Log entry time 18:00:38 on March 08, 2011

Entry number 351061

This entry is a followup to: 351053

To make the L1A come earlier than retiming, I did two modification:

1, Make the retiming signal goes into another discriminator channel
before it is sent into RT, so basically now retiming signal comes ~10ns
later than before. When I determined the delay values (hence the length
of delay cable) of retming signal, I made sure it comes about 50ns
earlier than the leading edge of ADC signals. I believe giving up this
~10ns won't cause any problem but it is needed to be checked latter with
cosmic data or real data.

2, L1A signal was sent into delay module in crate2.slot17, then sent into
RT. Although I set the delay value to be zero, it actually cause about
~8ns delay by the module itself(+delay by two short cables, so totally
~10ns.). I removed the cable that sends the L1A into Delay, and directly
send L1A signal into RT without going through the delay module.

After those two change, from the scope, I can see that the L1A comes
~20ns earlier than retiming, what makes the Retiming works again.