[Elecraft] K3 SO-239 Connectors

[Paul Christensen]

As indicated earlier, I included a link to a photo that shows my K3's 
upgraded Amphenol SO-239 connectors.

216.229.20.37/images/K3.jpg

The Mouser catalog P/N is 523-83-798.  The Amphenol P/N is 83-798.

In spending some time reviewing the placement of the K3's SO-239 connectors,
the RF connector mounting configuration can be improved with only minor
re-work as shown below. In the link to the following photo, you will see the
two panel-mounted SO-239 connectors with three (3) solder-lugs added as well
as the KAT3 ATU Board.

216.229.20.37/images/K3-1.jpg

The K3 comes with both the ANT 1 and ANT 2 SO-239 connections routed to the
KAT3 board with only a single-ended wire.  The connectors' RF return path
necessarily relies upon clean, tight, but unbonded connections through the
chassis and also through a path of screws washers, and an aluminum spacer
that holds the KAT3 to the K3' side panel.  In the photo, you will see that
the final RF attachment point to the KAT3 is through a zinc-plated,
Phillips-head screw located at the upper left of the board.  This is the
same attachment point whether its modified or unmodified.  Only, in the
modified layout, bus wire connects the SO-239 connectors straight to the
zinc screw connection point.  Removal of the KAT3 board is just as easy as
it was before.  It's now also possible to replace the aluminum spacer with a
ceramic or fiber spacer which, will force all RF current along the new buss
wire and keep high RF current from circulating through the K3's chassis.

At 100W of RF, I am not completely comfortable with the existing arrangement
given my past experiences with transceivers of other manufacturer who
similarly use single-ended RF wiring techniques and rely upon good chassis
and hardware connections for the RF return path.  In almost all cases where
RF ingress was an issue, those cases were resolved by creating a direct RF
return path from the antenna connectors, rather than allowing 100% of the RF
current to flow and circulate through the chassis.  True, in a bonded
configuration like this, the chassis will still allow for some RF current,
but the level of chassis RF current can be reduced by either: (i) running a
short buss wire from a solder-lug on the SO-239 connectors to the KAT3; or
(ii) using short (3") coaxial cable from the SO-239s to the KAT3 with
shields bonded at both ends.   I suspect that ANT 1/ANT 2 antenna port
isolation may be improved by using the coaxial option.

Since the wire length between the KAT3 and SO-239 connectors is so short, I
decided to use #20 AWG buss wire with solder-lugs, rather than use coaxial
cable, although ideally, something like 2 or 3-inch cuts of RG-174 with
Teflon dielectric would be ideal.  I may switch to that at a later point and
verify port isolation results.

Paul, W9AC