Re-Stuffing a Kellogg Capacitor Module

[Ed Morris]

The audio on this Kellogg 1000 Red Bar phone I recently acquired was very low, and the audio quality was somewhat fuzzy.  I checked the receiver element for resistance and it was close to spec at just over 49 ohms.  So I suspected there might be an issue with the capacitor in the audio circuit.

The Kellogg Red Bar has a plug-in capacitor module containing a paper multi-cap that provides 1uF to the audio/talk circuit and .5uF or 1uF, depending on ringer type, to the ringer circuit.  The ringer circuit uses a link to connect two .5uF elements in the multicap in parallel in the event 1uF is needed.  For a standard 20Hz ringer, only .5uF is needed.  See Photo 2.

When I unplugged the capacitor module from the network block, I could see that the capacitor had swelled and pushed the module cover off.  The capacitor is sealed inside with pitch, very similar to the old Philco block capacitors of the 1930's.  Somewhere over the years, the pitch seal had failed, and moisture caused the paper cap to swell.  See Photo 3.

Testing the capacitance of the module, I found the 1uF element measured over 3uF with an ESR of over 18 ohms.  One .5uF element tested over 2uF, and the other .5uF element tested a just under 1uF, both with high ESR.  See Photo 3.

In the next post, I will show the repaired module.

[Ed Morris]

After removing the old multi-cap, I scooped out most of the pitch, some of which I saved for re-sealing the module.  To correctly install the new capacitors, I downloaded and printed off the copy of the schematic from the TCI Library.  The pin numbers shown in the photo below correspond to the pin numbers on the Kellogg 1000 schematic and those impressed on the body of the module.

I then wired in a 1uF capacitor between pins 1 and 2 for the network, and a .47uF capacitor between pins 3 and 5 for the ringer.  The pins of the module are aluminum, so I was concerned about being able to solder them, but this did not turn out  to be an issue.  I did use a heat sink attached to each pin while soldering to avoid overheating the plastic.

I did not install another .47uF between pins 3 and 4 as it was not needed for my ringer configuration.  Two .47uF caps and a 1uF cap won't fit in the module, anyway.  If I had needed a 1uF cap for the ringer, I would have just used a 1uF cap between pins 3 and 5.

Once the new caps were soldered in to the module, I plugged it in and tested the phone.  Everything worked, so I dabbed a few blobs of pitch around the edges of the module, then snapped the cover back in place.

For more on this particular phone, see this thread:
http://www.classicrotaryphones.com/forum/index.php?topic=18315.0

[TelePlay]

Very nice. Definitely an expert radio man's work of art.

[Ed Morris]

Thanks, John.  I was still editing when you posted 

[unbeldi]

It is fairly common to find the covers elevated on these  #225 capacitor packs  and the induction coils (#113, #114) too.

However, I have a hard time conceiving how the capacitance of a condenser increases when it expands, as you suggest.  These are paper insulated metallic foil capacitors.   The capacitance depends on the total surface area of the foils, the thickness and permittivity of the dielectric.
So if it expands, the thickness increases, the plates are farther apart, and the capacitance should fall, not increase.
So, either the insulation is shrinking, causing the foils to be closer, or the permittivity is increasing.

If the dielectric becomes more polarizable, I think I expect increasing leakage too,  increasing the equivalent series resistance (ESR) too.
You state that you measured 18 Ω.  An ideal capacitor has ESR = 0 Ω. Looking up typical values for even new capacitors, many do have ESRs of several ohms.   So, I don't think that is terribly bad for an old unit such yours.

[Dan/Panther]

Good work Ed, that's how it's done in the Vintage Tube radio world. Thing that is neat, is you have no idea it's been modified.

D/P