Click for Icom America website.

A DSP IF filtering test any IC-756Pro/Pro2/Pro3* owner can perform

 by Rob Peebles, W8LX, Dublin, Ohio May 4, 2002

" [The 756Pro's 15 kHz roofing filters] are not  “narrow filters” by classical analog-radio standards. Although their primary purpose is image rejection, they are adequately narrow to ultimately prevent overload of the DSP A/D converter(s), and that is all that is required."  (From George, W5YR's "Notes on roofing filters")

Try this test to convince yourself that the above is true. 

  1. Take your IC-756PRO (or 756Pro II or 756Pro III) and set it up in the CW mode, with a bandwidth of 50 Hz (BPF indicator illuminated).
     
  2. Tune in WWV at a time when the signal is strong (currently 20 over S9 @ 15 MHz). When you are zero beat perfectly (i.e. sidetone pitch matches beat note from the WWV carrier) all you hear is the carrier (no modulation, voice etc.). 
     
  3. Now tune up 100 Hz (due to calibration errors, it might be easier for you to just move up with the RIT instead of the main VFO after you find the true zero beat) and listen to the BCD time code. Notice how well you can hear the bit pattern of the time code (and how you don't hear much of anything else). Set the AGC to fast and watch the digital S meter, paying particular attention at the top of the minute when there is no 100 Hz tone transmitted. Pretty impressive for being 100 Hz away from that strong carrier!
     
  4. Now go up to 150 Hz above the WWV carrier....don't hear a thing. See how low the noise floor is?
     
  5. At 200 Hz above the WWV carrier we see the harmonic of the 100 Hz BCD. At 250 Hz above it is spooky how much of the signal you don't hear!
     
  6. Now, go up to 500 Hz (even minutes) or 600 Hz (odd minutes) and listen to the standard audio frequencies. Pay attention to the ramp of the tone coming and going in conjunction with the ticks (during the period of the tick the 500 or 600 Hz tone is briefly muted).
     
  7. 1000 Hz is the fundamental frequency of the ticks. At that offset you see the ticks very nicely but nothing else - well unless it is an even minute then you hear the harmonic of the 500 Hz tone being transmitted!
     
  8. By tuning 10 MHz, I can barely hear WWVH in the background. WWVH's top of the minute tone is 1200 Hz, and I can isolate it in the passband too (it is so weak it doesn't move the S meter). I hear it much better on 15 MHz (where WWVH is stronger), and on 20 MHz (where I can't hear WWVH) it's gone.

When you tune through WWV in such a manner and see how you can pick apart and separate the various components of their transmission without overload from the main carrier, you get a pretty good feel for what kind of "close-in" ultimate selectivity and noise floor can be achieved with DSP filtering. What we have here is, in effect, the aural equivalent of a spectrum analyzer set to a narrow resolution bandwidth.

The above test procedure should enable you to try this test in your own shack.

I've just re-convinced myself I don't want to go back to life before DSP.

*Editor's Note: This test is also applicable to the IC-7800, IC-746Pro, IC-7400 and IC-7000.

 

Contents copyright © 2002 Rob Peebles W8LX. All rights reserved.
Last revised: 06/15/2018
A. Farson VA7OJ/AB4OJ edited W8LX's original text, and created this page, with his kind permission.