Message	Date	From	Subject
1820	2008-07-19 17:23:10	Glen Leinweber	AD8307 power meter shielding
Was re-checking amplitude calibration on my homebuilt wide band power meter, with a 7 MHz signal. The AD8307 and op-amp were built into a soldered-up box, with a BNC input jack, and three wires for power, gnd, DC output. Power and output are ferrite-beaded and bypassed just inside the box. The processor and A-to-D and display are on a separate board. Builders were cautioned that shielding and bypassing are critical, because of the high sensitivity and large bandwidth. Well, my efforts were insufficient when testing the soldered-up module on the bench. Linearity was fine down to -45 dBm, but below that was poorer than advertised. Unstable DC output with no input signal suggested that stray RF was getting inside. A clamp-on RF choke was added to the coax between generator and power meter. Ah, linearity extended lower. Another clamp-on choke around the three output wires extended linearity lower still. The clamp-on chokes extended linearity down to -70 dBm., very close to the advertised limit. This is another example where a 360-degree shielding box was compromised by having to connect power and output signal wires inside. This instrument can really test your ability to keep stray RF out, and the cautions about shielding are well-advised. Once re-installed into the complete box, perhaps the overall shielding was adequate, but that'd be a dangerous assumption.

Was re-checking amplitude calibration on my homebuilt
wide band power meter, with a 7 MHz signal.
The AD8307 and op-amp were built into a
soldered-up box, with a BNC input jack, and three
wires for power, gnd, DC output. Power and output
are ferrite-beaded and bypassed just inside the box.
The processor and A-to-D and display are on
a separate board.

Builders were cautioned that shielding and bypassing
are critical, because of the high sensitivity and large
bandwidth. Well, my efforts were insufficient when
testing the soldered-up module on the bench.
Linearity was fine down to -45 dBm, but below that
was poorer than advertised. Unstable DC output with
no input signal suggested that stray RF was getting
inside. A clamp-on RF choke was added to the coax
between generator and power meter. Ah, linearity
extended lower. Another clamp-on choke around
the three output wires extended linearity lower still.
The clamp-on chokes extended linearity down to
-70 dBm., very close to the advertised limit.

This is another example where a 360-degree shielding
box was compromised by having to connect power and
output signal wires inside. This instrument can really
test your ability to keep stray RF out, and the cautions
about shielding are well-advised.
Once re-installed into the complete box, perhaps
the overall shielding was adequate, but that'd be a
dangerous assumption.