2022-09-13.md

2022/09/13 - Microscope development meeting

Attendees

Attending:

• Danail Stoychev
• David Miguel Susano Pinto
• Ian Dobbie
• Veronika Pfannenstill
• Nicholas Hall

Apologies:

• Julio Mateos-Langerak

Notes

• Cockpit experiments fail with connection error

  Veronika and David have an issue in Cockpit where experiments end with a Pyro4 error. Ian suggested that the problem is that the cockpit server has not been setup correctly and never gets the reply from the RedPitaya that the experiment has finished.

  David suggests that checking for correctly setting the client should be done at start of cockpit to prevent this issues. David will open an issue about it.

• Non-int sizes in Cockpit wx code.

  Nick has been having issues with this. Ian recently pushed a change to the touchscreen to fix exactly this issue, may be present elsewhere. Nick will check again and open an issue about it if the issue remains.

• Cockpit memory leak on MacOS fixed (cockpit#815)

  Ian has had this problem for a while. Has recently upgrade wx and the problem seems to be gone.

• Adding generic digital IO device to Microscope (python-microscope/microscope#251)

  Julio and Ian want a simple device that just flips mirrors. This is just about setting a bit on some register. David things that it's not any different from the executor class and should instead implement an ABC for the executor in Microscope. There will be extra logic required in Cockpit and how it affects the light path.

• Cockpit assumes that executor starts with all lines in low

  David and Veronika found that when Cockpit starts assumes that all digital lines start low. This can be seen on the executor debug window. If one exits Cockpit leaving a line high and starts it again, the line remains high (for example, laser might still be emitting light) but the debug window will show it low.

  David will open an issue about it.

• Light sources with a third state beyond enable

  Danny has been working with new light sources and they don't work well with the API of microscope. The reason is that both TTL signal and software control whether light is being emitted. He needs a third state where the device is on and enabled (enable does nothing) but is not emitting light to work with Cockpit. He has something working for him with a new set_state method.

  David thinks that the issue is not on the Microscope API it's on Cockpit interface. There is an indirect third state on Microscope which is via the set_trigger method to control all of these cases but Cockpit does not make use of it. Hence the need to have enable doing nothing. But this is a Cockpit issue and it's the Cockpit interface or backend that should be adjusted. Danny agreed that if Cockpit interface adjusts then Microsocope wouldn't need change.

  The possible combinations of trigger types and modes is not clear but David made a table on another issue recently. David will put that table on the documentation.

• PVCam software trigger missing

  Danny found an issue where sending software trigger are missed if sent before the camera is ready. He has a fix and will open an issue about it to be merged.

• Cockpit experiments with software triggers only

  Danny is building a system where many devices do not take hardware triggers. He's been working on making Cockpit experiments that do not need an executor and only use software triggers. Ian advised to at least use the camera as master device to control light source exposure.

• Linkam stage with Linux

  A group in Physics is interested in using the Linkam stage but they are doing it all in Linux. David is helping them but no one has tried the Linkam SDK in Linux yet. They're not sure they'll buy it yet.

• Thorlabs stage

  Julio has recently reported by email an issue controlling a Thorlabs stage. Veronika had a problem before where it would sometimes crash during the LoadSettings. It was fixed by an unreleased version but that was almost one year ago. Veronika has not tested the official releases since so does not know if the fix is public.

• Vortan laser has issues changing settings when not enabled

  Veronika is finishing adding support for the Vortan laser. However, there is the issue that one cannot set the laser power or trigger type without first enabling the laser and possibly emitting light at whatever was the last used laser power for some time. Vendor says there is no fix and suggested fix of enabling external analog control.

• Prevent stages from moving to limits each time Cockpit starts

  Danny suggests that microscope stages need a separate home method that cockpit can call at start separate from enable. David and Ian have recently made a slightly different change and instead added a method that reports whether enable is likely to trigger a move. See cockpit#809 and python-microscope/microscope#246.

• DeepSIM stages have separate home methods

  The stages in DeepSIM have home methods that can be called manually. This was in discussed with the point above but its moot because neither is in Microscope, they're Cockpit devices only.

• Danny multiple branches

  Danny has been making a lot of changes but are in a state not yet ready to be merged. He will polish them and open issues to get them merged.