Note: this modification was developed by W8JI for his FT-1000D. Earlier
 production FT-1000MP's may benefit from implementing this mod as well,
 since similar circuitry was used.
 Warning: This mod involves soldering in very tight places - if you don't 
 have the experience or equipment, do not attempt this or you really do
 risk damaging your radio.
    This mod greatly improves strong signal IMD in the FT-1000. I used a
 two kHz test spacing in the measurements, so the 12 kHz wide 70 MHz
 first IF filter did not make results look artificially better. I ran two
 HP generators through a combiner into the front end, and measured IMD
 with a selective level meter with 50 Hz BW and also with the 1000's
 internal meter (which is about 2 dB per S unit below S6, not six dB as
 claimed). This was the primary source of IMD in my brand new FT-1000,
 and the difference was dramatic. Remember this is a sample of one unit.
    I carefully looked but I never found anything wrong other than the fact
 the receiver forward biases the first FET in the noise blanker fully-on
 when the blanker was shut off, and that IMD feeds directly back into the
 8 MHz IF stages. The result was a few volts of peak RF voltage on the
 FET drain, and considerable amounts of that voltage feeding back to the
 IF stage through stray coupling.  I am working on better mods to the front 
 end, but this is a substantial change. Be sure to use common sense and 
 turn off power when soldering, etc.
 1.) Remove the bottom cover, locate the noise blanker on the left rear
 corner of the IF board near the first group of 8 MHz filters.
 2.) Locate TP -2001. Connect a 1N914 or any other silicon switching
 diode through a 10k ohm resistor to ground from that point, cathode
 towards the resistor and ground, anode towards TP-2001. DO NOT use a
 high leakage diode like a germanium diode.  (Note: from N6BV: the
 FT-1000MP uses the same designation, as I recall, for the TP-2001 on its
 main board schematic.) (Note from N1EU: I put a solder lug under the
 adjacent pcb screw for a ground connection.)
 3.) Connect a pair of common cathode connected diodes (same type) with
 one anode to pin 1 (brown wire) and the other to pin 2 of J2001. I just
 pushed the wires down in the pin of the plug and tacked them in place
 with a tiny bit of solder. You can verify these are the correct pins by
 measuring the voltage when switching the NB and wide NB buttons. Either
 or both NB switches "on" applies about 8 volts to either or both pins.
 (Note from N1EU: I think it's much easier to tack solder the diodes
 directly to D2094 and D2095 per the schematic below - see photo below.)
 4.) Connect the common to the diode cathode and 10k resistor junction
 through a short hook-up wire.
 This mod removes forward AGC bias from Q2003 and 2004 when the noise 
 blanker is off, and prevents IM distortion fed back via C2041 and foil 
 traces from impacting the receiver when the noise blanker is off. It 
 does not affect NB performance when installed.