Mercury IPS

[Tom-Willemsen]

As a cryogenics support-person, I would like a Mercury IPS driver to be available and fully tested (including against the hardware) in IBEX.

The old IPS controllers are now being phased out and replaced with Mercury IPS controllers. These are available to be swapped onto beamlines if and when the old IPS controllers fail (Oxford instruments will no longer repair old IPS controllers).

Because these Mercury IPS controllers will turn up on IBEX beamlines at extremely short notice at some point in future, it would be useful to have an IOC ready (and tested with the real hardware).

Manual is available on share: \ISIS_Experimental_Controls\Manuals\Oxford Instruments Mercury IPS

[kjwoodsISIS]

/remind me Add_to_Sprint 28th May

[Tom-Willemsen]

Cryogenics are running this device in the magnet lab tomorrow - it may be useful for someone to go and see how the device is operated to give us a better idea of what is needed for this ticket.

[Tom-Willemsen]

@davidkeymer and I went to see the Mercury IPS running in the magnet lab to talk with cryogenics about it and get an idea of how it's operated.

• There are three power supplies in a rack. These are arranged in a master/slave setup, and we only need to talk to the master power supply. In effect, we can treat them as a single power supply unit.
• Based on the tests that cryogenics section have done, the Mercury IPS hardware is unreliable. All of the following issues are seen while operating the magnet purely from the front panel (no external control software):
  • The firmware will sometimes crash. To reset it the whole power supply needs to be power-cycled. This is obviously undesirable for a magnet power supply, and cryogenics are chasing OI about this issue.
  • The switch heater occasionally reports that it's ON when it's actually OFF
  • The power supply reports a voltage of ~9000V which is incorrect (a sensible voltage for this power supply would be around ~8V)
• The rate limits at a given field are defined in the controller, in the same way as the old-style IPS. We do not need to set the maximum rates for a given field from software
  • Like the old IPS, there is also a global maximum ramp rate, but the field-specific limits will override this if it's too fast.
• This PSU can be operated using the same logic as the existing mercury IPS. The heater wait times/ramp rates/voltage stability criteria will be the same.
• When we are ready to test, cryogenics can set up the Mercury IPS in a safe configuration to test things (the current leads can be shorted).

We should set the Mercury IPS up with a similar interface to the existing IPS IOC.

[Tom-Willemsen]

On closer inspection of the manual, the mercury IPS has two allowable command sets:

• An SCPI-based command set which is "recommended" as it potentially gives extra information (but I'm not sure we are hugely interested in the extra information it provides)
• A "legacy" command set, which it claims is backwards-compatible with old IPS models

I have had a look through their definition of the "legacy" command set and it looks like the commands are the same as those we currently use to drive the older IPS. I think it will be worth trying this approach as it would potentially save us a lot of work if the existing driver already works with these newer power supplies!

Edit: impeded for now; I have sent an email to cryogenics asking if they can set up a mercury IPS in a safe configuration so that we can try the "legacy" command set as noted above.

[Tom-Willemsen]

Also take someone else down for the tests

[Tom-Willemsen]

It was hoped that this magnet would be made available for some tests towards the end of summer. As there has been no sign of this happening, putting this ticket back into the backlog for now.

[Tom-Willemsen]

PRs:

• Mercury IPS mods EPICS-IPS#7
• Mercury IPS mods EPICS-ioc#757
• Ticket4224 #7523

This was tested with shorted leads by cryogenics (with help from @davidkeymer and I) and was found to be working appropriately. It is now going to be used on ZOOM for experiment starting 10/12/2022.

Note: I have added the built version that was tested on cryolab to \\isis\shares\ISIS_Experiment_Controls\data for tickets\Ticket4224 in case it makes patches on instruments easier. Note that the patches contain some changes to the compiled SNL code, so does need rebuilding - it is not sufficient to just change DB & protocol.