subject: summary of meeting re: special triggers

attendees: crawford, bieser, vu, judd, greiner, klein, engelage,
		schambach (phone)

the discussion centered around the "hardware" specifics of how trigger
would handle special trigger request from the different detector groups.
it was agreed that the current TCD would be slightly modified to have 
3 lemo outputs dedicated to "special trigger requests" an thus allowing
up to 7 different special triggers ("000" is not a valid request).  the
outputs from as many as 8 TCD boards would be input to as many as 3 TLA
boards. the TLA boards would route that info onto two 34 pin connectors
on the DSMI feeding the last DSM board.  The FPGA in the last DSM board
would "or" the 24 bits to determine if a special trigger was being
requested.  that FPGA would also determine which detector was requesting
the special trigger and pass that info, along with the initiating 3 bit
pattern, into the TCU via a 12 bit word (e.g. 0-2 are the three bit
pattern, 3-10 specify the detector, and 11 specifies a special trigger
request).  For simultaneous special trigger requests from more that one
detector system, a priority scheme in the last DSM's FPGA would
determine which request was passed to the TCU.	The priority scheme is
such that the FPGA will look at each set of 3 bits in a pre-defined
order (set in hardware) and select the first non-zero set. Note: the 12
bit word, constituting the special trigger request, will not be sent
from the last DSM to the TCU unless all the physics input bits for the
crossing were zero (i.e. physics takes priority over special triggers).
the TCDs issuing special trigger requests will latch the request (to
last DSM) until it (the TCD) sees the trigger command word (from the
TCU) that services that request. the action to be taken by the specified
3 bits will be determined by what is loaded in the TCU LUTs.

fred agreed to have the modifications to the TCD ouputs made before
jo comes to california.  jo agreed to come to california sometime after
october 18th and before nov 8th to test the system.  fred and jo will
work out the specifics.

eleanor agreed to modify her algorithm for the FPGA in the last DSM to
accommodate special trigger requests as specified above.

hank/jack agreed to discuss how detector systems get "software" requests
for special triggers to the TCD with onl (claude).