[evla-sw-discuss] tone detection parameters

[Ken Sowinski]

On Fri, 23 May 2008, Michael Rupen wrote:

>   I'm assuming the review will concentrate on general areas of design 
> rather than page-by-page detailed comments.  Many details will of course 
> arise when you read the beastie, but fiddly bits should probably be 
> discussed on evla-software rather than in a big committee meeting.

Herein the first fiddly bit.

This whole comment is, perhaps, moot since we have no pulse-cal
generators on the EVLA antennas and no plans to install any.  It
may be relevant if Pietown is connected or VLBA data is to be
correlated.

If we had implemented pulse-cal tone extractors before the delay
modules then the only parameter needed would be the frequency of
the desired tone relative to baseband.  Instead there is one per
subband after all the delays has been inserted, so the tone
detection, as noted by Brent and Dave, must take into account the
delay and delay rate imposed ahead of the detection.  The VCI
suggests this information is provided from the outside.  Rather,
all the information needed already exists in the station board
CMIB, so it should be able to decide how to drive the tone detector
knowing only the frequency of the desired tone as in the simple
case.  The CMIB must take into account the both wideband and
subband delay and delay rate and, I think, the mixer in the stage 2
FIR if that is in use.

Thus all the information that needs to be communicated through the
VCI is the frequency of each tone to be detected and an integration
time.  That information is relatively static; the polynomial
describing the phase of the tone is unnecessary.