Yesterday at 13:40 and at 22:40 PDT the CW stopped exciting the EX PCAL because its testpoint was cleared from awgtpman. The excitation slot in the awg remained even though the excitation was not active, so the CW sender (psinject process on h1hwinj1) was unaware that its excitation was no longer operational and did not try to restart. When this happened the CDS overview screen did show a red excitation error for h1calex (meaning that an excitation should be present but is missing) and the operator medm overview should have shown a missing CW error as well.
The problem was found to be a "feature" in diag which allows the user to clear every test point on every model with one command. The feature was accidentally acitvated yesterday when the dcu-id and the testpoint were interchanged in the tp clear command.
Jim is working on removing this feature from the current diag code to prevent this accidentally happening again during ER9.
Actually the fix is very intrusive (will result in new binaries for the complete GDS suite of tools; diaggui, foton, awgtpman etc) so we will delay the change until next Tuesday.
In the mean time, within a 'diag -l' session, please do not type the following commands:
tp clear *
tp clear * <dcuid>
tp clear * *
tpclear *
tpclear * <dcuid>
tpclear * *
Take it from us, they will clear all test points on all models and stop the CW hardware injection.