[R-390] Lankford Filter Mod

[Charles Steinmetz]

Bob wrote:

>Once you have turned [the heaters] off there should be a few watts 
>to spare off of the filament winding.

It's not just a matter of the load on the transformer.  To be truly 
"plug-in," a replacement would have to run on the 0 and 6.3vac 
supplied on Pins 3 and 4 of the V202 socket.  This means it would of 
necessity be a half-wave supply (ugh -- exactly the wrong way to go 
for high current and low ripple, and a design compromise I will 
almost never accept).  It also forces you to use the chassis ground 
at the tube socket, when some other ground point for the mixer 
circuitry may be better for low noise. **

If you allow use of the 26vac winding, you either re-wire the tube 
socket (leaving a booby trap for anyone who later replaces the new 
mixer with an original 6C4) or provide another connector for 
power.  And if you accept adding another connector, why use the 
internal 26vac with one end grounded when you can just add a small 
transformer and have exactly the AC supply you want?

>I suspect that relatively few people are running 26Z5W's rather than 
>solid state diodes.

I've been under the hood of several hundred 390As, and about 70% of 
them still have 26Z5s.

Best regards,

Charles


**  The fact that Collins distributed the filament voltage with one 
wire and used the chassis for the return is one of the worst aspects 
of the 390A, and even many low-cost receivers from lesser 
manufacturers did not stoop that low.  It ensures that there are high 
60Hz chassis currents everywhere (50Hz where applicable), and 
directly causes a rise of more than 10dB in the hum floor.  I once 
(emphasizing *once*) re-wired one with ungrounded, twisted-pair 
filament supplies everywhere.  The hum dropped by more than 10dB and 
the radio sounded much quieter in all respects.  It was a wonderful 
improvement, but even so I wouldn't repeat it.