RCA Model 143 Cathedral Restoration

The RCA Model 143 is a very large 8-tube superhet circuit radio that receives long wave, medium wave, and two short wave bands. The cabinet has a rounded top with shoulders - not really either a tombstone or cathedral style cabinet.

The schematic and a parts list for the radio can be found on-line at Nostalgia Air. Any part number references in the text below reference the model 143 schematic listed as "Schematic with Fidelity Change" (Riders page 5-86).  There are several different versions and a few changes documented in the RCA Service Notes in Riders. 

My antique radio restoration logs

Overview

The radio was purchased on eBay.  Externally it appeared to be all original, complete and in excellent condition. It was listed as not working, and the power cord had been cut off.  All tubes, shields and covers appeared to be in place. There was only minor corrosion on the chassis, but it was VERY dusty.  There was no rodent activity.  This was definitely the most difficult restoration I have even attempted.  The chassis was very cramped, and access to many components extremely difficult.  Partial disassembly was required in order to access some parts for replacement.  The chassis appeared to have been built in layers!  For example, there are quite a few parts that are UNDER the band switch.  These must have been installed before the band switch assembly was installed.  Of course, removing the band switch was not an option! Two large terminal boards contain many parts that have to be changed - one board has parts on both sides.  The boards have to be removed from the chassis in order to replace parts (neatly) and also to access parts underneath or  to provide room to work.

Several transformers were installed using bent-over tabs which are very difficult to straighten without danger of damage to nearby components. One of the capacitor packs was riveted to the chassis. The two capacitor packs and the driver transformer were all potted in tar. 

Previous Servicing

I always attempt to avoid purchasing radios that have been "restored" by collectors or flippers, and am looking for either all original examples or those which have been "lightly serviced" in the distant past by radio service shops.  This radio met that requirement.

Survey

Cleaning

The chassis was very dusty, but not rusty. The dust was then blown off of the chassis with an air compressor. Parts to be rebuilt or repaired were then removed to provide cleaning access: both filter capacitors, the two capacitor packs, the driver transformer, and tuning capacitor and dial assembly (for cleaning and replacement of mounting grommets). The chassis was cleaned using GoJo (white) hand cleaner and 00 steel wool. The chassis was then carefully vacuumed, and any remaining steel wool fragments removed using a magnetized screwdriver and masking tape.  

Restoration Strategy

Since all of the original parts (except C38) were still in place I decided to try to maintain the original chassis appearance to the extent possible. All original capacitors would be rebuilt in their original cases (restuffed), including the original filter capacitors and capacitor packs. I planned on finding an original RCA replacement for C38 in my junk capacitor stock, and restuffing it. I do not replace mica capacitors, but may test them in place if possible (usually this requires disconnecting one end of the capacitor).

Any out of tolerance resistors would be replaced with the same types if available. When I replace a component, I always remove the original part completely from a terminal.  Other good components connected at the terminal are protected from heat using old medical clamps (hemostats).  Excess solder is then removed using a solder sucker in order to expose terminal holes for reattachment of the rebuilt or replaced component.  

Repairs

Driver Transformer T2

Unless this part could be repaired or replaced, restoration of the radio would not be practical. Finding a replacement part would mean finding an RCA 143 parts chassis - which is very unlikely.  And even if found, the same transformer may be defective on the parts chassis. It was initially not known if the primary of the transformer or the reactor L29 was open, since they are in series. I also have not found anyone who knows the correct ratio of the transformer.  The first step was to remove the transformer from the chassis and attempt to remove the transformer and reactor from their case.  As usual with RCA/GE, the transformer assembly was potted in TAR!

Access to the bent tabs holding the transformer case is limited by a terminal board, wiring, and other components. The first step was to remove the terminal board.  All leads and components were disconnected from the 10 terminals. The board was then removed.  One of the nuts retaining one of the board's studs was hidden underneath a covered two-lug terminal strip on the rear chassis.  Several paper capacitor also had to be removed in order to access the transformer tabs.  The tabs were straightened and the transformer removed, with difficulty. As suspected, the transformer and reactor were potted in tar.

The transformer was placed leads up on an old baking tray in a 250 degree oven for about 1 hour - until the tar appeared to be bubbling.  The contents were then dumped onto the pan and heated for another hour in an attempt to remove as much tar as possible from the component.  While still hot, the transformer and reactor were removed from the pool of melted tar.  The components as well as the case were soaked overnight in mineral spirits. The parts were then cleaned as well as possible using the original mineral spirits followed by several washes with clean mineral spirits. The driver transformer, reactor, and various supporting hardware were then retrieved for inspection.  One lead of the transformer was soldered to a lug on the reactor.

The reactor L29 was OK - it measured about 3100 ohms. The primary of the interstage transformer was open.  Having nothing to lose, I did some further cleaning of the transformer and removed all of the E and I laminations for access to the coil. I removed the outer layers of the coil's insulation, exposing the connections from the wire  leads to the windings. But I never found the internal winding lead that was supposed to connect to the primary wire.  And the more I dug, the more damage occurred.  So I eventually gave up on the original transformer.  

I was now facing the task of replacing the original transformer.  But I did not know its turns ratio or other specifications.  It was about the same physical size as a Stancor A-62C or A-63C.  However, the winding DC resistances of the Stancor transformers were very much lower than that of the original transformer (only the secondary of the original was good and could be measured).  I had an A-62C transformer in stock which was good. It has a 1:2CT ratio, and it and the inductor L29 would both fit easily in the original can.  My concern was that the lower DC resistance of the Stancor transformer may imply lower inductance, and thus loss of fidelity.  I have experienced this effect on many other radios of similar vintage, such as the GE S-22, RCA Radiola R-7, and Crosley 124. Another concern using this transformer was that since its primary DC resistance was much lower than the original, it may not work well with the inductor L29 in series.  And also its ratio 1:2CT may not be correct.

One other possibility available was a Halldorson HVAT-2 universal driver transformer.  The HVAT-2 is a large transformer with lots of iron and lots of copper.  It handles either single or push-pull primary and secondary.  Push pull to push-pull  ratios of 1:1, 1:2 and 1:3 are available.  For single ended driver stages, many more ratios are available such as 1:1.2, 1:1.5 and 1:6 by selecting appropriate pairs of primary leads.  The DC resistance of this transformer's primary and secondary were similar to the original transformer.  For example, using a ratio of 1:2CT: 

  Primary Secondary
Original 1835 2270
HVAT-2 1620 2580

With its abundance of iron, it should provide excellent fidelity.  The problem was that if this transformer was used, there would NOT be room in the original case for the reactor L29.  It was not obvious to me what the purpose of L29 was.  Could it be eliminated? Being a large bank wound air-core coil with lots of Litz wire (3100 ohms) it would have some affect on high audio frequencies. I posted this question on Antique Radio Forums. The most likely answer I got back was that the purpose of L29 was to roll off the 10 kHz adjacent channel heterodyne whistle often present during nighttime skywave conditions where distant stations come rolling in that are only 10kHz away from local stations.  That was also my original suspicion.

By using the HVAT-2 transformer and eliminating L29 I could make all the primary leads available for access until the radio is restored, and then pick the turns ratio based on the best sound of the radio.  It is NOT practical to remove and then reinstall the transformer after restoration, as many components and connections have to be undone in order to access the bent-over transformer tabs (which will eventually break off if flexed several times).  So I decided to go with the HVAT-2 transformer and eliminate L29.  The transformer initially was installed with the ratio of 1:2CT.  If it turned out that L29 was really needed, I could install it under the chassis somewhere (outside of the T2 case).  New leads were used, since the originals were very stiff and would likely break if flexed. Not all the original colors used were available in modern cloth covered wire, so some other colors were used.  No tar or rosin was used to seal the transformer in the case - only strips of paper towel to hold it securely (the HVAT-2 is vacuum sealed).  Here is the repaired transformer:

The ends of the unused leads were insulated with shrink tubing and left in place, should a future owner wish to change the ratio from 1:2CT to 1:1, 1:1.2, 1:1.5, 1:3, or 1:6. A chart of available ratios vs. wire colors will be kept with the radio.

Resistors

The radio used older style "dogbone" type resistors.  Thirteen quarter watt, one half watt, and one 1 watt were out of tolerance. All were wound end types (lead wires wrapped around the ends of the carbon rod element and then the wraps soldered together). Two of the quarter watt resistors were in spaghetti sleeving, and thus normal 1/2 watt carbon composition resistors could be used to replace them, since they are not visible. I keep a stock of NOS and used "dogbone" resistors, and buy all I can on eBay and at radio swap meets (when reasonably priced)!  Of course, most of these resistors, even NOS resistors, have also drifted in value and no longer have their marked values.  My solution is to find a replacement resistor of the correct value and size as measured (ignoring the markings), and then repaint it to the needed value codes using enamel hobby paint!  I was able to find suitable replacements for all of the dogbone resistors.  The replacements were repainted using hobby enamel paint to the color codes of the original resistors.  Here are the replacement dogbone resistors (only five needed to be repainted) - some NOS parts in my stock were still in tolerance:

Paper Capacitors

All paper capacitors were either tubular types with their ends sealed with tar, or were in capacitor packs with multiple capacitors inside.  All individual paper capacitors were rebuilt in their original cases using modern 630 volt film capacitors in order to maintain the original under-chassis appearance.  Most capacitors were typical early RCA tubular capacitors, which have no markings other than the part number.  The part number on the capacitor apparently bears no obvious relationship to the part numbers in the published parts list! I remove each capacitor from the radio by unsoldering its leads.  The lead lengths and use of any insulating sleeving lengths are then noted.  I take notes of which lead goes where in the radio, and identify the capacitor using marking tags (which note the part identification and the location in the notes where it was removed).

The tubular types are very easy to restuff.  I simply pull the leads out of each end.  This usually also removes most of the tar used to seal the capacitor.  The capacitor foil roll is then pushed out one end.  Any tar remnants are removed using an Exacto knife with #11 blade.  I then place the cleaned up cardboard cover over a small alcohol lamp to melt any wax or tar left on the outside, and then wipe it off with a paper towel (this also removes dust and dirt).  Leads of the replacement axial film capacitor are then extended if needed (referring to the original component lead lengths).  The replacement capacitor is then wrapped in a strip of paper towel cut to the width of the replacement capacitor.  Enough is wrapped around the capacitor to hold it securely inside the cardboard tube and to center it.

Then ends of the restuffed capacitor are then sealed using melted rosin salvaged from servicing early RCA Superhet catacombs, which melts at a fairly low temperature.  Other restorers have used hot melt glue or even caulk to reseal capacitors.  The original insulating sleeve, or a replacement (black spaghetti tubing) is then installed.  Finally, the capacitor is tested before installation in the radio.

One capacitor, C59 rated at 400pf (voltage unknown) when removed from its case was found to be a mica type capacitor and tested OK (380pf with no leakage).  It was reused - reinstalled in its original cleaned up case and sealed with rosin.  There were several others of the same type with different values.  These all tested good and were also assumed to be mica capacitors.  All the remaining capacitors of this type were left in place and re-used after testing (by disconnecting one lead).

One original RCA wax/paper capacitor had been replaced (C58: the tone control capacitor, 0.01mfd). I noted that there were two other 0.01mfd capacitors in the radio.  Their RCA part numbers were 69037G2 and 69061G2.  I found an RCA 69037G3 in my junk capacitor box.  This part was restuffed and used to replace the missing RCA part C58.

Restuffed RCA Tubular Capacitors

Capacitor Packs

There were two capacitor packs used in the radio.  One was a small rectangular metal cased unit with wire leads containing C51 (0.5mfd unknown high voltage) and C42 (a 10mfd 25 volt electrolytic).  This unit was riveted to the chassis. The rivets were drilled out and the capacitor opened up, tar removed, and restuffed using a 0.47mfd/630 volt film capacitor and a 22mfd/50 volt electrolytic.  The two replacement capacitors were first wrapped in shrink tubing, then with strips of paper towels to center them in the original case.  Melted rosin was then poured into the assembly. The existing wire leads as well as the original cardboard cover were reused. The capacitor was reattached to the chassis using 6-32 hardware.

The other capacitor pack was a clamp mounted cardboard case dry electrolytic with wire leads.  This capacitor contained C45 (10mfd @ 50 volts), C46 (4mfd @ 450 volts) and C55 (8mfd @ 200 volts).   The original capacitor parts were sealed in tar.  The components were removed from the cardboard case by heating the outside of the case with a heat gun until one of the capacitors could be pulled out by its leads.  Once one capacitor was removed, the remainder of the parts as well as the tar could be cleaned out mechanically. It was restuffed with 22mfd/50 volt, 4.7mfd/450 volt, and 10mfd/450 volt electrolytics.  One of the five leads is in common with two of the capacitors, as shown in the schematic in Riders (make sure the polarity is correct!). The replacement parts were secured using shrink tubing then wrapped in strips of paper towel to secure and center them in the original case.  Melted rosin was then poured in to secure the capacitors in the original case. The existing wire leads and well as the original cardboard bottom cover were reused.

Filter Capacitors

The original filter capacitors C53 and C54 were still in place. The original capacitors were both 10mfd wet type clamp mounted units - working voltage unknown. There was plenty of room inside the case for mounting replacement capacitors. The restuffing process for this capacitor was as follows:

Tuning Capacitor

Prior to removing the tuning capacitor, the dial lamp sockets, the dial pointers, dial scale, dial pan, dial lamp diffusers, anti-backlash gears and vernier drive assembly were removed for cleaning and lubrication. 

The tuning capacitor had to be removed from the chassis for cleaning and to replace the deteriorated mounting grommets. Removal was very difficult and somewhat dangerous. There were four braided ground leads soldered to lugs on the chassis and five more soldered to the brass rotor grounding fingers (these braids went through holes in the chassis to grounding points underneath).  These braids were disconnected from the fingers, leaving the under-chassis connections intact.  The three stator connections were made using thick bus wire, the other end connected to lugs on the band switch!  One must use extreme care here to avoid breaking the band switch.

After removal from the chassis, the tuning capacitor was cleaned in my old Heathkit ultrasonic cleaner using dilute ammonia. Since the entire unit could not be cleaned all at once, several cleaning steps were done, with the capacitor at various angles.  After this, the capacitor was cleaned using soap, water, and old toothbrushes. It was then dried using a heat gun. 

The mounting grommets were replaced using GLg-Philco grommets available from Renovated Radios.  These are slightly too thick, but have the correct center hole and chassis mounting hole diameters. 

The vernier drive assembly was cleaned in my old Heathkit ultrasonic cleaner again using dilute ammonia followed by soap, water, and toothbrushes. It was then dried using a heat gun and was lubricated using automotive distributor cam lubricant (grease).  The assembly could not easily be disassembled for proper cleaning and lubrication. The two halves of the assembly were riveted together!  In addition, other parts were pressed or swaged together and could not be removed without damage. After the cleaning and drying, as much lube as possible was coaxed into the assembly using a toothpick.  It seemed to work OK at the time.  I will know more after the set is back together.  The anti-backlash gear and spring was also not working properly, even after cleaning and lubrication. It appeared that the spring was not properly engaging the gear.  Only a small amount of rotation would cause the gear to slip and disengage with the spring.  Like the rest of the assembly, everything was pressed together and could not be disassembled without damage. I'll just have to live with it the way it is.

The dial scale was replaced using a new reproduction dial available from Mark Oppat at http://www.oldradioparts.net/dials_intro.html (part #27255).  Holes in the dial lamp diffusers were repaired using scrap plastic and GC Service Cement.  I was unable to determine the correct pilot lamp to use (#40 or #46).  Some Antique Radio Forums readers advised using #50. I suppose that a #50 lamp would run cooler than a #46, since it draws .22 amps at 7.5 volts. I was concerned that the dial could be damaged by using the wrong bulb (I have seen too many Zenith black dial sets with holes burned in the dial scale). The diffusers on my radio already have holes, perhaps normal due to age and use. I did not have any #50 bulbs in stock, so I installed #46 bulbs.

75 Tube Grid Lead

The grid lead to the 75 tube cap was found to be shorting to the shield. The inner conductor was rubber covered and had deteriorated. All shielded cable in this radio had a cloth outer covering.  I wanted to try to reproduce this cable if possible. I found a suitable piece of shielded cable in my junk box. The outer braided shield was bare.  I was unable to reuse the original outer cloth covering.  So I used the outer cloth covering of a piece of cloth covered power cord.  I measured the original cable and attempted to reproduce it, since this cable is very visible on the back of the radio.  The reproduction turned out OK, even including the cloth tags where the outer cover was penetrated and the inner lead pulled through.  The color is close but not the same as the original (on the other hand, the original had likely faded).

Other Repairs

The type 85 tube was replaced with the correct type 75.  All other tubes were left in place, since they tested good (or only slightly weak).

The power cord was replaced by modern reproduction cloth covered cord, and an old style acorn type plug installed. 

The four top chassis washers were replaced by CW4 washers from Renovated Radios. The center holes were enlarged using a 3/8" Forstner bit in my drill press - results were not perfect but were acceptable.  This provided the needed clearance for the 3/8" steel spacers.  The four upper washers actually support the radio chassis, and were totally flattened as found.  The bottom washers bear no weight and were reused, even though they were deteriorated. 

Cabinet

The cabinet was in very good shape but with a few scratches and dings. It was cleaned with GoJo hand cleaner and 00 steel wool, followed by a coating of Johnson's Wax.  The front of the cabinet was slightly loose and was re-glued.

Testing

After the radio was completely reassembled, power was applied through a wattmeter and fused Variac.  Power was brought up slowly while monitoring the B+ voltage and the wattmeter.  The radio came alive and worked well on all bands - no assembly errors which is amazing considering the complexity of the radio and the degree of disassembly needed to access parts for replacement!  The radio sounded great with the initial driver transformer ratio of 1:2CT.  I later tried the 1:3CT ratio, but could not tell any difference in either volume or fidelity.  I left it connected as 1:3CT.

The radio was then aligned. Although the radio worked very well before alignment, many of the adjustments, including the IF transformers, were not even close to specifications.  Some of the alignment frequencies used were unusual.  For example, on band B the signal generator and dial are to be set to 5160kHz!  I have a frequency counter which I use to set my generator to the required frequency, since its dial calibration is not all that accurate.  But I do not know how I was supposed to set the radio's dial to 5160 if the dial resolution is only 100kHz at that frequency!  Why not use 5100 or 5200?  What is magic about 5160?  Needless to say I had to estimate the pointer position.  I was able to set the all the bands at the given alignment frequencies (but they were off slightly at other places on the dial) even though the padder and trimmer adjustments had been done correctly..

I felt that the120Hz hum level was excessive, even though the filter capacitors were doing their job, and were the correct values. I was advised by the good people on the Antique Radio Forums to check the balance of the output tubes. When I did, I discovered that one of the 42 output tubes had failed - it lit up normally, but had zero emission. It tested good during my initial survey and initial testing!  Once it was replaced, the excessive hum was gone.  

Restoration Results

Most of my restoration objectives were met, but not all.  There was no intention of restoring the set to factory new appearance!  My objective is usually to reverse any prior servicing and make the radio appear to have never been repaired.  I do not go so far as to artificially "age" solder joints, as do some collectors!  Nothing gives away a restoration faster than bright and shiny solder joints.  Here are some of my "misses":

Chassis Before Restoration

Chassis After Restoration