TNC2.10
 On the other hand, if
you are taking your output from a point close to the discriminator output
(which is the preferred method), then you will be bypassing the deemphasis
circuit.  In this case, the MF10 filter has the opposite effect to what would
be desired.  Incoming mark and space tones which differ by 6 dB (due to
preemphasis) will end up being 12 dB apart after passing through the filter.
The XR-2211 demodulator chip will not like this at all, and will probably
refuse to decode the signal.

Having convinced myself that the filter should be unnecessary, the next step
was to remove it from my TNC (one of the original TAPR TNC2 boards).  This is
very easy to do; you simply unplug the MF10 (it's a 20 pin device) and run a
jumper from pin 3 to pin 17 of its socket.  On-the-air testing confirmed that
copy of most signals was improved by bypassing the filter.  There were some
exceptions of course, but the net result was positive and my MF10 has remained
on the shelf.  I plan to do some further tests to try and determine how many
of the signals I hear are preemphasized and by how much, and how this affects
the readability of the signal.

CONTINUED IN FILE TNC2.3