...

Why does det.raw.image not work when we enable fex?  Need to set fex→prescale = 1 to get image on every shot, 0 to disable images.  det.raw.image() should still work, but doesn't.  Need to debug.  det.ttfex.image() might be an ROI for the background events (no beam).  det.ttfex.proj_ref() is a projection of that ROI.  det.ttfex.proj_sig() for every event beam. Not sure what det.ttfex.raw() is.

Piranha4

Be sure to scan the Opal section for experiences that may be relevant to the Piranha, e.g. FEB LEDs, fiber powers, etc.

The Piranha has an LED:

• Green: Normal, good state.
• Red: The camera is in an error state.  The output of the 'GCP' command lists the Built-in Self Test (BiST) value, which can be looked up in Appendix C of a recent version of the User's Manual.  'GCP' is executed by the cameralink class at start-up, so both devGui and the DRP will show its results.
  The camera may recover if you issue a Reset ('RC') command or power cycle it.  The LED will also turn red if the camera's internals get above 80 ºC.  For reference, normal operating temperature seems to be around 53 ºC.
• Blinking: The camera is powering up or calibrating.

I have found that in some cases the camera responds to its Three Letter Commands (TLCs) very slowly (upwards of 2 seconds per command) with respect to normal (50 - 70 ms per command).  This seems to be associated with being in External Exposure Mode ('sem 1' having been issued).  Putting the camera in Internal Exposure Mode ('sem 0') restores the fast behavior.  In the slow state, the DRP's Configure is on the edge of timing out, so I've arranged Configure's command sequence to be mostly exectued while the camera is in Internal Exposure Mode and left the switch to External Exposure Mode for last (if we want to operate in that mode).

The User's Manual has several warnings about not allowing the line rate to exceed 1 / (exposure time + 1 µs).  I have found that the camera gets into a bad state when this is violated.  All its parameters and the output of 'GCP' will look correct and normal, but readout will be accompanied with high deadtime.  For example, with a 71 KHz L0InpRate (as per groupca), L0AccRate will be 17856 Hz (1/4 the input rate).  The camera persists in this state until it is power cycled, reset ('RC', which takes 20-30 seconds), or the command sequence 'sem 0', 'set 4000', 'stm 0', 'stm 1' is issued.  The DRP does the latter during Configure in piranha4_config().  

Matt recommends configuring the AxiStreamBatcherEventBuilder with a timeout value near what you expect the latency to be.  If something in the triggering is causing you to drop images, you'll likely see it in the batcher event builder statistics (dropCounts).  If you are dropping images, then eventually those FIFOs will fill up waiting for image data, and then deadtime will be asserted.

devGui is often helpful for seeing the state of the system and testing/debugging.  Unfortunately, it currently issues 'spf 0' at startup, which puts the camera in 8-bit/pixel mode while leaving the FEB's DataMode alone.  Giacamo has requested that image acquisition be done with 12-bit/pixel mode.  Just be sure to have a consistent state before starting to run with devGui.

Restarting IOCs

(courtesy of Tyler Johnson)

...