Fault Detection and Camera Status

[Olliebrown]

While camera hardware error situations have not yet been observed, we anticipate they will happen eventually! This is because we have observed them with other programs that utilize the EDSDK (like digiCamControl and the included Capture program with the SDK).

There is a need to manage these situations and work around them, something that is not possible with the closed source tools but is possible for our custom server.

Considerations for this task:

• HW errors seem to arise from an improper understanding of the SDK commands
• The way that the 'shutter' button works is a good example: after pressing either fully or halfway from the SDK, it stays pressed/held until explicitly set back to 'none'. While in the held state, the camera behaves very differently.
• We can also approach things with verification: ensure commands have reasonable timeouts whenever issued and that all possible error responses are captured and relayed to the client.
• We may want to store cameras that have responded with errors so they can be queried by the client and detected before they become critical
• We may also want to "push" this information with a socket message instead of waiting for a client to poll for the info
• Lastly, once we have some errors states that are reproducible, we should look for software workarounds that allow resetting the camera or the SDK to get back to normal operation

[Olliebrown]

Once these situations are observed, please comment here with more details and consider documenting the specific situation as a separate 'issue' that we reference in a comment here.

[SamsonSmithC]

During a capture the cameras were primed via Trigger box and left in this state longer than usual. After roughly a minute of being primed some of the cameras suddenly triggered. All cameras then became unresponsive, displaying the 'ERR_DEVICE_BUSY' error until a full restart of the server & server monitoring program was performed. No further unusual behavior was observed. Issue has yet to be replicated.

The only known deviation from normal usage was the use of an inconsistent wireless mouse held by capture subject to control the C4 client. Will investigate potential of bad user inputs to C4 client causing issue.