Bc2000 (for the BCF2000 & BCR2000)

Another question related to manually programming the BCF with regards to
the Micron.  I believe I've come across what others have called
"negative NRPN values", and was wondering how to workaround this.  I
hear one solution is to use 2 different BC encoders per Micron parameter
that behaves like that.

Here's the situation: certain Micron parameters, like for example FX
mix, have 100 possible values, but the BC only can access half of them,
so to speak.

On the Micron, turning the parameter knob scrolls you through all the
values available:

FX Mix
100 % dry, 0% wet ===> 0% dry, 100% wet.

On the BC, using an encoder set to NPRN 65, you can only access the
'positive' half:

FX Mix
50% dry, 50% wet ===> 0% dry, 100% wet.

Taking a look in MIDI-OX, what's listed when you turn the BC knob makes
immediate sense (high to low, leaving out Status and Channel for
clarity):

D1 D2 -- Event
99 0 -- NRPN MSB
98 65 -- NRPN LSB <== #65 is 'FX mix' parameter entered into BC encoder
6 0 -- Data Entry
38 50 -- Data Entry <== 0% dry, 100% wet
...
99 0
98 65
6 0
38 49 <== 1% dry, 99% wet
...
6 0
38 1 <== 49% dry, 51% wet
...
6 0
38 0 <== 50% dry, 50% wet

And that's as far as the BCF encoder goes.  You cannot access anything
past this 50/50 mark.

When you monitor the Micron knob turning for the same parameter, it's
the same for the 'positive' values, but lo and behold, once you go past
the 50/50, the Data Entry parameters for MIDI CC 6 now changes from a 0
to a 127.

99 0
98 65
6 0
38 1 <== 49% dry, 51% wet
...
6 0
38 0 <== 50% dry, 50% wet
...
99 0
98 65
6 127 <== sudden change!!
38 127 <== 51% dry, 49% wet
...
6 127
38 126 <== 52% dry, 48% wet
...
6 127
38 79 <== 99% dry, 1% wet
...
6 127
38 78 <== 100% dry, 0% wet

So, is it possible to figure out what numbers are needed by the BCF to
make it access the 'negative' NRPN here?  I notice that MIDI CC 98 still
shows 65 for the parameter (FX mix).  Perhaps then the key is to
calculate the new numbers needed for the 'negative' range by taking a
look at the Data1 and Data2 numbers.  I suspect it has something to do
with converting to hexadecimal... I see numbers like 16284 and 16383
thrown around on other posts.  Seems to me the difference between "6
127" and "6 0" may be one bit flipped?

(darn, on a hunch, I tried using NRPN 64 but with data going from 973 to
1023, thinking that 64 1024 = 65 0, but it didn't work out... am I on
the right path though?)