From QUINN@UQUARK.PHYS.CMU.EDU Fri Apr  2 12:57:47 2004
Date: Fri, 2 Apr 2004 12:36:50 -0500
From: "Brian Quinn, (Reply to: bquinn@cmu.edu)" <QUINN@UQUARK.PHYS.CMU.EDU>
To: NAKAHARA@JLAB.ORG
Cc: QUINN@UQUARK.PHYS.CMU.EDU
Subject: NA cables "how to"

How to quiet flashing LTD lights (or prevent them from flashing in the first place)

  The LTD clock cables have a problem which acts like a poor ground contact.  
Attempts to locate and fix the problem have been only partly successful at best so we
will have to learn to live with it.

  Even with this problem NORMALLY THERE SHOULD BE NO FLASHING LIGHTS ON THE LTDs.
If lights begin to flash then you should try to treat the problem by manipulating 
cables as described below.  Better yet, if the worker manipulates the cables once 
per shift while doing the checklist, flashing lights may not come up for days 
in a row.  It appears that a relatively small motion of the cables scratches away 
a layer of oxidation and prevents bad contacts from forming.

To manipulate the cables:

  First make sure that either a run is in progress or you have pushed the 'Quinn 
button' to start the fake MPS generator.  (This will cause clock trains to go to the
LTDs.  They will then report any errors they see in the clock trains they are 
seeing by flashing the lower three lights on the LTDs and locking-on the upper ones.)


  Usually it is sufficient to grab the entire bundle of cables going to one of the 
LTD crates (holding it about a foot away from the ends going to the LTDs) and move 
it slowly up and down about an inch (maybe one cycle per second, you don't have 
to shake them rapidly).  At first this motion may cause lots of red 
lights to flash.  Just keep moving the cables.  After 4 or 5 seconds the red
lights should calm down.  This actually renders the cables insensitive to any 
additional motion.  You are not trying to find the 'right spot' to put the cables
in, you are just trying to flex them a bit.

  Then you can grab the bundle of cables coming out of the SDBs and do similarly.
Also do the same for the cables going into the SDBs (at the bottom front of 
the SDBs).  CAREFULLY and GENTLY also move the two cables going into the front of the
KGB.

  In each case you may cause lots of red lights to flash as you begin to move the 
cables but that flashing should calm down in a few seconds.

  Now go to the back of the rack and gently move the 5 cables coming out of the 
back of the KGB and the big bundles of cables coming out of the back of the SDBs.

  That should do it.  All the connectors should be happy now.  If there were any 
red lights flashing when you started, this should have calmed them.  Whether there 
were flashing lights or not, the LTDs should be calm now for 12-24 hours even if 
the cables get moved around.   If you've caused more problems, then it probably 
means you didn't move one of the cables for long enough (at first motion makes worse 
contact.  Then within a few seconds it should 'scrape away' the bad contact and 
make a good contact.   So just go back over the cables again.  In rare cases
you may need to hold one particular clock cable and give it a wiggle near its 
connector.  Be very gentle.  These connectors break off of the boards easily!

  Now you can reset all the latched error lights by pushing the red buttons.  
Be careful not to twist the buttons as you push them in.
A clockwise twist locks them and then the latched-error lights won't function.

Email: bquinn@cmu.edu         Physics Department
Phone: (412)-268-3523         Carnegie Mellon University
Fax: (412)-681-0648           Pittsburgh, PA, 15213    USA