"Larry W4CSC" wrote in message
...
On Mon, 10 Nov 2003 07:56:03 +0100, "Meindert Sprang"
wrote:
Any chance of using ferrite ringcores on the NMEA lines to surpress it ?

Meindert

Not as long as the unbalanced lines are all exposed like they are.
Every cable from every instrument uses a Belden foil-shielded pair.
But the radiation is going to happen, anyway, because the output of
some of the instruments is unbalanced, inserting a radiating ground
inside the faraday shield. Most instruments, you have to abandon the
shield (screen connection for you UK readers) at the point where the
instrument's unshielded power cable with its dangling data lead hangs
out. There's no way to complete the shielding to the instrument.

But you only connect the shielding on the transmitting side of the cable,
right? Never connect shielding on both sides unless it is part of the data
connection. It is always better to have signal and return in a pair,
shielded by a screen that is connected on the TX side.

Maybe Icom is right. Make all the NMEA connections via a coax
connector, unbalanced. M802 uses a BNC, in total abandonment of any
NMEA balanced concept. The shield of the coax to that BNC MUST be
connected to NMEA B (-) to get data on the radio's DSC display.

Hold on! That cannot be right. If the NMEA source you connect to the M802
has a NMEA A (+) and a NMEA B (-) line, it is a balanced output where both
wires are 'live'. If you connect that to the BNC of the M802, which is
unbalanced (BNC is grounded), you'll effectively short circuit the NMEA B to
ground. In such a case, you only connect the NMEA A to the input and the
ground of the NMEA talker to the ground of the listener. you should leave
the NMEA B unconnected.

So, I figured RF from the transmitter's case follows this odd ground down
into the network shield and screws it all up. But that didn't pan out
because the network does the same thing with the cable to the Icom
disconnected. Lucky for me that during short SSB transmissions, the
system components just ignore the trashed data, so the users don't see
there's no NMEA data for the 20 seconds the SSB is talking. I think
this is the reason more people don't mention or notice it. The
displays just freeze until you stop talking or take a breath when the
SSB output power drops to a very low level and data stream resumes.

Normally this kind of interference could be solved by looping the NMEA wires
through ringcores, effectively breaking the antenna formed by the wires.

Oddly enough, I don't notice this malady on the Seatalk unbalanced,
unshielded part of the system. This may be because its cables are
much shorter and all the Seatalk instruments are very close together.
(RL70CRC Plus, Smart Heading Sensor, WAAS-GPS). Even the GPS receiver
built into its antenna has a very short cable because I found a great
little unused place right on top of the helm to port of the crank
handle for the main sheet traveler to put both Garmin and Raymarine
GPS antennas. Coverage through the fiberglass hardtop on the cockpit
is excellent and noone uses the antenna for a grabhandle like they did
when the antennas were initially on top of the hardtop. It's cable to
the Seatalk plastic box with European screw terminals is only about
10" long. None of the Seatalk wires are over 24", making them much
too short to fit an 8 Mhz wavelength.

Indeed. And if the NMEA wires are longer, you can 'break' them by using
ringcores.

Ferrites won't stop the plastic boxes and unshielded cables with
square waves in them from radiating unless you rounded off the edges,
which would trash the data timing.

No problem to round off the edges. I have for instance, put RC networks on
the NMEA outputs. You can see the round-offs but a UART samples the signal
on several (mostly 16) positions within one bit time. So slow slopes on the
signal are no problem. It's not egde-driven but level driven.

Meindert