List of known firmware bugs (or desired changes)

Last updated on $Date: 2000/06/19 14:52:05 $ (UTC)

• To permit barrel-crossers to be handled correctly in the Track Fitter, we need to reverse the hit coordinate system in the odd half-barrels

• One ISPY address goes unused before each new event

• Introduced some time around October

• Reproducible in Quicksim

• SVT_INIT doesn't reset FIFOs in Load mode

• Introduced because I decided to synchronize the INIT signal inside the HitMan chip

• Still works in Run mode

• Logic bug in Merger's handling of "hold" signal when in "Merger test" mode

• "fifotest" works with original compilation of hitman but not with new compilation that makes E2 an input

• This is a useful test when board is known to have problems; would be nice to make sure it works

• DIP switch map needs to be rationalized in firmware, since we inherited screwy bit ordering from L1 component geometry

• Would make visual interpretation of board ID switches more obvious

• We decided that we no longer want to go into run mode on SVT_INIT

• Simple Boot change

• Stream-disable feature causes lost-sync error when stream 0 is disabled (others OK)

• Pedestal RAM bug (now fixed)

• Introduced in October's pipeline optimization

• Is reproducible in Quicksim

• Symptom is that non-zero pedram contents yield bogus output

• Different versions of hmaim.tdf in prototype consistently showed or didn't show problem; it's definitely a firmware issue

• For final prototype tests, we reverted to pre-optimized version of hmaim.tdf, and have the option of doing so for production

• Fixed 2000-02-05; one-line change in hmaim.tdf, to clear valid flag if pedestal subtraction underflowed

• For prototype, DS LEDs were stretched digitally by MOP; this is no longer needed

• Removed 2000-02-03