About boiling capacitors, or restoring the factory condition of tube radios

I published the article in 2018 based on material sent by the author, who unfortunately did not provide his data.
Years pass and I still ask the Author to contact me.

Old radios. No one can deny their charm and style. They often have a lot of history attached to them, or are part of it themselves. The radios found on the market are in varying conditions, some functional and playable, others sometimes looking as if they'd been pulled from a river. Each one is worth a moment of our attention.

A radio playing in the living room will add atmosphere to the room and the satisfaction of bringing it back to life yourself. Repairing a tube radio while maintaining original integrity is more labor-intensive and time-consuming than simply replacing damaged components. However, it's worth every minute of the effort. This will leave the receiver in a condition similar to when it left the production line. Dear readers, I'll present to you such a restoration process, based on a Pioneer radio receiver.

Let's get to know the patient.

A Pionier radio, type B, with a nice black box and yellow dial, and a rear panel in good condition. As for the good news, that's about it. Looking inside, we see a damaged power cable, rust on the chassis, and the speaker.

After removing the chassis from the case, we see protruding wires from the coils on the range switch and a few non-original resistors. Horribly, the tuning capacitor moves with difficulty, and its plates rub against each other.

Unscrewing the speaker doesn't bring good news either. There's so much dirt in the gap between the voice coil and the magnet that it prevents the diaphragm from moving.

Measurement of electron tubes – and this time the good news, electron tubes are alive and well.

The switch in the potentiometer works, and the range switch, after a little force persuasion, also started turning.

Knowing the condition of the most important components, I was able to determine the order of work:

1. Disassembling the chassis into its component parts, creating a wiring diagram at the same time.
2. Rust removal and chassis painting.
3. Range switch repair.
4. Cleaning the tuning capacitor and potentiometer.
5. Capacitor regeneration.
6. Replacement of non-original components.
7. Reproduction of details.
8. Assembling everything on the chassis.
9. Starting and tuning the receiver.

Receiver diagram:

As you can see, this is a simple, battery-powered superheterodyne. It has 6 tuned circuits, 3 ranges, and is powered by a 1.4 or 2V filament battery and an 80-120V anode battery.

If the chassis only has surface rust, we can remedy this by locally removing it and protecting the cleaned surface. Unfortunately, in this case, the rusting process was already very advanced. I disassembled all the chassis components and created a drawing of the connections and component layout. It might seem that a few photos would be sufficient, but sometimes components overlap and connections run at different levels, making it difficult to tell from the photo. The drawing is also resistant to damage to the hard drive or memory card (1).

I separate the disassembled components into their respective boxes: capacitors, resistors, mechanical components, and inductors. They will be subjected to various tests to determine their efficiency and suitability.

After removing the components, we can see large areas of rust in all their glory, including under the chassis.

I unriveted the bases and removed the rubber grommets, and placed the bare chassis in a box with other components for sandblasting. I started the electrical repair with the most important component – ​​the range switch. The success of the entire process depended on its repair. As you can see in the photo, all the coils, except the short-wave ones, were severely damaged.

I disassembled the switch's mechanical components, which, like the chassis, need to be cleaned of rust. The switch plates were coated with contact cleaner. I disassembled the tuning trimmers and cleaned the plates and mica spacers. Dirt accumulating in this area causes leakage and reduces their Q-value, and therefore the Q-value of the entire circuit they operate in. The mica spacers can be cleaned with window cleaner, but be careful as they are very brittle and can break easily.

Next, I removed the wax that encases the cores in the coils. This is usually not a problem; if the wax is hard, you can gently heat it with a hairdryer. The tuning cores are brittle, so use a well-fitting 1x3mm flat-blade screwdriver to remove them to minimize the risk of chipping the core. If the core refuses to unscrew, you can try screwing it inward so it falls out the other side. This note only applies to long and medium wave coils, as in some shortwave coil models, the wire passes through the center of the former, blocking the outlet.

Let's return to the problem of the "eaten" coils. First, I had to determine how many turns were actually damaged. During the inspection, I found damage to the coupling windings in the medium and long wave heterodyne coils, as well as damage to the antenna coils in the same bands. The inspection was confirmed with a multimeter; the replaced coils had no continuity. In the book Coils for Receivers by H. Borowski, we can find the winding data and inductances of these coils:

I started the repair with the antenna coils. On the longwave, the first 30 turns were chewed through, but that wasn't the end; the wire was also broken in the next layer. In total, I unwound 57 turns of the 1100 total. On the mediumwave coil, I unwound 34 turns of the 500 total. The heterodyne coils should have 30 and 25 turns for longwave and mediumwave, respectively. They were in shreds. I removed all these windings along with the paper separating tape. I prepared a new spacer and wound new windings using wire salvaged from the longwave antenna circuit coil.

I wound the coils en masse, adding three turns to each, and secured them with a pitch. The direction in which the windings are connected is important in an oscillator. If they are connected in phase, oscillations will not occur. Since I didn't know where the beginning and end of the coil should be connected, I soldered them randomly for now. I'll check if I got it right during startup and tuning. I also made a drawing of the switch with the connections:

After cleaning the switch's mechanical components, I reassembled it. To see how reducing the number of turns affected the coils, I measured their inductance. The measurements showed that the inductance was within the range described in the previously mentioned book.

Next, I tackled the next inductive components – the intermediate frequency filters. After powering up the cores, I took a closer look at the 200pF capacitors, which form intermediate frequency resonance circuits with the coils. These capacitors are causing significant problems; the silver layer sputtered onto the mica becomes sulphated over time, reducing capacitance. The contact between the mica board and the leads also deteriorates, resulting in a sudden loss of reception during operation or crackling noises during operation. The capacitors were desoldered (be careful not to break the coil leads in the process!), and I measured their capacitance, which proved to be correct, and bending and squeezing the board did not significantly affect the capacitance. Therefore, I deemed the capacitors functional and reinstalled them. If these capacitors have a low capacitance, you can try to compress the mounting rivets with pliers to improve the contact; if the capacitance differs significantly, you should desolder one of the capacitor leads and mount a 220pF/160V styroflex capacitor underneath.

The intermediate frequency filter housings have been cleaned and lightly polished.

Unfortunately, the intermediate frequency eliminator coil couldn't be saved. It has a broken lead at the bottom that can't be removed. I cleaned and slightly tightened the connectors to improve contact and removed the bitten wires. The whole thing will be installed as a dummy until a suitable coil is found. When I have some free time, I'll try to mass-wrap the coil and compare its parameters with the working original.

Next, I checked the potentiometer. I carefully bent the tabs holding the resistive track board to the housing and removed it. Then I drilled out the rivets securing the power switch board and looked inside. The contact board was stiff, so I lubricated it, and the contacts themselves were cleaned. Since both the axle and housing were rusty, they were set aside for cleaning and painting. The resistive track was in good condition, so I simply wiped it down with a suitable cleaning agent.

Since there were already a lot of components to clean, I had to take care of them, leaving the electronics aside for a while. The sandblasted components were coated with zinc spray. Once dry, it creates a durable coating that resembles the original. Over time, it will darken in color than in the photos.

I re-riveted the cleaned bases to the chassis. Tubular rivets are hard to come by, and if they are available, they're either too long or too short. For a good appearance, the rivet should protrude about 1mm above the material being attached. For this reason, I make the rivets myself; they're cut to the required length. I use specially machined center punches for riveting.

I also sandblasted and painted the potentiometer housing. I cleaned and lubricated the axle.

The rivets securing the switch section were replaced with new ones. When attaching the resistive track board, I always bend only two lugs, so that in the future I have two spare lugs if one of the lugs breaks during occasional repairs. Over the years, coil capacitors develop a high leakage current. This causes voltage changes at the tube electrodes, and the capacitors themselves begin to heat up, sometimes even to the point of bursting the glass or paper housing. Leakage current can be significantly reduced by removing water and ceresin oxidation products trapped between the capacitor plates.

To achieve this, I boiled the capacitors in paraffin. To determine whether this process improved the capacitors' condition, I measured their capacitance and leakage current before and after regeneration. As can be seen from the table, the leakage current decreased significantly, and the capacitance returned to its rated value.

"-" – means that the meter was unable to measure the capacitance due to too high leakage current.

Someone with a keen eye will notice that the schematic shows seven coil capacitors, but only six were boiled. I forgot about the capacitor placed parallel to the output transformer winding. It was boiled along with capacitors from another Pioneer, whose chassis we've already seen in the painting photos. I'll share the recipe for boiling capacitors in another article.

The mica capacitors from the range switch were in good condition. Unfortunately, the capacitors under the chassis were damaged – their capacitance was significantly reduced or nonexistent. I replaced them with working units from a spare parts collection.

The photo shows how capacitors become damaged. The black coating is silver sulfide. The tuning capacitor was mechanically cleaned of loose dirt. I then removed the remaining dirt using electrolysis. It's important to clean the wires (or plates, depending on the design) that contact the rotor axis. I checked for short circuits between the plates and then equalized the capacitances of both sections. The slotted, outer plates are used for this purpose. I installed a slewing wheel.

When I started repairing the speaker, I had dark thoughts – particles blocking the movement could damage the voice coil. Next, I desoldered and unscrewed the output transformer. The voice coil proved to be in good condition. I carefully unscrewed the magnet and removed its mounting. I decided to leave the diaphragm in place, as it was firmly attached to the basket. I sealed the holes in the basket to prevent dirt from getting inside and cleaned the basket of rust and old paint.

When assembling a speaker, the most difficult part is positioning the core in the slot so that the voice coil doesn't rub against anything during movement. When testing, be careful not to damage the voice coil. A refurbished speaker requires new/old fabric.

With all the main components ready, I started assembling the chassis.

The resistor near the speaker tube was attached using a strip with two terminal blocks, which was unfortunately broken. To make a new one, I glued several layers of black card stock together.

After cutting out the desired shape, I milled the holes and installed the original plates.

The next step was to try on the capacitors according to the previously drawn diagram.

Before assembling the components, their terminals were cleaned and tinned for easier soldering. The resistors were checked for resistance and contact at the mounting clips (they can sometimes break, causing the resistor to malfunction despite appearing otherwise).

The receiver contained three non-original resistors, which, moreover, had values ​​that were different from their intended values. Two of them were replaced with spare parts, and the missing S1 resistor on the 1S5T tube was replaced with one of the correct value (5Meg) and hidden in an oil-immersed sleeve to minimize its appearance. I made the connections using wire recovered from disassembling the radio.

After checking the correctness of all connections, the radio received a scale, "working" knobs, a set of test tubes and temporary battery connection cables with magnets on the ends.

I started the startup by applying 120V anode voltage and measuring the voltage at all important points in the circuit (excluding the tubes). Since the voltages were as expected, I could proceed with the receiver startup. It's important to note that the circuit uses very high resistances. Connecting a meter with a 10MΩ input resistance to the grid of two 1S5T tubes creates a voltage divider with a 5MΩ resistor driving it, resulting in a much lower measured voltage than the actual value. Therefore, it's necessary to either measure the voltage drop across this resistor or recalculate the obtained value taking the divider into account.

After inserting the 3S4T tube, I checked the low-frequency amplifier for proper operation, frequency response, and output power.

Next, I inserted the voltage amplifier tube and repeated the tests. The amplifier's sensitivity is approximately 30mV, the frequency response is 30Hz–10kHz, and the output power is approximately 150mW.

After inserting the 1T4T tube, I checked the IF amplifier. When tuning supercritically coupled intermediate frequency filters, remember that the untuned circuit must be attenuated by adding a small capacitance and resistance.

Starting the heterodyne, the 1R5T tube, I started by disabling the oscillator by unsoldering capacitor C22. Then, I applied a 465kHz signal to grid 3 and tuned the first filter, remembering to load the untuned circuit.

Tuning coupled circuits is performed in several steps, with successive approximations reaching the point where further rotation of the cores no longer increases the output signal, each time transferring the load from one circuit to the other.

After re-soldering C22, I checked if the oscillator was working correctly. It did work on shortwave, but I had to resolder the repaired coils because they were connected backwards.

Connecting a standard probe to a resonant circuit will detune it. I attached a Mosfet follower to the probe shown in the photo, allowing the oscillation signal to be viewed by bringing the probe close to the circuit, without the need for a galvanic connection and thus without detuning the circuit. The probe signal is fed to the oscilloscope and frequency counter.

First, I tuned the oscillators for each range according to the scale (tuning with coils at the beginning of the range, and with trimmers at the end). Then I connected the signal generator and tuned the input circuits.

The generator connection should be made using a "dummy antenna" to match the receiver's input resistance to the generator's output resistance. To avoid having to spend too much time figuring out which hole tunes which circuit during tuning, I created a special "cheat sheet."

After tuning, it was time for reception tests. Using a short, 3-meter antenna at home, I managed to pick up Warsaw and the Pyrzowice radio beacon (20 km) on longwave. On mediumwave, I received 3-4 foreign stations and a few more on shortwave. Cracks, grinding, hissing, and growling were everywhere. This is thanks to technological advancements; long live switching power supplies, LED bulbs, and broadband internet. So I switched to a large antenna and decided to replace the mains anode power supply with a battery-powered power source.

The small battery roughly in the center is the glow battery (1.2V 2000mAh), while on the right are four batteries (25V 5000mAh) providing a total anode voltage of about 100V. Changing the power supply and antenna (and turning off the bulbs in the workshop) significantly reduced the interference. Czechs and Germans could be heard on longwave, mediumwave revived, and on shortwave, the Chinese joyfully presented their version of the news on the Polish section of Radio China.

That's it for now. I can't consider the repair complete yet because the power cord is missing, and I won't have it for another two weeks. While I'm waiting, the radio has been placed in a box and received a set of temporary knobs.

While preparing the article for publication, I received a package with a cable.

I mounted it to the radio, but first I made some aluminum sheet wire markings and applied them to the cables. All that remained was to assemble the connectors and build a retro-style stabilized power supply.