[R-390] mechanical filter failure

[Charles A Taylor]

After opening up scores of the cylindrical nechanical filters that Collins
has in almost all R-390As, I found that the main mode of failure seem
to be the contraction of the black foam that encases the resonator
discs, and so allows them to migrate around in the encasing cylinder.

If yon take and shake a mechanical filter, and you feel something
moving around, you've got a filter that on its way to failure.

I've opened several filters and seen the "slug" loose. Just a shake
and the tiny #42 AWG wire will break....another broken filter.

And sometimes the lead breaks off at the bobbin, leaving no way
to splice some more lead in. Trouble.

After seeing several of these, I resolved to to do a little experi-
mentation.

I know of no source for plastic bobbins that Colllins uses in its
mechanical filter, so I looked through my many miniature plastic
  VHF coil forms.

I found one or two that look that they would take enough #42
wire to wind on then to couple to the end of the slug.

Starting first thing one morning, I wound 600 turns of #42 AWG
on one of these forms and measured the D-C resistance to
be nearly 50.0 ohms dead center.

If looking at the many 2, 4, 6, 8 & 16 kHz mechanical filters
transducers, I find that the transducer D-C resistance of both
the input and output transducers averaged 50.0 ohms.

Rarely does a transducer display 50.0 ohms D-C resistance.
There was a variation of as much as 5 ohms between one
transducer and another.

Which leads me to guess that the original bobbins were
hand-wound, and at considerable effort, too.

So far, after opening these transducers and finding a broken
lead or two, I've spliced a corresponding length of #42 wire
to the shortened lead and using a little of my wife's nail polish
to insulate the bare lead.

TO reposition and fasten down the "slug," I've cut 1/2" wide
lengths of fish paper and attached them to the transducer ends
cementing them in place with cellophane tape and Duco cement,
and rolling up the strips of paper until they just fit into the cylinder.
  ]
Then I use a piece of cellophane tape and Duco cement to
fasten the ends of the coils of fishpaper.

Finally, I slip the "slug" into the cylinder and use a little
RTV to fasten the fishpaper at the end of the slug into
place so that the "slug" is firmly fastened to the case.

Lastly, fasten (using weezers to wrap the #42 wire around
the tiny terminal), make a solder connection with a 25-w soldering
iron and #20 lead/silver/tin solder.

Finally, I resolder the ends closed with a 250-w soldering
gun quickly so as not to fry the innards.

Then I run a signal generator through the mechanical
filter into an oscilloscope to see if the filter works and
displays the nearly rectangular bandpass that makes the
mechanical filter so valuable.

ANY QUESTIONS, PLEASE?

Perfesser Chuck