RS TDC fix

                                                
                                                          E. Ramberg
                       Summary of RS-TDC work, 8/28-8/29/01. 

I arrived at BNL the evening of Aug. 27.  The next day there was no beam
and the DAQ was relatively free, so I decided to work on integrating the
remaining two TDC's into the readout stream.  These are TDC numbers 30
(crate 19, slot 14) and 18 (crate 20, slot 14).  These have been left out
because if you add them to the DYC chain of 5 other TDC's, then the system
hangs.  During my stay at BNL, I investigated the behavior of both of
the non-working TDC's thoroughly and came to a fairly firm conclusion that
the problem is the termination of REN/PASS signals into the last TDC of the
chain.  Here are some of the tests I performed during my stay:

- I tried to simply put all 6 of the TDC's in the readout to see if it
  would work miraculously.  The readout still hung up.  The problem is 
  that the DYC doesn't send along all of the data in its buffer after every
  event.  In other words, the "EMPTY" light goes out almost immediately.
  If there is a large amount of data going into the DYC - RSMON flasher 
  signals, for example - then the DYC becomes "FULL" immediately and hangs
  the readout.  If there are only headers going into the DYC, then the RSTDC
  system keeps on going until that DYC fills up.  It doesn't seem to fail
  on every event, but I didn't verify that.

- On a non-working chain, with all 6 in the readout, I then individually 
  swapped out every TDC in that chain, to no effect (i.e. there were still
  readout problems).  This means that it really isn't a TDC problem.

- I find that I can cause readout problems in any chain of 6 TDC's by simply 
  putting in a "bad" TDC in the 6th, or last, position.  I tested 3 chains of 
  6 TDC's with several "bad" TDC's.  They all followed this pattern.  Thus
  I come to the conclusion that the real problem is termination of control
  signals.

- I swapped REN/PASS cables between a chain that was working and one
  that was not working.  This had no effect on the readout.  

- I got one of the non-working chains of 6 to work again by finding a spare
  TDC that worked with the termination resistors in.  The other non-working 
  chain still doesn't work.  There are only 5 TDC's in that chain.  I tried 
  all 3 of our spare TDC's in the termination position, with appropriate 
  terminator resistors, and the chain still doesn't work.  

- In my investigation into the termination characteristics of the TDC's, I
  find that there are two types of LRS3377's - revision A and revision B.
  Revision A type has an extra resistor pack soldered onto 3 pins of the 
  reverse side of the control line SIP (R17).  Revision B type does not have
  these.  I verified with Mike Behnke that this should be the case.  I noted
  that the 3 remaining spare TDC's are all rev type B and that the one I put
  in that worked was a rev type A board.  I communicated this and Sasha 
  consequently brought up new spare type A boards with him for his current
  visit.

- Conclusion:  TDC #18 is now in the readout chain.  TDC #30 still is not.
  The new spares should be tried out in the termination position on that
  chain.  I suspect one, if not all, will be found to cause the system to work.