fermi-ad / controls

Central repo for reporting bugs, making feature requests, managing RFCs, and requesting seminar topics.
https://www-bd.fnal.gov/controls/
2 stars 0 forks source link

CLX57E/NML56C device timeouts #21

Open awattsFNAL opened 9 months ago

awattsFNAL commented 9 months ago

From FAST/IOTA:

We are using a script for automated scans in IOTA. When reading out the machine state we often timeout on the skew corrector coil devices on CLX57E.

awattsFNAL commented 9 months ago

From John Wieland:

The devices are in general in the Sequence N:I{#}{ID}I where # is coil number and {ID} is magnet identification tag. For example, N:I1A1RI is coil one on magnet A1R.

awattsFNAL commented 9 months ago

From Keith:

The device N:I1A1RI is hosted by the NML65C front end as per D80. Mike Kucera is the keeper and Dean Arveson is listed as the backup.

I reached out to Mike and Dean via email to pass this issue to them.

awattsFNAL commented 9 months ago

From Mike:

NML65C is the node that contains the DAC and ADC hardware for controlling the magnet coils. These are combined function magnets and Jianming You, working with Chip, developed an OAC for the combined function magnet control (back in 2018).

If you’re getting a timeout on skew corrector coils I would look into CLX57E (OAC) which provides the derived values control (and maybe touch base with Jianming). With all the recent changes/upgrades to the CLX nodes something may not be playing nice here.

We don’t see anything obvious wrong with NML65C.

awattsFNAL commented 9 months ago

From Jimmy:

There are a lot errors in the log file, due to the following power supplies unreachable: cmts1-magpower1-xcor.fnal.gov cmts1-magpower1-ycor.fnal.gov cmts1-magpower1-quad.fnal.gov

All these do not respond to ping.

From Mike:

Those are power supplies are for the LCLS-II cryomodule internal magnets. Last week they completed their acceptance testing and everything is shut down in preparation for a cryomodule swap at the test stand.

Jimmy commented out those supplies in the front end startup code to see if they were the cause of the "lag".