Last modified: 7-OCT-1999
There is a very obvious crosstalk amongst the 16 channels associated with a given SAS/SCA chip pair; at about 1.5%, this is large compared to benchtop measurements of order 0.3%. The difference is not understood. However, the crosstalk appears to be larger on some 16-channel groups, than on others.
The apparent crosstalk within groups of 16 channels, as seen in the early TPC tests, was traced to a digitization problem by Chinh Vu and Spencer Klein. By varying component values on the board, one can trade-off between noise and "digital" crosstalk; the latter was reduced considerably at the expense of increasing noise from about 1.05 ADC counts to about 1.35 ADC counts.
The crosstalk between various combination of channels within a group of 16 seems to vary considerably. This needs to be studied systematically and quantified.
Crosstalk signal appear to be partially differentiated version of the real signals; this needs to be systematically studied and quantified.
The crosstalk from FEE charge injection seems to contain a larger differentiated component than does that from cosmic-ray signals. There also appears to be a signal induced by the on-board pulser. This needs to be systematically studied and quantified.