Philco 90B Cathedral Restoration

The Philco 90B Cathedral is a large 9-tube AC Superhet circuit radio that receives only the broadcast band.  This example was the single 47 output tube version. 

The radio had seen a moderate amount of servicing in the past.  Six wax-paper capacitors had been installed to replace likely leaky or open Philco bakelite block capacitors. The copper Mershon filter capacitors had also been replaced.  But it appeared, at least initially, that all repairs were reversible.  That being the case, I decided to try to maintain the original top and bottom chassis appearance and to reverse any previous repairs to the extent possible.  

The schematic for the Philco 90 can be found on Nostalgia Air.  Any part numbers will refer to numbers on that schematic.  The schematic at is annotated with parts values in addition to part numbers, and is indispensable. 

My antique radio restoration logs


This radio was purchased on eBay at a reasonable price (these models go for big bucks, especially if refinished or have excellent original finish cabinets).  My example cabinet was presentable, with no damage or serious scratches or dings, but the finish was heavily oxidized, and very dark.  It had its original knobs and grille cloth, and the cloth was in good condition.  The power cord was original, but frayed in places.  All tube shields were in place.  When the radio was removed from the cabinet, it appeared that mice had been living inside!  There were lots of seeds and some nuts present.  But I saw no evidence of component damage or excessive corrosion due to their urine.  After cabinet cleaning, I did discover where small nails had been used to secure the arch brace and the cabinet base.  These were hardly visible and were left in place, since attempting to remove them would likely result in more damage to the finish.  The eBay seller did an excellent job at packing (triple boxed!) and there was no shipping damage. Here is what the radio looked like as received:

Previous Repairs

The radio had seen a moderate amount of repair in the past, and all repairs were far in the past, judging from the components used.  This was a well loved radio that someone paid a lot to have repaired, and several times. Six wax-paper capacitors had been installed to replace the usually leaky Philco bakelite block capacitors.  In three cases, the replacement capacitors had been installed in parallel with the existing capacitors, which had not been disconnected (perhaps they had failed open?) In the other three cases, wiring and other components had been disconnected from the bakelite blocks and left hanging in space.  In one case, one of the lugs of a bakelite block capacitor had been cut short.  But NONE of the original bakelite blocks had been removed, fortunately.  Once restoration started, and the bakelite block capacitors were removed for restuffing, I found that in 3-4 cases a terminal had been cut and then soldered back to the stub left on the capacitor!  I assume this was done by  the service tech to test the capacitor.  This was not obvious until the solder on the terminal was melted in order to remove a component or wire.


My usual restoration procedure is to first make a complete survey of the condition of all components.  The survey results guide my restoration strategy.  If major and unique components are defective or missing and cannot be restored or replaced, I may elect to sell the radio rather than restore it.  I always assume that all paper and electrolytic capacitors are leaky and thus should be replaced (I always "restuff" the original containers if possible).  Any mica capacitors are assumed OK until testing proves otherwise.  I never apply power to an unrestored radio, even through a variac or "dim bulb" tester.


The chassis was very dirty, and had seen mouse activity - two holes in the chassis top left when the filter capacitors had been removed provided perfect access!  No nest was found.  All tubes and shields were removed and dust was removed using an air compressor. Before starting repairs, I took photos of the chassis top and bottom so that routing of wiring and component placement could be restored.  Lead dress is often critical in radios.  The tuning capacitor was removed for cleaning, for replacement of the mounting grommets, and to remove the paraffin wax underneath.

I then removed all of the non-original capacitors, documenting their locations and connections.  In the process, I documented using a sketch where all the original components (resistors) were likely connected (to terminals on the bakelite block capacitors) When I replace a component, I always remove the original part completely from a terminal.  Other components such as mica capacitors and in-tolerance resistors 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.  

The top and sides of the chassis were cleaned with GoJo hand cleaner and 00 steel wool.  Since this process may leave metal residue, I then went over the chassis with a vacuum cleaner followed up by a small magnet and masking tape.  The tuning capacitor was disassembled and its parts cleaned using soap, water, and old toothbrushes. Some parts (or portions of parts) were also cleaned in my old Heathkit ultrasonic cleaner using dilute ammonia. After drying, the bearing surfaces were lubed with automotive distributor cam lubricant.  The tuning capacitor grommets were replaced by GLg-Tuner grommets from Renovated Radio, thick side up.  These worked OK, but the center hole was slightly large and allowed some movement of the capacitor.


Bakelite Block Capacitors

All original Philco bakelite block capacitors were removed from the radio (one or two at a time, making careful notes of connections), their contents removed, cases cleaned using lacquer thinner, and restuffed using modern 630 volt film capacitors.  Before removing the contents I unsolder the internal capacitor leads from the outside terminals and clean off the terminals. I use mechanical methods of removing the bulk of the potting tar and capacitors (small screwdrivers).  One must be careful NOT to pry against the bakelite case, as it is easily broken.  Some collectors use heat to remove the contents.  Here is the restuffed capacitor 52 (the line bypass capacitor, Philco part number 3793E), which contains two 0.015mfd capacitors:

Metal Cased Bypass Capacitor Blocks

The two metal cased bypass capacitors 24 and 30 were restuffed using new film capacitors.  The capacitor 24 had three identical capacitors inside, so there was no need to identify the leads.  I did measure the length of each lead as it was disconnected, since at this point I did not know if all the leads could be salvaged.  Another reason to measure the lead length (they were all different) is to maintain the original lead dress and routing.  In the case of capacitor 30, it contained three different capacitors, and originally had three different color leads.  Unfortunately, the colors had faded or were very dirty and the colors could not be identified.  They were identified by where they connect in the schematic, and were tagged using marking tags.  Again each lead was measured as it was disconnected.  The Riders schematic has no component values marked, so the leads of this capacitor had to be identified by referring to the schematic at

I was able to salvage all of the original lead wires and reuse them with one exception (a frayed wire).  The original insulators inside were also retained and reused.  These capacitor types are very easy to rebuild by straightening the metal tabs, removing the fiber top cover, unsoldering the internal capacitor common lead from the case, freeing up the contents by passing a thin blade between the insulator and metal case, and pulling out the contents.  The tar covered capacitor inside the insulator was discarded after removing the original lead wires for reuse (scrape away enough tar to unsolder the lead from the capacitor body).  Capacitor 24 was originally three 0.25mfd capacitors.  It was restuffed using three 0.22mfd/630 volt axial film capacitors.  The original leads were soldered to the shortened replacement capacitor leads, and insulating spaghetti tubing was applied to prevent shorts.  The three capacitors were then wound with a strips cut from paper towels, which was secured with masking tape in order to fill up the space inside the original can and original insulator.  A common ground bus was soldered to the side of the case where the original common ground was removed.  The original insulator is important since it supports the fiber board cover, which has a hole for the three leads to emerge.  Here is the restuffed capacitor 24:

Capacitor 30 was constructed the same way, and it was restuffed using the same method.  It contained a 0.1mfd, a 0.25mfd a 1mfd capacitor.  It was restuffed using a 0.1mfd, a 0.22mfd and a 1mfd film capacitor all at 630 volts. I only stock 630 volt film capacitors.  Its original leads were also reused.  

Mershon Copper Filter Capacitors

Both original Mershon copper filter capacitors (6mfd) were missing.  Fortunately I had several copper Mershon capacitors in stock, complete with their mounting and terminal nuts. Since Mershon capacitors were so distinctive in appearance, I purchase all I can find that are reasonably priced on eBay or at radio swap meets.  I had one 6mfd unit, one 4mfd unit and two10mfd units.  The 10mfd units were much larger than the originals (these were possibly used in the 25-40Hz versions of the Philco 90).  The 4mfd unit was about the same size as the 6mfd unit, and both were about the same size as the original 6mfd units in my restored Philco 70B.  So I used the 4mfd and 6mfd Mershon capacitors for my Philco 90 restoration.  An observer would have to have very good eyes to tell the difference between the 4mfd and 6mfd units. I was missing the insulating fiber mounting bushing and negative terminal for the one capacitor that was insulated from the chassis.  I was also missing the correct positive terminal lugs for both capacitors.  I was able to obtain all of these parts from the friendly folk on the Antique Radio Forums Classified.  

The replacement copper Mershon capacitors were restuffed.  Some collectors have been able to remove the tops of these capacitors in order to restuff them.  I have never been able to do this easily, and did not want to risk damaging the very nice looking copper capacitors.  Plus, with that method, the positive stud and electrode have to be removed as a unit (I wished to retain the original positive stud). Before restuffing, the cases were polished using Brasso metal polish. To restuff them, the capacitors were chucked in my small Unimat lathe (gripped by their mounting nuts) and their cases deeply scored about 1" up from the bottom.  The cuts were then completed using a hobby razor saw and cleaned up using an Exacto knife with a sharp #11 blade.  This leaves only a thin line on the case - hardly visible.  The original contents were then removed and the inside of the capacitor case cleaned. The original positive electrode foil was removed and the stud was cut short and then drilled to accept a ground lug and 4-40 screw and nut.  The positive lead of a replacement 10mfd/450volt capacitor was attached to the ground lug.  The negative lead of the new capacitor was extended, insulated using spaghetti tubing, and routed though a small hole drilled into the hard rubber base.  This lead was attached to the original ground lug (the insulated capacitor) or clamped between the capacitor base and the chassis.  The two parts of the case were then rejoined using 3/4" PVC plumbing couplings and epoxy.  I used a wrap of masking tape to ensure a snug fit inside the capacitor halves before applying the epoxy (Devcon 5-Minute Epoxy).  Here is one capacitor during assembly and when completed:

Tone Control

The tone control rotary switch contains three capacitors: a 0.015mfd and two 0.01mfd units.  It was restuffed with new capacitors using the same technique as used for the bakelite block capacitors.  The existing capacitors, which were embedded in tar, were dug out mechanically.  The existing wiring to the switch contacts was removed, since the rubber insulation would fall apart if moved.  As much tar as possible was then removed mechanically using various tools.  The bakelite case was then cleaned using lacquer thinner.  The capacitors were replaced using 630 volt film capacitors, and the wiring to the switch contacts was replaced using bus wire and thin spaghetti tubing.  The control was then sealed using melted rosin (left over from servicing RCA Radiola Superheterodyne catacombs) to hold the capacitors in place.  This wax melts at low temperatures and will not damaged new components.  Some collectors use tinted hot glue or even caulk for resealing components.


Most resistors in the set were a mix of early cast-end dogbone type resistors and later dogbone types having wire leads wound around their ends.  Seven resistors were out of tolerance by more than 20%.  Five were early cast-end type dogbones, and two were wound end type dogbones. I maintain stocks of NOS and used dogbone resistors of both types, and buy all I can find that are reasonably priced.  I also NEVER throw away a used dogbone resistor, even if out of tolerance.  Searching my stocks, I found suitable replacements for all seven (selected by measured value and size (wattage) rather than marked value, since most resistors of this age will have drifted, even if NOS). In the photo below, the 70K Philco cast-end resistor (top row, right) was in tolerance, although its leads were short.  The NOS 50K 1/2 watt wrapped end resistor next to it was also in tolerance. The original resistor was much shorter but was the same type.  I found one other cast-end resistor in my stock marked 70K that was close to 90K in value.  It was repainted as a 99K resistor using hobby paint (bottom row, right).  Other wrapped end resistors were found that were close to the remaining needed values.  These were also repainted to the needed values using hobby paint.  The final result was that three of the original cast-end dogbones were replaced by wrapped end dogbones.  This of course was a compromise, but reproducing cast-end resistors is beyond my skill level - some collectors do it - I have tried and failed.  I have seen other Philco model 90 sets where almost ALL of the resistors were wrapped-end vs. cast-end types.  So apparently there was a transition from the earlier type to the later type.

The volume control worked smoothly, and its resistance looked reasonable (160K).  Its resistance is not documented in the Riders schematic, but is marked as 500K on the Philco 90 schematic found on

Other Repairs

The rotary power switch  was removed from the radio and flooded with Big Bath cleaner while rotated many times (there are openings in the body of the switch where the cleaner can be applied).  This failed to produce a reliable ON condition as measured by my DVM.  I next attempted to clean the switch in my old Heathkit ultrasonic cleaner using dilute ammonia and water, followed by drying using a heat gun.  This also did not help.  I repeated the Big Bath cleaning step, and the switch worked about 90% of the time.  I then tested the switch using a 60 watt light bulb at full line voltage.  It seemed to work perfectly!  I suppose it needed to switch some current to work reliably. It was returned to the radio for testing after restoration.  I did have a suitable replacement in stock if later testing proved to be a problem - a NOS rotary switch of the same type, but with a longer round shaft and DPDT rather than single pole.

The power cord was original and somewhat stiff.  If folded sharply, the conductor insulation would break.  There were two areas where this had occurred, plus areas where the outer cloth covering was worn off.  In order to maintain as much originality as possible, the cord was repaired using shrink tubing.  An old style plug was installed.

The speaker cable had insulation breaks in places, but it was not replaced.  Replacement of the wiring would be difficult since the speaker plug is riveted together!  And of course, any replacement wiring would not look anything like the originals.


The cabinet was vacuumed and then cleaned using GoJo (the white type, not pumice) hand cleaner and 00 steel wool.  This removed some of the dirt, but the dark oxidized finish remained (mostly on the arch and sides).  The cabinet was then waxed using Johnson's paste wax. The grille cloth was simply vacuumed.  The chassis washers were replaced using CW-5 washers from Renovated Radios.  These seemed to be close to the correct size, as the control shafts were reasonably centered.  The next thickness available was much too thick.  The small nails securing the base and arch brace were left in place.

Testing and Alignment

Once the radio was reassembled and the tubes installed, power was brought up slowly using a variac. A DVM monitored the B+.  The radio came alive immediately and worked.  Globe tubes were installed in place of some of the newer ST types that came with the radio.  But I did not have enough good globe 24's or a good globe 47 to convert the set to ALL globe tubes.  I also did not replace one of the 27's with a globe tube, since it was hidden inside a shield. The set was then aligned.  The alignment went well, except for the low frequency adjustment.  By rocking the dial and adjusting the low frequency padder, I found the output was continuing to rise with the padder at maximum capacity (fully tight).  I did measure the parallel fixed mica capacitor (40), and it was right on specification (as was the parallel resistor 41).  I suppose that the coils may have changed inductance over time, and with moisture incursion.

The power switch worked reliably - I made a mental note not to rely on DVM resistance measurements in the future!  I suppose some switches (with dirty or oxidized contacts) may need some current through them to work reliably.  The set had excellent sensitivity all across the dial, and its tone was excellent.  The B+ was slightly high - 267 volts vs. 250, likely due to using 10mfd filter capacitors instead of the original 6mfd units.  Like all of my radios of this age, I operate the set using a fused bucking transformer at 111 volts (which was typical of 1931).  Line voltage at my home is 122 volts (ouch!)

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":

Before and After Restoration Photos

Chassis Before Restoration

Chassis After Restoration