unfastened the little rear panel that holds three thumbwheel controls for brightness, vertical hold, and horizontal hold. I also unplugged power cords for the TV and clock, and cables for the yoke, tuner power, and picture tube.

At this stage, you can slide the chassis back a couple of inches to unplug more things.

Reaching inside, over the main circuit board, unplug two speaker leads from the audio output transformer and one wire from the Video output pin on the board.

After removing one screw, you can lift up the HV cage lid, remove the 1B3GT HV rectifier tube, and then unplug the second anode lead for the picture tube. It simply slides into a hole in the 1B3GT socket.

If you have recently played the TV, use a clip lead to ground that HV lead to the chassis before you touch it. This discharges any latent charge held by the picture tube and prevents you from receiving a painful (although not fatal) shock.

Also in the previous view, you can see a green wire coming up from the main printed circuit board and grounding behind a screw on the chassis. Predictas are notorious for bad contacts on the main circuit board, and evidently this was one serviceman's solution. The photo also shows the two speaker leads that were unplugged from the audio output transformer.

Peering into the cabinet at the left, you can see the cloth-covered shielded cable that brings the signal from the tuner to the IF (intermediate frequency) section. Unplug this from its phono style jack.

Finally, reaching in with a nut driver, dismount the power/volume control and gain control after removing two nuts inside the front panel.

In the photo, I numbered the leads and cables that connect the chassis to the rest of the TV:

Notice that the three thumbwheel controls and the dual front control stay with the chassis, connected by long leads. When moving the chassis around the workbench, be careful not to twist or yank those leads.

Next, an overhead view shows the basic layout. At lower left is the power transformer. Next to it is the HV cage, which contains the HV rectifier tube and flyback transformer. At upper left is the IF section with its three tubes. To the right are the main circuit board, a large can capacitor, and the vertical output transformer at upper right.

The main board looked in decent shape, although the area around the horizontal tubes (lower left of the board) was coated with dark oily stuff. Tabletop Predictas create lots of heat in a small space and it's common to see some evidence of overheating near hot tubes.

Turning the chassis over, you can see why the Predicta was cursed by generations of repairmen. Although there are a few cutouts in the chassis frame, much of the main board's underside is inaccessible:

A blue cap from an ink pen was wedged firmly between the board and the chassis. Perhaps it was accidentally swept into that spot when someone reinstalled the board, or a repairman intentionally stuck it there for some reason. Let's hope this is not a caveman "repair" for a problem such as a cracked trace or intermittent ground contact.

How Cool Is The "Cool Chassis?"

When you look closely above and below the Predicta chassis, it's evident that Philco's "cool chassis" tag was a bit of a misnomer. Yes, there are vent holes underneath and on the edges, but the entire perimeter of the main circuit board is crammed full of connecting leads, so there is less through-chassis ventilation than suggested in the dramatic magazine ad.

While the metal of a conventional chassis conducts and distributes heat, the synthetic circuit board material is more of an insulator. It also can be damaged by heat over time.

Philco did a reasonable job of ventilating its tightly-packed cabinet, but it's still far more cramped than the roomy console cabinets of most older TVs. Size alone means that these TVs inevitably run hot. Some present-day Predicta owners install little muffin fans to combat overheating, although I personally find the fan noise distracting.

Cleaning and Inspection

I start every project by cleaning and testing components. Every tube is removed and checked on my tube tester, and then I clean its pins and every hole in its socket.

As often happens, all of the tubes tested good. Since the TV worked, at least marginally, that wasn't a huge surprise.

Don't overlook this little item when cleaning. As the marking indicates, it's a dual diode used in the horizontal comparator. I removed it to make sure its pins are clean, like any other plug-in component.

I inspected everything else under strong light, looking for trouble signs like burned components, broken leads, or blobs of leftover solder from sloppy repairs.

This Predicta looked unmolested, with signs of minor service in the past. On the main board, three capacitors had been replaced by snipping the original leads above the board and soldering new caps to the stubs. Two of the replacement caps were the relatively modern yellow type. Perhaps the prior owner did a minimal make-it-play service within the last 20 years.

I normally prefer a neater replacement method, but during the Predicta's normal service life, "pigtailing" above the board would have been reasonable to address minor problems. Unwiring the main board circuit board merely to make things neater would have added a lot to the repairman's bill.

I'm less enthusiastic about this repair. See how the serviceman piggybacked a new resistor in parallel with the old one, rather than replacing it.

This is a dropping resistor in the noise inverter bias circuit, specified as 22K ohms with a 5% (more precise than usual) tolerance. No doubt the original resistor's value drifted upward with age. Wiring another resistor in parallel would somewhat reduce the resistance, but I doubt whether this quickie combination fell within 5% of the needed value. When I remove the main board for recapping, I'll redo this the right way.

I then cleaned the entire chassis using isopropyl alcohol, paper towels, and a soft brush.

Next, I took out the tuner mechanism and removed its shield, which gives access to the contacts inside.

Most old TV tuners benefit from an internal cleaning to remove grime and oxidation from their contacts, but when doing so, follow the motto, "less is more." Avoid spraying cleaner all over the place and don't use anything abrasive, which can destroy the precious metal plating on contact surfaces. I apply liquid DeOxit with a Q-tip, rubbing the contacts and operating the tuner from time to time.

While the tuner was out, I replaced a paper capacitor on its top. You ordinarily don't replace any resistors or capacitors inside a tuner, however. Its internal parts are extremely reliable and replacing any of them may force a needless and difficult realignment.

Replacing the Speaker

The TV's audio was faint and much distorted, and a quick speaker swap revealed the cause. Connected to another source, the Philco speaker sounded awful, while a test speaker connected to the TV sounded great. I happened to have on hand a new speaker that fit, so I popped it in, reusing the leads that plug into the audio output transformer.

The old speaker looks fine on the outside, but something inside (perhaps the voice coil?) needs help. Maybe on some rainy day I'll see whether it's repairable.

Replacing Electrolytic Capacitors

The Predicta chassis is cramped, so I restuffed all of the electrolytic cans, rather than install new caps outside of them. In many TVs and radios, you can install replacements under the chassis, but this chassis has no "underneath" in the usual sense.

The electrolytics are mounted outside the main circuit board, so you can replace them before or after removing the board.

The E2 electrolytic can was slightly tricky. It contains four caps, but only three of my replacements will fit inside. Here, I have left the 82-mfd cap outside the emptied can:

I drilled holes in the can base for the new capacitor leads and drew a little diagram to help reconnect everything. (At this early stage of the project, I was using Sams part numbers rather than Philco part numbers.) Numbered tags are attached to the disconnected leads.

After restuffing E2, the can is reinstalled and the 82-mfd cap is snuggled next to the can's terminals.

Three new caps fit easily inside the E1 can. In the first photo, the rebuilt unit has been rewired and remounted on the chassis. Then the can is glued back on.

Don't overlook capacitor C41, the AC line filter hidden behind the power terminal. It should be replaced with a modern Class X safety capacitor.

Another slightly hidden cap is C42, the B+ decoupling capacitor for video IF. It is mounted on a terminal strip just outside the main circuit board. The arrow points to my replacement; the original was a large "black beauty" molded paper type.

Servicing the Main Circuit Board

It's possible to replace some components on the main circuit board using the "snip and pigtail" method from above, but a thorough restoration requires removing the board.

As with my other Predicta, I unwrapped the leads from the board lugs, labeled each lead with a tag made of tape, drew a diagram showing all of the connections, and unmounted a few parts, such as the range switch and picture tube cable socket, to facilitate removal.

In the next photo, the board has been removed and placed next to my diagram. Blue tags identify the connecting leads that remain with the chassis.

Later in the project, I bought an original Philco factory service manual, which has two diagrams showing everything that I had drawn by hand, and more. The first one identifies every component and lug on the board. The second page identifies every lead and it also gives voltage and resistance values for tubes on the main chassis.

Removing the board is tedious and it increases the risk of wiring errors and damage to the leads or terminals, so I was determined to avoid pulling it more than once. In addition to paper capacitors, I also replaced mica capacitors in the horizontal and vertical sweep circuits.

I also tested all of the board's resistors and replaced any that measured more than 20% off the specified value. In this I was informed by my earlier Predicta restoration, where I had discovered some dodgy resistors after pulling the board a second time.

I debated whether to replace couplates on the board before reinstalling it. This was a pure judgment call. On one hand, some of these multi-component packages are known to be troublesome, and there's no practical way to test them other than by substitution. On the other hand, the TV had produced a passable picture before restoration, so for all I knew, the couplates might all be good.

Adding Quick Disconnects to the PC Board

While discussing this project in the VideoKarma forum, someone asked whether it would be useful to attach "quick disconnects" to the board leads. In theory, this sounds great. Instead of laboriously unwrapping and unsoldering three dozen leads, you could simply unplug them and pop the board out!

This idea had been raised before, but I couldn't recall anyone trying it, so I thought it was worth investigating.

At a nearby surplus store, I found some Molex connectors that looked usable. After I installed one, however, it seemed larger than I liked:

One connector doesn't take up so much space, but multiplying that connector by three dozen might create a mess.

Next, I got a bunch of Anderson power pole connectors, thinking they might fill the bill. But when I laid several of them near the places where they'd connect, the sight was dismaying.

If each connector was wired with a short lead to its pin on the board, you would end up with lots of leads and connectors above the board. This could really interfere with everyday service. With so much stuff in the way, it would be hard to see a tube, much less replace it. And it would complicate diagnostics like snaking an instrument probe or a lead in to a test point deep inside, to measure something while the TV was turned on.

The best idea that came out of this discussion was to leave all of the board pins bare and solder a female Molex .062 connector on the end of each lead, insulating it with a bit of shrink tubing. Then you could plug each lead directly onto its pin, just as the Siesta's video lead plugs onto a pin.

On reflection, I rejected that method, too. In my opinion, the advantage of being able to quickly pull the board is outweighed by the risk of creating intermittent connections. The Predicta board is prone to flaky connections under the best circumstances. It seemed unwise to surround the board with three dozen connectors that might be nudged loose by accident during routine service, and whose contacts may become corroded or dirty over time.

The last thing I wanted was to make this finicky board even more finicky! I abandoned the quick-disconnect scheme and proceeded to restore my set the old fashioned way.

Loosening the Main Circuit Board

An alternative to adding quick disconnects is to free the main board from its ground lugs and carefully push it away from the chassis, as far as the slack in its connector leads will permit.

This view from below shows the board loosened from the ground lugs and pushed away from the chassis:

This doesn't give access to the entire foil side of the board, but it provides enough elbow room to reach many more components through the access holes. This is how I replaced three of the Predicta's couplates, as explained below.

Unsoldering the ground lugs takes only a couple of minutes. A solder sucker will remove excess solder, and then the lugs slide through their slots when you lift the board.

Servicing the High Voltage Section

After recapping the board, I removed the high voltage cage, cleaned everything with isopropyl alcohol, and inspected the flyback transformer and the socket of the 1B3GT HV rectifier tube.

I also measured and recorded the resistance of the windings of the yoke and flyback transformer, comparing them to the values given in the manuals. As an additional check, I tested these components with my EICO 944 flyback/yoke tester. They all passed with flying colors.

While in the neighborhood, I replaced a couple of resistors. Here's one (R71) on the yoke socket, normally hidden inside the cage:

The black wax coating around the flyback transformer had a few surface cracks but it looked better than many flybacks I have seen before. Resistance checks of its windings gave no cause for alarm and the TV had produced a nice, bright picture, so I left the flyback alone.

I replaced the rectifier socket and cage cover, reinstalled the main PC board, and powered up the TV.

Loss of High Voltage

After I remounted the board, the TV's audio worked, but the high voltage supply fizzled. The screen flickered to life for one or two minutes and then went dark.

I first checked a number of obvious things, substituting tubes, testing voltages in the horizontal and high voltage circuits, and so on. The grid voltage on the 6DQ6 horizontal output tube was deficient—zero volts where you'd expect -45 volts—and the waveform at the grid did not match the model given in the service manual.

Since I had been methodical about replacing components and reinstalling the main board, my first impulse was to suspect that the N7 couplate had failed.

A couplate is a 1950s-style integrated circuit with a few capacitors and resistors in a flat envelope. The N7 couplate processes the signal from the horizontal oscillator tube and delivers it to the grid of the horizontal output tube.

Original factory couplates are no longer available, and I doubt you'd want to use one anyway, since its old resistors and capacitors might degrade with age, just like their larger 1950s counterparts. It's possible to build a substitute, however, and I had done that when restoring my first Predicta. Before going to that trouble, I called on the VideoKarma forum, in case someone had a better idea.

Testing by Horizontal Signal Substitution

That discussion produced several good suggestions, including using my BK 1077B Television Analyst to substitute a drive signal for the signal that should be sent to the grid of the 6DQ6 horizontal output tube via the N7 couplate.

The idea is to zero in on the trouble spot. If the high voltage reappears, then I'll know that the problem had to be somewhere upstream of the 6DQ6 grid. That upstream territory includes the N7 couplate. On the other hand, if HV is still lacking after I provide a known-good signal at the grid, that indicates a problem downstream, such as a bad 6DQ6 socket or bad flyback transformer.

To illustrate, here's a portion of the Sams schematic, whose horizontal section is a little easier to follow than the Philco version (the Sams part number for the couplate is K7).

The red arrow shows where I'll inject the test signal. The blue rectangle roughly outlines the upstream zone; if HV reappears, then we suspect a problem there. The green rectangle includes downstream components; if HV doesn't come back, something must be funky down that way.

I first checked the test signal from the BK 1077B Analyst to make sure it was operational and approximately the right shape. So far, so good:

A label on the PC board showed where to connect the test signal to the grid of the horizontal output tube (the blue in the photo comes from my LED flashlight, not a strange condition in the TV):

The substitute drive signal brought back the high voltage, and with it, the raster. We have a picture again!

(In the previous photo, the image height is deficient. That was also true when I tried the TV before starting restoration, so I deferred the height issue for later consideration.)

In the course of servicing the main board, I had already replaced all of the capacitors and resistors around the 6CG7 horizontal multivibrator and 6DG6 output tubes. This substitution test convinced me that the N7 couplate was at fault, so I proceeded to build a replacement.

Replacing the N7 Horizontal Couplate

Now, I faced a decision: whether to pull the main PC board a second time or try to replace the N7 couplate from above. Turning the chassis over, I could see that the foil-side connection points for the couplate would be reachable if I unsoldered the board's ground lugs and pushed it out as described earlier.

Removing the old couplate was a bit pesky. Its four lugs fit very tightly into the holes in the board, and even after sucking away excess solder, it was impossible to remove the couplate intact, so I cut it into pieces above the board and then removed each lug separately. Here are the remains of the couplate and the parts that will replace it:

The couplate has five terminals and it contains three resistors and three capacitors. Below is its schematic from the Philco manual. The components inside the couplate are bordered with a dashed rectangle.

I cut a piece of perforated project board and marked where its five terminals will be located. (Although the N7 terminals are numbered 1-6, terminal 2 is not used.) Now, I'm trying the new board for size before populating it:

Here's the new N7 couplate under construction. To save space, I used both sides of the perf board.

As often happens, the most convenient physical layout doesn't mirror the logical diagram in the schematic, so a little mental effort is needed to see how the components match the diagram.

After you build a couplate replacement, you could wrap some tape around it or dip it in an insulating material, if you like. I left this one bare for the time being, until I confirmed that it worked.

Installing the new couplate was easier than removing the old one and it worked like a charm. The high voltage and picture returned and the horizontal lock was stable:

Deficient Height

When servicing the board, I had replaced all of the capacitors and resistors around the 6DR7 vertical tube, so I assumed that those new components were good. Now, I made voltage and resistance tests on all of that tube's pins and I checked the values of the potentiometers for height and vertical linearity.

While the chassis is in the cabinet, you might think that much of it is inaccessible for powered-up testing, but that's not entirely true. Yes, the underside of the circuit board is out of reach, but in many places on top of the board, you can sneak in a clip lead or probe for testing.

Extenders make it possible to reach a tube's pins for diagnostics. In this photo, I have plugged the 6DR7 tube into the blue extender and taped a clip lead to pin 3. This pin is normally inaccessible when the tube is plugged in, far back in the interior, but now I can reinsert it in the chassis and bring the lead out to my oscilloscope or multimeter.

Resistance and voltage tests didn't turn up any smoking guns, so I decided to replace the N2 and N3 couplates. These sometimes fail in Predictas, and the only reasonable way to test them is by substitution. Even if replacement doesn't fix the problem, it should help to isolate the cause by removing couplates (with a total of 10 internal components) from the equation.

Replacing the Vertical Stage Couplates

Seen in the schematic, the N3 couplate is the vertical integrator, which processes the signal from the sync separator tube and passes the resulting waveform to the vertical oscillator. The N2 couplate returns feedback from the 6DR7 pin 1 plate to its pin 7 grid.

When restoring my first Predicta, I had replaced its vertical feedback couplate, so building N2 was familiar, although this one used slightly different resistor values.

Again, I found it convenient to put capacitors on one side and resistors on the other.

The N3 couplate is more complex, with three capacitors, two resistors, and a resistor-capacitor combination shown as a 90K resistor (R3) with a .004-mfd capacitor (C2) distributed across its length.

Following advice from my earlier Predicta project, I replaced that component with a 90K resistor (actually, two 47K resistors in series) and two .002-mfd capacitors mounted on either side of it.

You can also build the R3 component with two 47K resistors in series and three .001-mfd capacitors, one connected between the resistors and the other two at their ends.

Testing by Vertical Signal Substitution

With the new couplates in place, the vertical stability was great, but the height was no better than before. Rats! Testing voltages and double-checking my replacements didn't uncover any obvious cause for the problem.

I renewed the discussion on VideoKarma, wondering whether I should suspect the vertical output transformer, since I had replaced just about every other component involved with height.

As before, I got many useful suggestions, which you can read in the full discussion if you're curious.

Taking a cue from my earlier experience with the high voltage issue, I remembered that my BK TV Analyst could also generate a vertical plate drive signal. Perhaps substituting that signal would help isolate the problem.

The Philco manual showed a model waveform for the signal at pin 1 of the 6DR7 tube, and the signal produced by my TV Analyst looked similar:

I pulled out the 6DR7 vertical output tube, thus removing its signal from the circuit, and then injected the vertical plate drive signal from the Analyst at the pin 1 (plate) connection point. The results were ambiguous:

The first photo shows the screen image with the 1077B's Amplitude control set about midway. The height is still deficient. The second shows the screen with the Amplitude control turned up all the way.

Although the second image filled the screen, the lower half of the screen was distorted and the overall linearity was dreadful. As one forum member remarked, perhaps the excessive amplitude merely masked a problem in the output stage or elsewhere.

Since the 1077B also provides a vertical grid drive signal, I tried substituting that signal (at C19). The results were no better than when injecting a plate drive signal.

In short, signal substitution for the vertical stage didn't reveal an obvious culprit, as it had with the horizontal problem.

Testing and More Testing

In the next photo, I have disconnected the vertical linearity potentiometer (VR2) and its bypass capacitor (E1-3) and substituted a fixed resistor and a new 100-mfd cap:

Substituting fixed resistors of different values didn't improve the height, either.

Again, no cure. But this substitution told me that the problem wasn't caused by a defective linearity pot, and that the new electrolytic I had replaced earlier wasn't defective or miswired.

In addition to testing voltages and waveforms on the 6DR7 vertical tube, I used my HP 428B milliammeter to measure its pin 9 cathode current. Here, with the TV's Width control turned fully clockwise, it measures around 38 or 39 milliamps:

I double-checked that reading with an analog multimeter and got the same results. The cathode current was basically OK, approximately 30 milliamps as given in the Sams manual if I dialed the Width control down to an average setting.

Another forum member suggested temporarily shorting out the thermistor (R64) which is connected between the two windings of the vertical yoke.

In previous measurements from the yoke plug, I had determined that the winding resistances looked OK, with the thermistor contributing about 7 ohms of resistance. That was slightly higher than the specified value of 4 ohms (when cold), but it showed that the thermistor was not open or shorted.

Strange things sometimes happen when a component heats up, however. To eliminate the thermistor as a trouble source, I slid the yoke off the picture tube neck, shorted the thermistor with a clip lead to remove it from the circuit, and reinstalled the yoke.

Eliminating the thermistor didn't improve the height, either, so I removed the clip lead and kept plugging away.

Height Problem Solved

While discussing the couplates I had replaced, one forum member asked about capacitor C16, which is attached to pin 3 of the N3 vertical integrator.

That capacitor had been replaced when I recapped the main circuit board. Double-checking it now, I saw that I had used the wrong value, installing a cap labeled 683 (.068 mfd) rather than 682 (.0068). A rookie mistake—Arrgggh!

This illustrates the risk of replacing many parts at once. When restoring a more traditional TV, you can usually power up the set for a quick test after replacing every few parts, so a flub like this is quickly caught and corrected.

That's how I normally work, but you can't restore a Predicta that way, unless you want to uninstall and reinstall the main board a dozen times or more.

VideoKarma member DaveVM recommended a tactic to help avoid such errors: before installing a new cap, check its value on a capacitor tester. This ensures that the new part isn't defective, and more importantly, it's a final safeguard against inadvertently installing the wrong part.

Vertical Linearity and Stability

The previous screen shot looked promising, but a test pattern showed that the vertical linearity was less than ideal.

The upper half of the screen is stretched vertically, making the circle somewhat pear shaped. This is the best I could get when adjusting both the height and linearity controls for the best compromise (they are interactive).

Many old TVs have slight defects in screen geometry, which don't matter a lot in everyday viewing, but it was harder to overlook the lack of vertical stability. Vertical lock was hard to obtain, and when the vertical sync was far off, it blinked to a bright horizontal line (momentary loss of all vertical deflection) every now and then.

After replacing C16, the height was actually a little excessive. I substituted slightly higher values (reducing the height) until I was able to fit the entire image on the screen, with only slight underscan as needed to obtain better linearity.

I also adjusted the value of two other components, particularly R14, which, I had earlier been advised, affects the vertical hold range. These changes improved the stability as well as linearity. Here's a test pattern at this stage:

When working on linearity, don't forget about the centering adjusters, located on the yoke. I found that the vertical centering control had been adjusted way off, perhaps to compensate for some aging component. After I centered that, it was much easier to make the picture right using the height and linearity adjusters.

Horizontal and Audio Alignment

At last, the picture looked satisfactory, so I finished by working through the horizontal and audio alignment procedures, and then played a couple of movies to make sure the television was stable.

This Predicta's audio is surprisingly good. Perhaps the new speaker is better quality than Philco's original issue. This set lacks a tone control, but you can adjust the tone by changing the value of C17, the tone compensating capacitor. Increasing the capacitance cuts the treble, this improving the bass response. You don't want to increase it too much, of course, or the tone becomes muddy.

Speaking of audio alignment, fellow collector Russ Barnett shared some advice about the correct tools and procedure. Transformers T5 and T6 both have concentric cores of different sizes. In some transformers, you would adjust one core from the top and the other from the bottom. With this type, you adjust both cores from the top using different size tools, shown here:

On top is a 3/32" (.09375") Delrin plastic hex tool used for the top cores. The tool on the bottom is also 3/32" on the tip but it then tapers down to about .07". This is used on the bottom cores. You can get such tools from Antique Electronic Supply or Vetco.

These photos show the approximate location of the bottom core in the transformer can:

Russ notes that the hex cores may not be lined up perfectly with each other, so extracting the tool after tweaking the bottom core is a delicate little dance. First you pull up until you feel the tip disengage from the bottom core (about 1/4" to 1/2"). Then you pull up gently until you feel the the tip touch the underside of the upper core. Rotate the tool until the hex tip lines up with the upper core. Then you can pull the tool straight up and out

After adjusting the tone to my taste, I declared the electronics done. I gave the dial parts and knobs a cleaning and polish before putting it all back together.

Final Thoughts

On the whole, this was a pretty straightforward project, not counting the detour caused by installing one part with the wrong value. I got a kick out of making this Predicta operational, after seeing it in a memorable collection years earlier. Perhaps the previous owner would appreciate the effort spent in making it work again.