CCWN 75:16            
                                                To synchronize
the filter, the transmitting station sends a string of  dits, and
the phase of the 100 millisecond-spaced control pulses to the 
integrators and sample/holds is adjusted for maximum contrast
between dits and  spaces. 
                LEARNING TO KEY A CCW RIG
        Ray Petit and Andy McCaskey have noticed an interesting
characteristic of  CCW sending. Unlike a standard keyer, which
starts sending a dot or dash AS soon  as the contact is made, a
coherent keyer must follow an established time  reference. For
example, using .1 second bit length, it would be possible that a 
.1 second interval had just begun, so that the dot or dash could
not start until  the next .1 time interval begins. The lag of the
transmitter can be from 0 to .1  second. Therefore, you must send
an average of .05 seconds ahead, or .5 bit ahead  of the
transmitter. This results in a significantly different experience
in  sending. We do not know how important it will be.        Ray
has given some thought to digital circuiting which could remove
this lag  by allowing the operator to hear a signal corresponding
to his sending while the  transmitter is actually an average .5
bit behind his sending.        Woody thinks it could be overcome
by using a standard keyer adjusted to the  appropriate speed. The
operator hears the results of this keyer. The output of  this
keyer is then used to key a CCW keyer. The CCW keyer keys the
transmitter  anywhere from 0 to .1 seconds (average .5 bit )
behind the operator's standard  keyer.  TV COLOR-REFERENCE AS A
FREQUENCY STANDARD        The 3.58 MHz reference frequency on
network color TV broadcasts is being  referenced to a rubidium
oscillator and therefore is a potential very-accurate  reference
frequency. The error between the local oscillator and the
standard is  displayed as a color bar on the TV picture. The
change in color of the bar  indicates the frequency of the error,
and has the interesting property of showing  whether the
difference is high or low in frequency by the order of the color 
changes. It appears accuracies in the order of 10-10 are easy to
obtain.        If you are interested in the circuit and a reprint
about this process,  write Sandra Howe, Time and Frequency
Services Section, National Bureau of  Standards, Bolder,
Colorado, 80302, and request the article "Calibrating  crystal
oscillators with TV color reference..." and other relevant
materials.