Version for API 1.6.0 or 1.8.0 and Eiger1 or Eiger2 #38
Conversation
This is to make sure git tracks the rename before the file is split into to child templates
Fix units on detector distance Add new parameters * Header Detail * Counting Mode * Auto Summation * Initialize Command * Image Bit Depth
In SIMPLON API version 1.8.0 as of EIGER2 software version 2020.1, some (all?) API JSON responses that had previously returned a value_type of "string" now return a value_type of "list". To accommodate this, add "list" as a synonym for "string". Example: Before installing EIGER2 2020.1: ---- $ curl http://10.42.41.10/detector/api/1.8.0/status/error {"value": [], "value_type": "string"} ---- After installing EIGER2 2020.1: ---- $ curl http://10.42.41.10/detector/api/1.8.0/status/error {"access_mode":"r","value":[],"value_type":"list"} ----
In SIMPLON API version 1.8.0 as of EIGER2 software version 2020.1, some
(all?) API JSON responses that had presumably previously returned an
empty response now return an empty JSON string (i.e., two double-quote
characters). To accommodate this, treat an empty JSON string the same
as an empty response.
Example:
After installing EIGER2 2020.1:
----
$ curl -X PUT -H 'Content-Type: application/json' -d '{"value":"disabled"}' http://10.42.41.10/monitor/api/1.8.0/config/mode
""
----
The $(P)$(R)BitDepthImage_RBV PV is a longin record for the bit_depth_image SIMPLON API parameter, with DTYP=asynInt32, but the INP parameter is EIG_DCD_bit_depth_image which indicates an Asyn type of asynFloat64. This causes the EigerParam::fetch(double &, int) member function to be called to read the SIMPLON API parameter. The member function expects a JSON value_type of "float", but the bit_depth_image parameter JSON value_type is either "int" or "uint", so the member function fails with an error indicating that the JSON type is unexpected. Change the Asyn type of the bit_depth_image parameter to asynInt32 to allow it to be read without error.
The mDCUBufFree parameter is only created in the eigerDetector constructor if mEigerModel is EIGER1, but the mDCUBufFree parameter is fetched unconditionally later in the constructor which will result in a crash if mEigerModel is not EIGER1. Fix this by only fetching the mDCUBufFree parameter if mEigerModel is EIGER1.
The SIMPLON API for the compression parameter has changed over time, in particular the GET response "allowed_values" value. In 1.6.0, it is ["lz4", "bslz4"]. In 1.8.0, it is ["bslz4", "lz4", "none"] (i.e., order changed and "none" added). In 1.8.0 after the EIGER2 2020.1 release it is ["none","lz4","bslz4"] (i.e., order changed). Accommodate all three variations of the API. Example: 1.6.0: ---- $ curl http://10.42.41.10/detector/api/1.6.0/config/compression : Due to lack of access to a 1.6.0 SIMPLON API for testing, the full : response for the query is unconfirmed, but it is believed to include : the following: {"allowed_values": ["lz4", "bslz4"]} ---- 1.8.0: ---- $ curl http://10.42.41.10/detector/api/1.8.0/config/compression {"value": "bslz4", "value_type": "string", "allowed_values": ["bslz4", "lz4", "none"], "access_mode": "rw"} ---- 1.8.0 after installing EIGER2 2020.1: ---- $ curl http://10.42.41.10/detector/api/1.8.0/config/compression {"access_mode":"rw","allowed_values":["none","lz4","bslz4"],"value":"bslz4","value_type":"string"} ----
Added code to determine the API version using the Simplon API. Added a public method to return the API version.
…command. Added new eigerModel_t enum. Added new private member variable holding the Eiger REst API version.
The eigerModel is now determined by parsing the "description" parameter from the Simplon API. Improved the logic so the API version and the detector model (Eiger1 or Eiger2) are decoupled. Some things depend on the API version and some depend on the model. Changed so ADFirmwareVersion is the firmware version, not the API version. Change so ADSDKVersion is the API version. Previously it was unused.
|
Hi @MarkRivers An issue I ran into was the EIGER2 X 1M detector failing to correctly disarm when trying to interrupt a running series of exposures. We run a mode that we call "freerun" scans where we set (in the API language) the trigger mode to manually-triggered internal series, with This was working prior to 2020.1 (I think) but when I asked Dectris about it they suggested to add a Is it included (or possible to include) into this ADEiger version to send CANCEL then DISARM when issuing a Acquire=0? Thanks, |
Theshold2Energy, Threshold2Energy_RBV Threshold2Enable, Threshold2Enable_RBV ThresholdDiffEnable, ThresholdDiffEnable_RBV CountrateCorrEnable, CountrateCorrEnable_RBV ExtGateMode, ExtGateMode_RBV HVReset HVState_RBV Added "External Gate" choice for TriggerMode and TriggerMode_RBV
This reverts commit 43644cd.
Threshold2 (threshold 2 energy) Threshold2Enable (threshold 2 enable) ThresholdDiffEnable (enable computing difference image of threshold1 - threshold2) Add External Gate option to ADTriggerMode (extg option; in next firmware release) ExtGateMode (external gate mode; in next firmware release) HVResetTime (high voltage reset time) HVReset (reset high voltage for the specified time) HVState (current state of high voltage) NumExposures (nexpi parmaeter; in next firmware release)
|
@MarkRivers said:
I have now added support for 2 energy thresholds, and I have model-specific medm screens and auto-converted OPI screens. I have also added support for ExternalGate feature that will be in the next firmware release in early 2021 but is already available for testing. I just tagged R2-7 which had all changes prior to adding Eiger2 support. I think this branch is now working well enough to be merged into master, so I am doing that now. |
|
I am close to being ready to tag a new release. I would appreciate if people can test the master branch on both Eiger1 and Eiger2 detectors. Remaining things to be done:
|
This PR is based on the work by Gary Yandel at DLS, with some improvements.
There is still additional work to be done:
Since this version appears to work with both APIs and both models it can be merged into master while additional work is done on another branch.