Focus Difficulties		24 Apr 00

    Some problems getting ZFOCUS to converge were reported recently.  I 
suspect that the problem has been with us all along, but was not noticeable
when one used a bright planet.  Currently the only planets not within the
sun avoidance zone are Uranus, Neptune, and Pluto.  This will be the case
until late June.

    The problem is worst around 270 GHz.  There's no problem at 230 GHz, nor
at 346 GHz with the SIS345.  (No other SIS345 frequencies have been tried
as yet.)  Even at 270 GHz, the XFOCUS seems to be OK (but we haven't done much
testing in X).  Focus drift scans show that the problem is due to a variation
of the receiver output with Z position such that a complete cycle is made after
moving about half a wavelength. (See attached plots.)  The amplitude of the
variation is about 5000 counts and is even present if one tries to focus on
blank sky.  Since that's about the amplitude of the Uranus signal, it confuses
the ZFOCUS fitting program. 

    The best way to get the ZFOCUS position under these circumstances, would be
to make measurements at more than three focus steps, or even do a focus drift
scan.  But until Chef is modified, there's no convenient way to process
those data.  Toolbox seems to deal with more than three steps (but not a
drift). But Toolbox appears to assign the data to a step number (1-n) rather
than to a focus position.  The center ("C=...") value of its fit would have to
be converted to a new Z position by hand, using the focus step size obtained by
dividing the focus distance by one less than the number of steps. 

    For the time being, one will have to either accept the default value for
FCZ or tune the SIS230 to a lower frequency while focusing.  Chris Walker
thinks it might be due to a standing wave originating from reflections off the
dewar window.  The window is grooved to suppress such reflections at 230 (?)
GHz.  Perhaps at 270 GHz, the grooves don't suppress the reflections. 


-Bill