MuSR: 3He Refrigerator Testing - Githubissues

ISISComputingGroup / IBEX

Top level repository for IBEX stories

5 stars 2 forks source link

MuSR: 3He Refrigerator Testing #6243

Closed JamesKingWork closed 2 years ago

JamesKingWork commented 3 years ago

As a MuSR scientist, I would like the 3He Refrigerator IOC to be sufficiently tested against hardware.

Acceptance criteria:

[ ] Test plan and outcomes are documented in this ticket
[ ] Communications to the device is tested against hardware
[ ] Control logic of the IOC is tested against hardware

Notes:

The 3He refrigerator is a Heliox/ITC503 based system
LabVIEW VIs are at C:\LabVIEW Modules\Oxford Software\Muon Heliox

3He Refrigerator tickets:

Document VI control logic: https://github.com/ISISComputingGroup/IBEX/issues/6239
Implement control logic: https://github.com/ISISComputingGroup/IBEX/issues/6240
Communications: https://github.com/ISISComputingGroup/IBEX/issues/6241
OPI: https://github.com/ISISComputingGroup/IBEX/issues/6244
Machine setup: https://github.com/ISISComputingGroup/IBEX/issues/6242
Testing on Device: https://github.com/ISISComputingGroup/IBEX/issues/6243

JamesKingWork commented 3 years ago

They do want to use this for the upcoming cycle and cryogenics have confirmed that they can set up the machine shortly after the software is completed

JamesKingWork commented 3 years ago

Impeded waiting on https://github.com/ISISComputingGroup/IBEX/issues/6294 and https://github.com/ISISComputingGroup/IBEX/issues/6240. We could do an initial test after the communications are reworked and then test the autocondense separately

JamesKingWork commented 3 years ago

The testing for this device is currently being postponed to enable EMU mercury itc and MUSR soak testing

JamesKingWork commented 3 years ago

The current plan is to move back to SECI if this device is needed in the next cycle. Removing it from the board and impeded, will propose again when MUSR want to test

rerpha commented 2 years ago

so far we have been working on OPI: https://github.com/ISISComputingGroup/ibex_gui/tree/Ticket6243_hardware_testing support: https://github.com/ISISComputingGroup/EPICS-hlx503/tree/field_test_fixes IOC: https://github.com/ISISComputingGroup/EPICS-ioc/pull/688

rerpha commented 2 years ago

we may need to use https://github.com/ISISComputingGroup/IBEX/issues/7143 if possible as the state machine seems to be sending commands in sequence too quickly!

rerpha commented 2 years ago

OK, there was a race condition that was causing things to happen in no given order when setting a SP (though mostly just the sorb SP) which was caused by DLYs in sseq that called other sseq records with the same delay, ie

sseq1
 - set this pv
 - delay 0.1 seconds
 - do sseq2 
 - delay 0.1 seconds 
 - set another pv    <---- this might happen before sseq2 finishes as the delay is the same inbetween sseq

sseq2
- set this pv
- delay 0.1 seconds 
- etc

I have bodged this for now by setting the outer delay to be longer (0.5) than the inner delays (0.1 each) added up in total, and this seems to work quite nicely. I don't want to change this without another opportunity to test with hardware but we should probably do something like use a WAITn instead - though this might not wait for the nested sequence to finish which would be something else to test.

rerpha commented 2 years ago

im going to be a bit cheeky and merge the above branches into master so we can have a built copy in case my copied-over version does not work in the morning. hopefully the build completes!

rerpha commented 2 years ago

in a nutshell, everything seemed to work and we managed a recondense as well as skipping through the first part of a recondense. setting setpoints for sorb/he3 pot now work well. I will be setting up on WISH today so will put this in review once they are happy

rerpha commented 2 years ago

repointing - some other things are needed:

It would be better if the threshold for switching between the HE3POTLO and HE3POTHI thermometers was separate to the threshold used for switching between high and low temperature mode. I think we discussed this when we implemented the thermometer switching, but ended up using the same threshold for speed. Unfortunately this didn’t work out during real usage.
There are a lot of communications drop outs for some reason which causes the reported temperature to drop down to zero.
The switching the heater off every time a new setpoint is sent is messing up the PIDs, causing very noticeable temperature drops and hence time delays each time the setpoint is changed (again, see graph).

rerpha commented 2 years ago

actually, i will create a new ticket for this, as the hardware testing has now happened and the other things can be considered separate tasks.

Reviewing this one won't really require much reviewing as WISH used the sorb but would like some tweaks made. I might just close it, but feel free to open it again and review if you disagree

rerpha commented 2 years ago

https://github.com/ISISComputingGroup/IBEX/issues/7168 is the tweaks ticket