Open sghose opened 5 years ago
CC @tacaswell
Depending on how this is implemented this may require a release at XPD.
depending on the purpose of the scan, my interpretation of this scan can be either of the following:
t-series at each spatial point, iterate through points
independent t-series while scanning at spatial points
seems more aligned with the description but it doesn't guarantee your data will be collected at the consistent spatial points.
@sghose is that what you have in mind?
@chiahaoliu independent tseries while scanning at spatial points is the one we need. @CJ-Wright I don't understand what do you mean by a new release? If it is that difficult might not be possible by 22nd.
@tacaswell Tom is working on the BS plan for this setup, we will be testing the same on Thursday and Friday. If the BS plan works we will then step up the implementation with xpdAcq
Once the device is ready, BS plan is ready we would like to use the xpdacq & xpdAn protocols for data collection and Analysis.
If we have issue with BS plan, then we could run the device epics and start tseries and position scans with tseries. All I need from xpdAn is to write the output files with time stamp and position information.
There is a potential way to do this without a release (use the save server template to take the timestamp from the timestamps['pe1_image']
data in the events, which should be correct). The main issue is that the analyzed events (which have the data we want to save) have timestamps associated with the time of the analysis. But we might be able to access the original timestamp via the combined raw + analyzed data stream.
Please try to help us on this on Friday if Possible.
@sghose you can run this test yourself when you have a few spare moments. You need to modify the save server template
save_server --template='{base_folder}/{folder_prefix}/{start[analysis_stage]}/{start[sample_name]}_{event[timestamps][pe1c_image]}_{__independent_vars__}{start[uid]:.6}_{event[seq_num]:04d}{ext}'
Once you are ready take a few shots and see if the timestamp (which will be in seconds since epoc) matches the raw data events
Note that you will need to change the [pe1_image]
to be the detector's readback name (eg pe2_image
)
Users need to collect data and use the pipeline using tseries for about 5-10 minutes for every 1 sec interval while scanning the sample in ID for 10 positions. The output data file should contain the time stamp as well as position value.
Expected Behavior
We will run xrun with a sample and plan. Plan contains a tseries for the whole time (5- 10 min) while scanning through several positions ( Pos_Start, Pos_stop_Pos_interval) back & forth until the tseries ends.
Possible Solution
user a position change plan and tseries plan to combine to a master plan and run with sample. Have the xpdAn handle the file naming.
Steps to Reproduce (for bugs)
1. 2. 3.
Context
We have a flash sintering setup commissioned at XPD beamline which applies a electric field continuously upto a Current limit reached ( runs for 5-10 mins total). This is EPICS ready but not Ophyd ready( DAMA is working on it). We will run this filed in parallel with xrun ins a pseudo automatic run. Once we have time stamp in the file and position we can later match them with the Filed value from time stamp. In future DAMA group will make it a device to run in synchronized with xrun.
Priority
We have a user starting on 22nd of March (this week), So, all we need to write the files with timestamp and position. Should be a quick way.