|The Crosley 125 "THE
LITLFELLA" is an 5-tube AC superhet circuit
radio in a small cathedral cabinet. As purchased, the radio was
stated as not working. It receives only the broadcast band.
The radio had been serviced in the distant past, but no recent restoration judging from the age of the parts used. All repairs appeared to be the normal service shop type. I decided to attempt to reverse all previous repairs to the extent possible and restore the original above and below chassis appearance. The schematic for this radio can be found on-line at Nostalgia Air.
My antique radio restoration logs
This radio was purchased on eBay. The cabinet was in good original condition, as were the knobs and grille cloth - just the usual wear, dings and scratches. There were several major scratches or wear on the front, unfortunately. The radio was sold as not working. There was no external evidence that the radio had been restored. Even the line cord and plug were original, and in good shape (but with some breaks). I always avoid knowingly purchasing a radio that has been restored, as many collectors take shortcuts such as removing the original capacitors and filters.
The Crosley 125 is a 5-tube superhet circuit radio that receives the broadcast band only - typical in 1931/32. It has no RF amplifier stage, but does have a tuned pre-selector (3-gang tuning capacitor). This additional tuned circuit is needed to prevent excess image response with such a low IF frequency (175kHz). Only one of the IF transformers is tuned. The second IF transformer is an untuned bifilar wound affair, likely with broad response characteristics. The second detector is a tetrode anode bend type that provides gain in addition to detection. This type of detector was commonly used in early superhet circuit radios prior to the introduction of the type 75 duo-diode-triode tube. The anode bend or plate detector is used to reduce the tube count, but unfortunately has quite high distortion. The Crosley 125 is a non-AVC set, which means that the volume control must be manipulated while tuning to prevent blasting, overloading, or completely missing weak stations.
The radio had been serviced perhaps several times. This was a "well loved" radio, and someone had paid a lot of money to keep it running!
Two 124 tubes were blue Arcturus globe type tubes branded Crosley - they were good and likely original. The other three tubes were replacements and were not branded Crosley. All tubes were good.
One original tubular paper capacitor remained, and one had been replaced - the tone control capacitor. An AC line to chassis bypass capacitor had been added. It is not shown on the schematic for this radio or for the three other early Crosley chassis I have restored. My theory is that this capacitor was added by a service man so that the radio could be operated without a ground connection. Since the volume control shunts the antenna coil, a ground connection may be needed for this to be more effective.
All of the bypass capacitors in this radio were originally in two metal clad "bathtub" units, one with two capacitors and the other with four. The dual unit was intact. The quad unit was still in place, but two of its lugs had been cut off and tubular capacitors installed, mostly just hanging in space. These replacement capacitors were older units, perhaps circa 1940-1950.
The original dual filter capacitor and its mounting bracket had been removed and had been replaced by two tubular units under the chassis. The original was obviously leaking!
One original resistor had been replaced (247 tube grid bias resistor). The original was likely a flexible wire-wound resistor.
The speaker plug was missing, and the speaker cable had been simply passed through drilled holes in the socket and soldered to the lugs on the back side. The plug used has a history of shorts and intermittent connections to the speaker socket, so it was likely eliminated during servicing.
There were signs of an attempted repair of the volume control. There was some solder between the two halves of the case, and also some solder on the resistance wire inside. The outside lugs of the control had also been clipped off (probably so that the control could be removed from the chassis).
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 for parts 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.
Before starting repairs I made BEFORE photos of the chassis bottom. I use these photos to ensure that replacement parts and wiring are placed as close as possible to their original positions. Some radios are subject to problems such as oscillation or motor boating if wiring is re-routed or lead dress is not the same as the original. It was later found that there were several deviations from the published schematic in Riders, and especially the second detector plate circuit and plate choke and associated capacitors. There was a 100pf mica capacitor across the detector plate choke that was not in the schematic.
All tubes and the tube shield was removed. The tuning capacitor and dial assembly was removed for cleaning. All non-original parts were then removed and additional photos taken. These photos were likely closer to the chassis's original appearance than when first received.
The volume control and switch on this radio was original. It shunted the antenna coil primary on one end of the resistance element and the other end varied the bias on the IF amplifier stage (the center or rotor contact was grounded). In order to provide smooth control of volume in such non-AVC sets, these controls often have two different types of resistance wire, and are almost impossible to replace with standard controls (effectively, they have a reverse taper). Measurements indicated intermittent operation and no continuity of the element. The usual failure mode is a defective weld or joint between the two different wire sizes used. The control was disassembled, and the resistance wire contact area and wiper were cleaned using lacquer thinner and an old toothbrush, followed by a wipe down with paper towels. There was a solder blob at the point where the two different resistance wire types met. So a repair had been attempted and failed (solder will NOT stick to the nichrome wire used).
This control was repaired by inserting a paper thin sliver of brass about 1/8"x3/8" (material normally used for model railroad construction) between the two sections of the resistance wire that had previously been joined (between the resistance wire and the insulator). The repair was successful - the control measured about 4.5K ohms and the resistance varied smoothly as the control was rotated. This was fortunate, since there is really no suitable replacement for this type of control. About the only substitute would be a 5K linear taper potentiometer with switch, or even better, a control with reverse audio taper. These are VERY difficult to find. My old IRC Control Catalogue does show a 10K reverse taper control existed. About the only likely source of one of these would be Mark Oppat.
The schematic shows that the control (Crosley part number W-23618A) has a minimum resistance built in. My control had a minimum resistance of 166 ohms, but there was a 300 ohm dogbone resistor in series with the lead that goes to the IF amplifier cathode. This was likely NOT original, and it was removed pending final testing. The part number of my control is W-23618 - with no "A" suffix.
The power switch was mounted on the back of the control. It was a conventional bat-handled switch operated by an arm attached to the rear of the volume control shaft. The switch was bad - it measured high or infinite resistance, likely due to dirt and corrosion. Fortunately there are some openings where cleaner can be injected with having to disassemble the switch. The control was first flooded with Big Bath cleaner, followed by Caig D-5, followed by more Big Bath. The switch was operated repeatedly during the cleaning operation. After this treatment, the switch worked reliably.
The tone control measured open, and was also intermittent in operation. Unfortunately it was the type with a paper element coated with a thin layer of carbon! There were worn areas where the carbon coating was completely gone. Any attempt to clean these controls will remove the carbon coating. So essentially the control would have to be replaced. I first attempted to hide a NEW IRC Q type control inside the original case (the original wiper assembly, shaft, and bushing were removed from the case). The original was 25K. I used a Q13-123 (50K, audio taper) to replace it. But there was simply NO room for the new control with its rear cover attached. The rear cover serves as a stop for rotation, so the control could not be used without the cover. I was forced to install the IRC control inside the original case with its rear cover in place. As much of the original back cover terminals and rivets were removed as possible using a Dremel tool. Several layers of electrical tape were used to insulate the exposed terminals and rivets from the back cover of the IRC control. The original control's back cover was then attached to the rear cover of the IRC control using epoxy. Bus wire leads were used to connect the lugs of the IRC control to the original control's lugs. This unfortunately left about a 1/8" gap between the original control body and its back cover. But this was better than using a visible new control! And the gap is hardly visible from the back.
Eight original dogbone resistors were out of tolerance by more than 20%. One was 1/4 watt, one was 1 watt, and the remainder were 1/2 watt sizes. In two cases (300 ohms 1/4 watt and 300K 1/2 watt) I had a NOS replacement available that was in tolerance. I collect NOS as well as used dogbone resistors just for this purpose, and buy all I can find on eBay and at swap meets. In the other cases, I did not have replacements available that were in tolerance. In these cases I attempted to find a replacement that was the correct size and had the correct measured value (within 20% tolerance) but not the correct markings! I then repaint the resistor with the value required using hobby enamel paint. In a couple of cases I was forced to use 1/4 watt replacements instead of the original 1/2 watt sizes. For example, the 15K 1 watt dogbone below was in reality a 10K that measured 14.4K ohms (well within tolerance for 15K). All I had to do is paint the end color green (it was black).
The bias resistor for the 247 output tube had been replaced by a newer 5 watt wire wound resistor. The original was likely a large flexible resistor, value 350 ohms. Here is an example of the likely type used from another Crosley radio:
I did not have a suitable flexible resistor in stock. So I decided to attempt to reproduce one. I used a piece of black spaghetti tubing (#11 gauge) for the body, which was then painted a maroon color using model railroad lacquer paint. The grid bias for the 247 output tube is developed across this resistor. The Riders schematic says that the bias for the 247 output tube is -16 volts. The current through the resistor would then be 46ma. The power requirement for the resistor was therefore 0.73 watts. I used three 100 ohm and one 47 ohm 1/2 watt metal film 5% resistor in series, and placed them inside the spaghetti tubing. The maximum power that needed to be dissipated by any one resistor was thus 0.21 watts. The ends of the tubing were wrapped with fine bare bus wire and tinned, much like the original. Here is a photo of my reproduction:
All of the bypass capacitors in the radio are in two metal cased "bathtub" type units with either 2 or 4 capacitors inside. These capacitors were still in place (riveted to the chassis). Two of the lugs on the capacitor containing four 0.1mfd capacitors (W-22412) had been cut off and two tubular paper capacitors installed in its place. The other capacitor containing two 0.5mfd capacitors was intact (W-24355).
The metal capacitor cases were removed from the chassis by drilling out the rivets and their contents removed mechanically, retaining the original terminal boards and remaining lugs. The capacitor cases and terminal boards were then soaked in paint thinner and cleaned with old toothbrushes, followed by a final cleaning with lacquer thinner. On the four-capacitor unit, two original terminal lugs as well as an intact terminal board remained. The terminal board on the opposite end was damaged and the lugs cut short. I was able to find suitable lugs and a replacement terminal board from another donor metal cased capacitor from my junk box. Once all parts were cleaned up, the terminal boards were attached to the cases using epoxy. Suitable axial film capacitors were installed inside the cases (four 0.1mfd at 630 volts in one, and two 0.47mfd at 630 volts in the other). One end of each replacement capacitor was crimped and soldered to each terminal lug on the inside, and the other end soldered to the capacitor's metal case. The capacitor cases were then filled with melted rosin/wax salvaged from servicing 1924 RCA Superheterodyne catacombs in order to stabilize the contents. This wax melts at a low temperature and will not damage components. The capacitors were reattached to the chassis using 6-32 hardware and star washers.
Only one original paper tubular capacitor was still in place. It was restuffed using a 630 volt film capacitor. It was a Dubilier Cub type, which are difficult to restuff since they have a wooden dowel down the center! Here is my method of restuffing Dubilier Cub capacitors. For the missing capacitor (0.05mfd, unknown voltage) also assumed to be a Cub type, I found a 0.1mfd 200 volt Dubilier Cub capacitor in my junk capacitor stocks. This was also restuffed with a 0.047mfd 630 volt film capacitor. I did not have a 0.05mfd Cub type capacitor in my stock. It was positioned in the radio so that the markings were not obvious (but may confuse some future restorer!) Here are the restuffed Dubilier Cub capacitors:
The original dual filter capacitor had been removed and had been replaced by two tubular units under the chassis. There was gunk and corrosion on the side of the chassis where the original capacitor was mounted (and had leaked). This provided a clue as to the size and shape of the original, which was a rectangular cardboard-cased unit with three terminals, clamp mounted to the chassis. On-line searches failed to find photos of any original examples. But someone who had perhaps read my Crosley 124-M PLAYTIME restoration log sent me an email with photos of an original Crosley 124 filter capacitor along with its measurements! The Crosley part number for the models 124 and 125 are the same: W-23801. In my case, the original clamp was also missing (there were holes in the side of the chassis where it had been mounted).
I formed a cardboard case using thin cardboard from the back of a writing tablet. Two 10mfd 450 volt electrolytics were mounted inside, and their wire leads connected to three ground lug terminals mounted on one end. The cardboard case was assembled using carpenters wood glue and rubber band clamps. The finished case was then painted using satin black semi-gloss enamel. A label was fabricated using the correct Crosley part number and values. The Cornell Dubilier brand and logo was used, since several original capacitors were CD brand Here is the result (the clamp is a reproduction made from aluminum stock):
From a lead supplied on the Antique Radio Forums Classified, I ordered a reproduction Crosley speaker plug from Barry Dagestino. It was very similar in appearance to an original plug, although it lacks the spring type prongs of the original. It uses what looks like 4-40 screws for contacts, which were slightly too large initially and had to be filed down some in order to insert the plug into the socket.
The original line cord and plug were reused. There were three places where the (very stiff) cord had broken and bare wires were exposed which could short out or cause a shock hazard. One break was at the chassis grommet. The two additional breaks were repaired.
The two extant Crosley branded 124 tubes were Arcturus Blue tubes. These were good and likely original. The Philco 80 was a globe type. These would be reused. The 47 and 35/51 were shouldered or ST types. I did not have a globe type 35 or 51 in stock, and this tube was hidden by shielding, so it was reused. The 47 was replaced by a globe type 247.
The cabinet only needed a good vacuuming inside and then cleaning on the outside with GoJo and 00 steel wool. It turned out definitely presentable.
Once the radio chassis was reassembled and the tubes installed, power was brought up slowly using a variac. AC power consumption was monitored using a watt meter, and a DVM monitored the B+. The radio powered up and worked immediately. The shields were installed, and then the radio was aligned. The radio performs well, is quite sensitive and has very good tone due to the large Magnavox speaker. The rebuilt volume and tone controls work smoothly. The 47 tube bias voltage was measured at -15.4 volts at 110 volts AC in. The schematic mentions -16 volts (close enough). The B+ measured 260 volts at 110 volts in (specification was 230 volts). This was possibly due to my using a 10mfd input filter capacitor vs. the specified 8mfd.
A ground connection definitely improves performance and reduces noise. I experimented with connecting the chassis to an AC socket neutral and hot through a 0.01 and 0.1 mfd capacitor (the radio had a 0.01mfd line to chassis capacitor installed originally, which was NOT original). Reception was improved, but a solid ground connection to a grounded AC outlet worked much better.
I was able to successfully reverse all previous repairs and restore the likely original appearance of the radio under the chassis. Of course, I did not know the exact appearance of some of the original parts. The reproduction filter capacitor was the correct size, but coloring, labeling and terminal type and configuration was just a guess. Exact placement of components and wiring was also unknown, but there were some clues such as some original component lead lengths. The speaker plug used was a reproduction.