address
[Reported by Slawosz, EPC]
During the commissioning of systems for the LHC and testing of power converters in our test zone A7, our colleagues from EPC discovered an interesting failure mode of the NanoFIP address while disconnecting the WorldFIP plug.
In short:
• when the WorldFIP plug is removed slowly with the top pins first, it causes the disconnection of the ID dongle but not the communication and the FGClite can pick up the data for another converter. This happens with addresses that have "0" on the 3 least significant bits. • we reproduced the problem on the R2E-600A in A7 with the address 17, 18, 19... • we reproduced the problem on the LHC60A in A7 with the address 8 ('01000') and the problem does not appear on the devices with the address 7 ('00111') as the WorldFIP ID's least significant bits don't change. This particular problem comes from the design of the NanoFIP in which the address changes are read directly from the connector without any specific filtering.