Atwater Kent Model 184 Restoration

The Atwater Kent Model 184 is a small 4-tube AC superhet circuit radio. It receives the standard broadcast band only. It uses a very unusual circuit which incorporates regeneration in the second detector in lieu of a separate IF amplifier stage There is a regeneration control on the back, just like mid-1920's sets! And yes, if set incorrectly, it will squeal! When set correctly, the set is very sensitive for a 4 tube radio.

The set had seen minimal servicing in the past. Only one paper capacitor and one resistor had been replaced, and replacement filter capacitors had been tacked in below the chassis (the original filter capacitor cans were still in place). That being the case, I decided to try and maintain the original top and bottom chassis appearance if possible and reverse previous repairs to the extent possible.

The schematic for the AK184 can be found on Nostalgia Air. Any part numbers will refer to numbers on that schematic.

My antique radio restoration logs

Previous Repairs

The original filter capacitor cans were still in place, and tubular electrolytics installed under the chassis.
Only one original Atwater Kent resistor and one wax-paper capacitors had been replaced.
The power cord plug had been replaced.
Deteriorated insulation on several wires, as well as the crumbling insulation on filament shunt resistor R14 had been covered with tape.
Only one tube, a type 42, was branded RCA-Cunningham, and could have been original. The remainder were replacements.

Overview

The chassis for this radio is very compact and is difficult to service. Some components are buried beneath others that must be removed for access. Most wiring was rubber covered and deteriorated - insulation would fall off if the wire was moved even the slightest amount. The chassis was VERY dirty as found:

Survey

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.

The AC power switch on the volume control was bad - dirty and/or oxidized contacts likely. The control itself was OK after cleaning.
The speaker field and cone, output transformer, power transformer, and IF transformer were good. The speaker field measured about 2000 ohms vs. 1000 ohms as documented in the Riders schematic. I have restored other model 184's, and they had 1000 ohm field coils. This was a bit worrisome. The speaker may have been replaced. For example, the speaker for the earlier (and virtually identical) model 984 is about 2000 ohms!
The antenna coil and oscillator coil were OK.
Lots of rubber covered wire was bad - the insulation would crumble and fall off if the wire was moved even a slight amount. Most of this wiring would have to be replaced.
All resistors were in tolerance - rare indeed!
The original filter capacitor cans were still in place, and tubular filters had been tacked in under the chassis.
The power cord was original, but had several breaks. The plug had been replaced.
The ground lead (black) had been removed.
The tinfoil covered cardboard tube shields were missing, which is usual in this model. In fact, I have only seen ONE set with original shields. These were likely removed by servicemen to replace the tubes. Since they have a wire soldered to the foil and to a chassis ground lug, which has to be removed, they were likely NOT replaced. Without these shields, the set will break into oscillation if the regeneration control is advanced too far.
C7, 120pf mica, was defective (measured only about 50pf).
R5 (1 meg) had been replaced by a 250K resistor.
C4 had been replaced.

Repairs

All tubes were removed. 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. Components were identified and marked on the photo with a red felt-tip pen. I then removed all capacitors, documenting their locations and connections. Some resistors were also removed for access. The original filter capacitor cans were removed for rebuilding and chassis access. The volume control and tone switch was removed for access and for cleaning. The tuning capacitor was removed for cleaning access to the chassis. The antenna coil T1 and IF transformer T3, along with their shields, were removed for cleaning and for lead wire replacement.

When I replace a resistor or capacitor, I always remove the original part completely from a terminal. Other components connected at the terminal are protected from heat using old medical clamps. 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 of the chassis was cleaned with GoJo hand cleaner and 00 steel wool. The tuning capacitor was cleaned in an old Heathkit ultrasonic cleaner with dilute ammonia. After drying, the bearings were lubed with distributor cam lubricant. The dial drive reduction mechanism was removed, disassembled, cleaned, and relubed using automobile distributor cam lubricant.

The original power cord was shortened in order to avoid breaks and fraying. An old style bakelite plug was attached.

The power switch on the volume control was flooded from the front with Big Bath cleaner and the switch cycled. It eventually worked OK. I was NOT able to remove the cardboard rear cover in order to properly service the switch (it was stuck to the volume control, and attempting to remove it would have likely destroyed it).

Tube Shields

The original tinfoil covered cardboard tube shields were missing, which is usual. Here are what the originals looked like:

Repeated Want To Buy postings on the Antique Radio Forums Classified resulted in no hits for original replacement shields. In previous restorations of this model, I installed Goat type metal shields on the two 6C6 tubes, with a wire soldered to the shield and to the extant chassis ground lug. This time I decided to try something different. I first tried to reproduce the shields using cardboard tubing and aluminum foil. The results were not satisfactory, as the foil tends to wrinkle, and I was unable to find the correct thickness of cardboard tubing. I had in my parts stock two covered tube shields along with their shield bases, so I decided to install these shields. In the case of the 6C6 First Detector, the rivet holes for the tube socket matched the shield base exactly. But not so for the 6C6 Second Detector - it had a much larger socket for some reason. So I drilled holes through the chassis as well as the socket (between pin connections) for the tube shield base. The shield bases were installed using 4-36 hardware. The screws were installed on the bottom (tube socket side), and small nuts on the top (tube shield base side). Any excess screw length was cut off using a Dremel cut off tool.

Resistors and Capacitors

All the original Atwater Kent paper capacitors were rebuilt in their original cases using modern 630 volt film capacitors in order to maintain the original under-chassis appearance. I reseal the cardboard tubes using rosin salvaged from RCA catacombs (it melts at a low temperature and will not damage the replacement capacitors). I did not have an exact replacement for the missing type 208 (.05mfd) AK capacitor, so I fabricated one using an old Sprague cardboard tube and fabricated a label using Microsoft Word. The Sprague tube was restuffed and the label glued on.

Reproduction Cap	Original AK Cap

The original filter capacitors cans (wet type) were rebuilt using new parts. C14 (10mfd/475 volts) was rebuilt using a 10mfd/450 volt tubular electrolytic. C13 (16mfd/300 volts) was rebuilt using a 22mfd/450 volt tubular. The cans were scored deeply about 1" up from their base on my Unimat lathe, and then cut in two using a hobby razor saw. The positive electrodes were then removed, retaining the stud. The flat portion of the stud was cut off short and a hole drilled for a screw, nut, and solder lug. The replacement electrolytic capacitor was then connected to the solder lug. The negative lead, extended with bare buss wire and with insulating sleeving, was routed through a small hole drilled in the base. C14 was covered by a cardboard insulator. In this case, the upper part of the capacitor was not reused - only the base section was reinserted into the cover, which was then retained by a clamp on the chassis. The two halves of C13 were rejoined using a PVC plumbing coupling, and secured with epoxy cement.

I normally test mica capacitors if one lead has to be disconnected anyway. C7 (120pf) was tested and was bad - it measured only about 50pf. It was replaced with a very similar in appearance 100pf unit.

The insulation on the filament shunt resistor R14 was crumbling and falling off. Someone had wrapped it with friction tape! The resistance wire was OK. The center tap lug was first uncrimped and the resistance wire removed. The rivet was drilled out and excess solder removed. Two layers of spaghetti tubing (small and large) were then placed on both legs of the resistance wire. The terminal ends were insulated using heat shrink tubing. The two ends of the resistance wire were then twisted and secured using a small 4-40 screw and nut through the center tap lug. Again, heat shrink tubing was used to cover the repair and insulate the terminal lug.

Wiring Error, or Design Change?

When C8 was removed for restuffing, I noticed that it was NOT connected per the schematic. It was connected directly from the plate of the second detector 6C6 tube to the grid of the 42 output tube. In the schematic, it should connect to the junction of R4 and R6. I assumed that the purpose of R4 is to provide a voltage drop and regenerative feedback voltage through the regeneration control A5 to the "tickler" winding on the IF transformer T3. If C8 were connected directly to the detector plate, the feedback voltage would be reduced due to the shunting effect of C8, C10 (when bass tone is selected), R10, and the grid of the 42 output tube. C8 was original, and there was no evidence of it having being moved. The parts placement diagram supports the schematic. I'm wondering if this reduction in feedback voltage was ON PURPOSE and a design change. Perhaps there were complaints from customers that the radio was too unstable if setup as designed. I have found that one must connect the antenna and ground and THEN adjust the regeneration control. If the antenna or ground is changed in any way, the set may break into oscillation. I restored the set per the original schematic.

One other model 184 I restored was wired per the schematic, as was a model 984 I restored.

Wiring

Most of the original rubber covered wiring had to be replaced. The insulation would crumble and fall off if the wiring was moved even the slightest amount.

Testing and Alignment

Once the radio 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 came alive immediately and worked. The set was then aligned - no surprises. The radio worked well and picked up lots of stations using only a 20' piece of wire in my basement. The radio is VERY sensitive, and especially for a 4-tube set.

The B+ was about 246 volts - spec was 225 volts. But line voltage here is quite high, about 122 volts, and the radio was originally designed for 115 volts. But obviously the 2000 ohm speaker field worked fine and did not reduce the B+. Perhaps the schematic was incorrect! I did not think to measure the voltage drop across the field before reinstalling the radio in it cabinet. That would have provided a good cross check (schematic says the B+ drop is 55 volts).