Errata, Chapter 5, EMRFD 
  (10March11)

  p 5.4, left column 2.5 inch up from text bottom:   Replace the word "average" with "mean."   (21March10)  

p 5.8,  left column, 1.8 inch down from text top:   Change the difference expression (Frf-Fif) to a product (Frf*Fif).   (21March10) 

p 5.15, middle column, 3.5 inches down from top of text.   Add a sentence right after TUF-1 and SBL-1.   The sentence should be:   "Additional information on mixer operating levels is found in the discussion of phasing transmitters in Chapter 9."   (10March11)    

p5.18, Fig 5.46 and related discussion:   This is informational and there is no error that needs fixing.   HC series flip-flops are used in several places throughout EMRFD as frequency dividers to generate square waves that then provide harmonic outputs.    In this discussion we used some really old data suggesting that the output impedance of a CMOS part was around 1K.   That was the end resistance used for filter design.   We recently did some measurements with a HC inverter where a variety of loads were placed on the part.   I measured an output R of 39 Ohms for a single gate.    This drastic difference means that the bandpass filter will have different components and should be redesigned accordingly.   There is no need to change any of the example circuits, for they were all checked with a spectrum analyzer when they were constructed.
     There can be a problem with this low impedance:  When we see a low value, we might be tempted to think that we can do an impedance match to this value and extract a lot of power.    That is probably not true, for there could be stability problems.  Heavily loading the HC output could cause high currents to flow internally in the IC, causing it to exceed power ratings.  A heat sink would then be required.     Loading will usually be light when using the chips for frequency multipliers, for the usual bandpass filter will attach to the IC with a small series capacitor.   This will present a low impedance at the harmonic, but a relatively high Z at the fundamental.    The IC will not be heavily loaded by such a circuit.    An ideal way to design might be to place a 220 Ohm resistor in series with the output.   The composite output impedance would then be about 250 Ohms.  If this value is used for filter design, the result would be worst case 500 Ohm load on the IC which should not cause problems.     (25Oct06)        

p5.21.  Reference 8.   Chris Trask informs us that he did not write this early data sheet for the NE-602.  The data sheet in my files has no listed author, but I'm pretty sure it was Bob Zavrel, W7SX, for he did much of the early Signetics applications work on that part.  That was an era when application notes just appeared from the mist and didn't need anyone to write them.  (24Feb03)  (This has NOT been confirmed with W7SX.)