Goal:
EPICS (software) interlock that will close V1 (main 40m vacuum valve) under abnormal conditions:
1, P1 above 3 mTorr OR
2, P2 above 1 Torr OR
3, N2 below 60 PSI OR
4, V4 closes OR
..........because TP2 slowed down or failed
5, Maglev fails OR
6, Power fails
Condition 1 & 2 has to be bypassable during pump down.
Note: NO COMM. error message disables interlock. You have to reboot by reset.
Also would be nice if we had code and hardware that would:
Turn off PZTs (two sets of steering mirrors, OMC) when pressure is in transition region. (Need relay box and EPICS IO channel).
1) Some code for this ALREADY EXISTS; it may not be right or properly implemented, or handle all the above cases.
The code can be found in /cvs/cds/caltech/state/from_luna
This is bunk. Where is LUNA?
See: /cvs/cds/caltech/target/c1vac1/VacInterlock.st which is a soft-link to code on luna: /opt/CDS/d/epics/apple/Caltech/40mVac/40mVacScipe/dev/src/VacInterlock.st
2) I read through VacInterlock.st for jollies, and added comments. See /cvs/cds/caltech/target/c1vac1/VacInterlock_commented.st
PLEASE START WITH THIS CODE; WE NEED COMMENTS!
3) The EPICS state code used at the 40m is in /opt/CDS/d/epics/apple/Caltech/40mVac/40mVacScipe/dev/src which bugs me. It should ALL be in /cvs/cds/caltech/target . The problem is, as of 2004, one couldn't build EPICS state code on a 40m machine... I think it had something to do with a lack of a Solaris c compiler. I Think this was since been rectified. Can we build EPICS state code on, eg, op440m?
4) Compiling EPICS state code: see (maybe obsolete) instructions in http://www.ligo.caltech.edu/~cit40m/Docs/40mProcedures.pdf
5) Looking in /cvs/cds/caltech/target/c1vac1 I see that (on Aug 31, 2007) the latest saverestore.sav file is dated Feb 3 2006. Does that mean that this vme cpu hasn't been rebooted since Feb 2006?? We have had power outages; is c1vac1 really on UPS? It definitely should be! Or is it that the saverestore facility doesn't work?
6) I just now tried "telnet c1vac1" and got no response. I wonder why. In the past, one could telnet in to EPICS vme cpus.
7) When changes are made and a new VacInterlock.o is built, we need to load it into the vme cpu by rebooting it (correct?). Will this cause any undesireable hiccup in the vacuum controls? We need to be sure about that!
_
Progress (9/7/07):
Steve says that the N2 pressure transducer is NOT getting into EPICS, which might explain why the interlock hasn't been working (since the code in VacInterlock.st seems to be ok).
Meanwhile, can I modify and rebuild VacInterlock.o ?
ssh ajw@luna
source /opt/CDS/c/barker/setup/cshrc.definitive
setup epics/release/r3.12.2
cd /opt/CDS/d/epics/apple/Caltech/40mVac/40mVacScipe/dev/src
epics -nw VacInterlock.st
add comments!! Change N2 threshold from 55 to 60.
cd /opt/CDS/d/epics/apple/Caltech/40mVac/40mVacScipe/dev/target/mv162/obj
make VacInterlock.o
Hey, it worked! Now if I wanted, I could log on to rana,
cp /opt/CDS/d/epics/apple/Caltech/40mVac/40mVacScipe/dev/target/mv162/obj/VacInterlock.o /cvs/cds/caltech/target/c1vac1
Then if I dare, I can restart c1vac1 by pressing the red button on the VME module. It should load the new version. But I worry that it will mess up the vacuum system in some way... First, fix the N2 readback!
_
Test VacInterlock:
ezcawrite C1:Vac-N2pres 50
and indeed V1 and VM1 close. So the interlock works.
ezcawrite C1:Vac-N2pres 70
and re-open V1 and VM1.
Further, the N2pres transducer works, and the correct value displays on the Omega analog-to-RS232 box. So the problem is from this box to c1vac1, or in the software VacNitro.st. To debug this, need to reboot c1vac1 to be able to telnet into it. Hope that saverestore works; else, autoburt restore with burtgooey . We'll try this next week.