[Gb-ccb] forwarded message from Martin Shepherd

[John Ford]

Martin Shepherd writes:
 > 
 > >  > On Wed, 20 Aug 2003, John Ford wrote:
 > >  > > Martin, don't forget about the RFI filters that will be in the line.
 > >  > > they will change your numbers somewhat.
 > 
 > Indeed. The 32MHz Pi-filtered DB25 socket that Spectrum Control sells,
 > has a 100pF shunt capacitance. That unfortunately makes it impossible
 > to arrange for a frequency-independent match over the frequencies of
 > interest. I'll try to figure out how significant the resulting
 > frequency dependent reflections would be on a 100ns rise-time
 > step-function, and if need be, change the matching resistors to move
 > the frequency of the perfect impedance-match away from DC.

Remember that the purpose of the opto-isolated lines is to drive
switching signals, and so Rich's pulse shaping method is probably more
appropriate than creating a perfect match.  I suspect that the
pi-section filter with its ferrite bead will knock down any
reflections.

The continuum signal lines will be bandwidth limited by the bessel
filters, and so should be no real problem to handle.

 > 
 > > No, they are still necessary to filter the lines out of the box.
 > > *every* line in and out of the box must be filtered.
 > 
 > Sorry, I guess that you have said this before. I wasn't thinking.
 > 
 > Martin
 > 
 > _______________________________________________
 > gb-ccb mailing list
 > gb-ccb at listmgr.cv.nrao.edu
 > http://listmgr.cv.nrao.edu/mailman/listinfo/gb-ccb