A Predictable Predicta

"That's the set TV repair people hate more than any other set," said a NJARC member familiar with the Philco Predicta, "and it always has been."

The Predicta has the most widely admired vintage television design, on the outside, but its electronics have been considered from the time it first came off the assembly line among the worst. Realizing this after accepting one to fix from a collector in New Jersey had me sweating bullets. It turned out, though, the problems I encountered were familiar to other Predicta servicers. After noting the symptoms, and making some painful studies of the chassis, I pulled over to asked for directions.

Symptoms

The owner's complaint was that the picture would not sync, neither vertically nor horizontally. The set showed only a dim raster only an inch tall appeared. Sound was fine, video appeared to be present, but there was no sign of any synchronization.

Diagnosis

Lacking a power transformer in the set or an isolation transformer on the bench meant it was not safe to view any signals with an oscilloscope. The owner loaned me the Riders service notes, which listed the resistance values to ground or B+ to every tube pin. The vertical circuit of this set is a single two-section vacuum tube and its collection of resistors and capacitors. It's a simple multivibrator with the output of the vertical sync integrator connected straight to the plate of the vertical oscillator. A check of the resistances to these tube pins led me to the first defective component - an open vertical linearity potentiometer.

The controls on this set are unusual. On the front of the cabinet are the usual channel and fine-tuning knobs, and on/off/volume and contrast knobs. On the right side are three small knobs: brightness, vertical hold, and horizontal hold. There appear to be no controls inside or outside for vertical height or linearity, but they are in fact hidden inside the external controls. Remove the brightness knob and insert a small screwdriver in its hollow shaft, and you reach the vertical linearity trimmer. Inside the vertical hold control is hidden the vertical size trimmer, and inside the horizontal hold control is - another horizontal hold control! (The internal one is for coarse adjustment.)

The vertical linearity control is a 1K, one watt pot that carries all of the current to the cathode of the vertical output tube. No doubt years of intense heat and power consumption had burned out this item - but where to find an exact replacement? While I searched for a source, I inserted a fixed resistor of ½ the value of the pot so I could continue diagnosing the set.

This temporary fix resulted in a full height picture, but sync was still scrambled. Adjusting the coarse horizontal hold brought a good horizontal lock. Attempts to lock the vertical resulted in a double picture - two half-height squeezed pictures, one on top of the other, with a vertical interval in the middle of the screen. This meant the vertical oscillator was running at half speed, or 30 cycles per second instead of the normal 60.

A complete recap had not been suggested. The assignment was to fix the problem, not overhaul the set for long-term reliability. All those black beauties and wax-dipped paper capacitors were suspect, as were the electrolytics in the vertical section. The task seemed daunting because the sync/video circuit board is soldered to the steel chassis and offers no access to its solder side. Tidy capacitor replacement would have been very labor intensive.

After running the set for a few minutes, I gave the paper caps a "finger test" to see if any became warm. Capacitors have little significant heat-producing resistance, so a capacitor that warms up is almost certainly one with significant electrical leakage. One capacitor in the vertical section, a 1000 volt .05 uf oil-filled black beauty, was warming up so I snipped it off the board. There turned out to be a slick of oil under it. It was not just leaking electricity but oil. Luckily, I had an orange drop of appropriate value in stock, so I soldered it to the leftover pig tails, and the vertical section started acting normal for the first time.

A Parts Disappointment

A search on google for "Predicta Parts" found a fellow near Los Angeles running a Predicta repair and parts business. Email exchanges resulted in a promise to send an old pot from a "junker" for $20 plus Priority Mail shipping. Two weeks and one Postal Money Order later, I had another brightness/vertical linearity pot in my hands, but a check with an ohmmeter revealed it to have precisely the same defect as the bad one. I resolved, with the owner's approval, to implement a kludge. Antique Electronic Supply sold a 1000 ohm 5 watt pot, which I ordered along with another batch of orange-drop caps.

The "horizontal width" control on this set is a pot on a tab that reaches up from the chassis (it's in reality a horizontal drive control). Right next to it was a similar but unused tab perfect for mounting a trimmer pot, so I drilled a hold in it and mounted the new one there, wired it to the vertical circuit, and now we had an internal vertical linearity adjustment working, ending up looking almost factory-made!

Linearity Out Of Range

Vertical linearity was not in the range of the control, so I did a complete recap of the vertical section with no resulting improvement. Then, on a hunch, I substituted the vertical tube, and linearity came into range! Later, I found out that this was another known Predicta problem: Vertical linearity cannot be achieved with certain brands of tubes … even new ones!

Dots at the Top

Now I had an excellent looking test pattern, but when connected to a commercial VHS tape, some white dots of retrace lines appeared near the top of the picture. These were from the digital information broadcasters and editors insert into the vertical interval right after the sync pulse that carries digital time code and other information. Its presence in the Predicta's retrace indicated that vertical blanking was failing to suppress it.

Attached to the vertical deflection coil in this set is an RC network integrated as a single part labeled "retrace suppression" on the schematic. Could paper caps in this item have gone bad? The fellow in L.A. mentioned that he built replacements for these circuits, creating them with discrete components, dipping them in epoxy, and then painting them brown. He offered them for $9 each. I first wanted to find out if it was the source of our problem, so I built a replacement RC network with parts on hand (the schematics showed the internal parts in these). The set performed exactly the same - the original and my handmade replacement dimmed the white spots but did not eliminate them.

A check with the radio-phono Usenet group was helpful. The dots were not in television signals in the 1950's when the set was made. Engineers didn't predict them for their Predictas. The owner and I decided to leave it in its original state, sometimes showing retrace dots, rather than to engineer improvements.

Burn-in Reveals a New Problem

After the set was back together, I ran it every time I watched TV to keep an eye on it for reliability. The set showed a new problem. The picture slowly became darker only minutes after turning it on, until within a half hour, it was too dark to watch. After yanking it open again, I tested the paper caps in the video section by finger, and sure enough, the last coupling capacitor, a black beauty connecting video out to the CRT, was warm. Replacing it with an orange drop resulted in a severely contrasty picture with extremely unstable sync. Now what?

Stupid Predicta AGC Tricks

It turns out that the front panel contrast control is really an AGC (Automatic Gain Control) adjustment. Turning it indirectly changes the size and shape of the sync pulses (and even affects the sound). There is also a "local/fringe" slide switch on the back by the antenna inputs that is in fact a coarse AGC adjustment. It was on "fringe." Switching it to "local" brought the contrast and sync back to normal.

Final Adjustments

For good measure, every paper capacitor that seemed to be warming up was replaced. My experience with this set confirmed the prevailing wisdom that paper caps of that vintage should all be replaced even if they seem OK. Finding and changing only the bad ones can be more time consuming, and the ones that seem good can be considered ticking time bombs ready to ruin the expensive or irreplaceable parts they are wired to.

Horizontal and vertical widths were set, and adjustment of the centering magnets by the deflection yoke resulted in a nice normal picture. A new #44 channel selector lamp from Radio Shack was the final touch. The owner was happy to get his set delivered and working nearly like new, at which time I got to see his incredible collection of blue ribbon antique radios, TVs, Hi Fi's and test equipment. That made it all worthwhile!