Minutes of the MVD evo, March 21 2018


Partecipants: Gianni, Jonny, Tobias, Alessandra, Daniela


Agenda: 

1. brief report of the ToASt development status 
2. strip readout architecture: ToAST, flex, MDC and GBTx
Routing of all these parts and timing
Clock distribution
ASiC configuration


1. 
Jonny present some slides with the picture of the analog blocks under study 
(see the table)
He remindr that the bit value to enable the global calibration circuit is 1.
Discussion about the calibration circuit and its resolution.
Daniela requires again a clear list of specifications for ToAST (written!)


2.
Daniela explain the implemented lines in the two J connectors of the optoelectronic
board designed for the connection to the pixel modules.
Each connector implements 5 e-links for data transmission (@320 MB/s), 
one clock line, 
two configuration lines (in/out) working at 80 MB/s.
(see the slides)

A subdivision of 6+4 e-links can be possible (to be verify anyway). 

With 5+5 and 6+4 we can connect the GBT to the MDC chips of the strips.
There are several warning about the fact that some stave MDCs have to share the 
optoelectronic board with disk MDCs (6 links for the first ones and 4 for the second ones),
of course using one of the two connectors separately.
Is it possible that we meet noise, interference problems? following
such connection layout? To be investigated

It is important to study (to verify) the distribution of the clock line and the 
configuration lines from the J connector to the MDC and then to ToAST chips.
Can we implement a daisy chain or a 'derivative' distribution... ??? 
the answer to this question decides also about some things to be implemented in the ToASt chip !!

However, how to distribute clock and configuration lines has to be studied for both 
staves and ,if we are developing the ToAST for both systems.
The solution of the routing can be easier for the stave, more complicated for the disk part,
because on the stave we have a long unique flex (is it that ?), 
on the disk we have several flex circuits, one for each trapezoidal sensor.
It is important to verify and solve the problem.

Question: did we consider the problem of the squew in the data coming from several ToAST
chips ? in particular concerning the timestamp information and in the hit finding ?