Restoration/repair of a 1947 RCA Victor TV set.

[SeanB]

Going to be a lot of carbon composition resistors to change.  Do you use a modern film replacement ( probably massively overrated to keep the size similar) or do you look for old ones which have drifted to a more or less correct value?

[Vgkid]

Good job with the restoration so far. Keep us updated.

[digital]

https://www.eevblog.com/forum/Smileys/default/smiley.gifThank you for sharing the pictures and info.Cheers

[N2IXK]

Going to be a lot of carbon composition resistors to change.  Do you use a modern film replacement ( probably massively overrated to keep the size similar) or do you look for old ones which have drifted to a more or less correct value?

I generally use 5% carbon film for 1/2 Watt resistors, and metal film flameproof types for anything larger. The 1/2 Watt ones I use are about the same size as the old carbon comps.

I keep a small stash of NOS carbon comp resistors around for use above the chassis or where otherwise visible, but other than that I generally go with modern components. I like to keep the appearance original if the back is removed, but don't go as far as restuffing paper capacitors or using carbon comps under the chassis for originality.

Old sets like this use a mixture of 5%, 10%, and 20% tolerance resistors depending on the circuit. I generally go through and replace any resistor that has drifted outside its marked tolerance. The cost difference must have been substantial back when these things were built, as 5% resistors only show up in the critical sync and oscillator circuits.

I generally find about 50% of carbon comp resistors of this age end up outside tolerance.  The worst offenders are resistors larger than 1/2W, and resistance values less than 100 ohms or greater than 1M. Apparently the resistive mixture wasn't all that stable with very high or very low amounts of carbon in the mix.

[N2IXK]

OK, time for another progress report if anyone is still following this project.

Went through the entire set and replaced any resistors outside of their marked tolerance. There were a couple in the suspect circuits, but replacing them did nothing to get the horizontal frequency back down where it should be.  This left only the handful of mica caps in the oscillator and sync circuits, and replacing them one at a time gradually improved things, In particular, C165, a 5 pF 1500V cap used in the feedback loop caused a major improvement.   Tests OK using a simple LCR meter, but I guess it was breaking down under the huge pulses it sees coming off the damper tube.

Altogether, ended up replacing C122, C123, C135, C162, C163, C164, and C165.   Also found that C169 was missing completely, with lead tails on the terminals where it had been clipped out, probably at the same time that the cap was added across the flyback as a previous "repair" for low width. All caps were replaced with modern silver micas of 500V or 1kV rating, eccept for C165, which was replaced with a 3kV ceramic disc, because I didn't have any 1500V micas around. With all caps replaced, I managed to get a pretty decent image, with good width, decent horizontal linearity, and HV reading right where it should be.  Sync locks nice and steady both horizontal and vertical.   At this point, it looks like the sync/sweep/HV and video stages are working at 100%, and I essentially have a video monitor.  See attached photo for screenshot before I tweaked the vertical linearity in. I will be adding an external composite video input jack, so the set will not be dependent only on the long discontinued analog NTSC broadcasts.  Still need to get the tuner and IF strips working at this point, so the set can be used to receive signals from a low power RF modulator for the full "vintage TV experience". Gave the tuner a thorough cleaning with deoxit, but didn't help, with nothing at all coming through on any RF channel.

Next up is to repair the damaged speaker cone, reinstall the speaker, and see if the sound IF strip and detector is working any better than the video IF is. This set is somewhat unusual in that it uses 2 separate IF strips for sound and video, rather than the more common "intercarrier" IF system introduced a few years later.

[SeanB]

Looking a lot better, and getting a good image as well.

[Len]

I like the period-accurate test pattern. 

[N2IXK]

I like the period-accurate test pattern. 

That's off of a DVD that I found a while back. The actual video signal isn't exactly 100% period accurate, as it is actually a color signal, and does include a 3.58 MHz burst pulse, and the scanrates are 15,734 and 29.97, rather than 15,750 and 30.00.  The original monochrome TV scanrates were shifted ever so slightly to avoid intermodulation with the chrominance signals when the NTSC compatible color standard was adopted in the 1950s. The slight frequency shift was well within the sync lock-in range of the existing monochrome sets, so essentially nobody noticed.

At some point, I would like to cobble together a simple board (Raspberry Pi?) to output the Indian head pattern (still frame from a .JPG or whatever), as a period accurate monochrome composite video signal.  A true modern equivalent to the original RCA TK-1 pattern generator that originally produced this pattern.

For those that don't know, broadcast test patterns in the early years were usually produced by a specialized tube called a "monoscope", which was essentially a CRT containing a silkscreen printed image plate rather than a phosphor screen. Much cheaper and easier to operate than using a standard camera focused on a test chart, such as was used for testing the cameras themselves. There are a handful of surviving TK-1 units, but the number of still operational monoscope tubes is very small. More here:

http://pharis-video.com/p2794.htm

[Vgkid]

That looks very good.

[N2IXK]

Got back on the set a few days ago, and made some progress.  Repaired the damaged speaker cone and reinstalled the speaker in place of the temporary filter choke I installed for testing.  Repair technique I used is the Tacky glue/coffee filter method Bob Andersen shows on his YouTube channel:



I had to apply my patches on the front of the cone, as the big field coil on a small speaker makes access through the rear of the basket difficult. 

Also took a few minutes to reinstall the cardboard covers over all the rebuilt filter caps. they get attached using a thin smear of non-acetic cure RTV sealant inside the end of the tube, and get snapped down over the aluminum can "stumps" on the chassis. This hides all the modern caps completely, prevents a shock hazard from the non-grounded cans, and keeps the topside of the chassis looking original.

Removed my temporary video input circuit and reconnected C103 to the video detector tube. Set my signal generator for a channel 3 signal, and set the tuner to channel 3. As before, no signal on the screen, and no sound from the speaker at any setting of the fine tuning besides a low hum at max volume indicating that the audio amp was alive. Tapping a screwdriver blade on the volume control terminals gave a loud 60 Hz hum from the speaker, so the whole audio amp seems to be OK. So both the audio AND video IFs seem to have no signal coming through them.  This would seem to indicate a problem in the tuner itself, before the separate video and sound IF outputs are generated.  Tried swapping the 6J6 converter (mixer) tube, and discovered that by wiggling the tube in the socket, the signal would flash onto the screen.   We have a mechanical problem, not an electronic one.  One of the pins of the tube socket seems to have an intermittent connection. I sprayed cleaner into the socket, and got the set to work pretty well, but the signal periodically cuts out, and thumping the tube or the area around the socket hard brings it back.  Looks like I'm going to have to dig into the tuner itself to get a look at that socket and the soldering under it.  Unfortunately, working on old time VHF tuners is pretty intricate stuff. the placement of everything affects the alignment, and access to the tube socket terminals may involve a LOT of disassembly on this particular set, as it uses a stack of wafer switches, rather than a single removable turret drum with the coils like many later designs.

I DO have a complete donor chassis, however, with a nice clean tuner assembly.  Perhaps a transplant might be the way to go here?  Will have to look into it. For now, have the tube wedged in place and working.  Sound and picture both tune in correctly, in the same place on the fine tuning control. The alignment of the IF and tuner seem pretty much spot-on after all these years! A multiburst test pattern gives a video response out to ~3.5 MHz, which is typical for a set like this.  It improves a bit when injecting video after the IF stages and tuner, as expected.  The set is pretty much working, and the screen shots posted here are all from an RF input signal, on VHF channel 3.

Next up will be a modification to make the set more usable in this era now that analog OTA signals are mostly unavailable.  Line level A/V inputs and a "line/air" input switch will be installed to allow the set to be fed from a standard composite signal without needing an outboard RF modulator or a repurposed VCR to act as one.  Will also pull the tuner from the donor set to see what is involved in the tuner swap or repair decision.  Looks like the electronic portion of this restoration is almost complete. 

[N2IXK]

OK, got the video and audio line inputs installed. Drilled 3 holes in the rear chassis apron to accommodate an RCA jack for audio, a BNC jack for video, and a DPDT toggle switch to select between line input or OTA reception.  As the audio circuitry is up at the very front of the chassis, I used shielded cable (RG-174 coax) to make the connections between the audio circuitry and the added input switch.  The cable is dressed neatly into place with the existing wiring, and the shields are connected to chassis ground right at the input jacks.

The video circuitry connects via short lengths of insulated wire and extending C103, run directly to the switch to minimize effects on video quality. A 75 ohm 1/2W resistor is connected directly across the input BNC, to provide a standard video termination.  DaveCAD schematics attached. 

Input circuitry works well,  and allows the set to be used with any DVD player, etc with a composite output. Just have to label the added jacks and switch at some point.

The tuner intermittent is a real cute one, an internal crack in a tube socket contact between the part that grabs the actual tube pin and the solder tab underneath.  Would require the replacement of at least the contact if not the entire socket, and the solder side of the socket is BURIED inside the VHF wiring of the tuner. Hard enough to get an ohmmeter probe on it for testing, much less a soldering iron and pliers.  Looks like a tuner swap is coming up....

Got the donor tuner out pretty easily (about a half dozen electrical connections and a bunch of screws), so I will carefully clean it up, replace any drifted resistors that can be accessed, and get it ready for transplant in the next day or so.

[SeanB]

Can you not just pull the remains out from the top and push the remainder out the back of the socket, then simply make a donor pin with a thin end and solder it to the remainder. It will not be strong, but should survive the tube being inserted a few times, and might make good contact mechanically alone without soldering.

[N2IXK]

The tuner needed to come out of the chassis whether to attempt tube socket repair or to just swap the whole tuner assembly, so I carefully pulled the unit out to get a really good look at things.   Pulling the tuner exposed several more carbon comp resistors, a couple of which were out of tolerance, and digging through the junkbox didn't find any tube sockets of the same style to grab a new contact from. These are a somewhat unusual micanol molded saddle type socket with an unusual insulator shape, presumably designed for low loss at high VHF channels.  I was able to pull the broken contact out from the top of the socket, and tried reinserting it with a small dab of silver-filled conductive epoxy applied, to make contact with the stub left inside the socket.  It worked well enough on the low band channels (2-6) but apparently it becomes lossy on the high band (7-13), as signal strength slowly drops off now at higher channels. Oh well, gave it a shot but decided on a straight tuner swap with the donor unit.  Gave ait a thorough cleaning, replaced a couple carbon comps and installed it into the chassis.  100% success, and the set is now rock-solid on 2-13.  Couldn't test Channel 1 due to limitations of my signal generator, but the set is not likely to EVER see a signal down there...

At this point, I'm planning to let the set cook on a corner of the bench whenever I am working in the shop, and declare it electrically restored and ready to go after it has clocked 12-18 hours without incident.  I will get on the cabinet restoration at some point next year, and will post a few pics of that stuff, but not a lot of details as this isn't a woodworking forum.    I may try temporarily installing the 10BP4 tube somehow (prop it up on a block of foam?), to see how things look with the real 10" CRT.   Don't anticipate any problems there, as the test tube generally does a good job and if anything the contrast ratio will be better and retrace lines less apparent on the 10BP4, due to differences in grid cutoff characteristic.