Do not attempt to do anything you have not been trained to do by ePSIC staff. |
Please do not attempt to change any settings on the Merlin/Medipix software other than those explicitly referred to in this page. Doing so may produce unexpected results! |
There is a qt GUI on TEMCentre PC that enables Users trigger 4DSTEM data collection on Merlin detector, also capturing the state of the microscope in an hdf file.
The steps to set this up:
Here is the structure of an example metadata file:
{'experiment:NXentry': {'data:NXdata': {}}, 'metadata': {'A1_value_(kV)': 3.869999885559082, 'A2_value_(kV)': 1.1100000143051147, 'aperture_size': 1, 'current_OLfine': 33673, 'deflector_values': {'ANGBAL': array([32768, 32768], dtype=int32), 'CLA1': array([15776, 8032], dtype=int32), 'CLA2': array([31487, 29314], dtype=int32), 'CLS': array([33245, 29480], dtype=int32), 'Correction': array([44800, 44800], dtype=int32), 'GUNA1': array([28329, 35414], dtype=int32), 'GUNA2': array([31312, 34299], dtype=int32), 'ILS': array([32768, 32768], dtype=int32), 'IS1': array([32768, 32768], dtype=int32), 'IS2': array([32768, 32768], dtype=int32), 'MAGADJUST': array([10752, 10752], dtype=int32), 'OFFSET': array([32768, 32768], dtype=int32), 'OLS': array([32768, 32768], dtype=int32), 'PLA': array([33332, 31092], dtype=int32), 'ROTATION': array([32768, 32768], dtype=int32), 'SCAN1': array([15776, 8032], dtype=int32), 'SCAN2': array([31487, 29314], dtype=int32), 'SHIFBAL': array([26944, 25792], dtype=int32), 'SPOTA': array([30768, 38788], dtype=int32), 'STEMIS': array([32189, 29485], dtype=int32), 'TILTBAL': array([45056, 45056], dtype=int32)}, 'field_of_view(m)': 2.0000000000000002e-16, 'ht_value(V)': 15000.0, 'lens_values': {'CL1': 33536, 'CL2': 0, 'CL3': 29952, 'CM': 33059, 'IL1': 24576, 'IL2': 12288, 'IL3': 23631, 'OLC': 44636, 'OLF': 33673, 'OLSF': 2048, 'OM': 0, 'PL1': 54016}, 'magnification': 1000000, 'nominal_camera_length(m)': 0.04, 'nominal_scan_rotation': 0, 'spot_size': 6, 'step_size(m)': 7.812500000000001e-19, 'zero_OLfine': 33673}} |
If 4DSTEM data is collected using the GUI, follow the following steps to convert to hdf5:
python batch_convert_cluster.py e02 <year> <session_ID> |
This should start multiple cluster jobs, converting mib files into hdf5. The output appears in /dls/e02/data/year/session_ID/processing/Merlin.
Under Config Files, browse for the command file: Desktop\Command files\Live and press Run. In a few seconds live display should be active under Image (Start Acquisition button turns red and shows Stop Acquisition). Note that this script removes the File Saving option. The live image is set to a frame time of 10 msec with 6-bit counting depth. If any of these parameters require changing, stop the acquisition, apply the changes in the Image tab [**Frame Time and Counting Mode (depth) only to be changed**] and restart the acquisition.
Using live mode, decide on the suitable frame time (minimum 1 msec, no practical maximum) and counting depth for your acquisition (1×2, 2×6, 2×12 or 24-bit). Note that only one of the two counters will be in use (Counter 0, except in 24-bit mode where they will be filled sequentially). Note that settings can only be changed while the acquisition is stopped.
To record a single frame:
This mode of acquisition is suited for capturing a series of frames in TEM mode. In Continuous Mode, the two counters are read alternately with no gap time and written to disk.
Press Start Acquisition
In this mode, the acquisitions are triggered by the clock from the JEOL scan generator. Firstly, set up the live view, as described in (1), position the beam as required and select a suitable counting depth. The counting depth will determine the minimum dwell time in your STEM scan and the maximum size of your scan array (see Table 1).
Table 1- Acquisition conditions that have been tested
Counter bit depth | Fastest scan dwell time (μsec) | Scan array | mib file size (GB) | Running condition |
---|---|---|---|---|
6 | 650 | 512x512 | 16.9 | headless |
6 | 500 | 512x512 | 16.8 | headless + RAW |
6 | 500 | 1024x1024 | 67.5 | headless + RAW |
12 | 1000 | 512x512 | 33.65 | headless |
12 | 1000 | 1024x1024 | 134.6 | headless |
note there is some variation run to run on the max frames, perhaps +/- 5%
Counter bit depth | Scan dwell time (μsec) | Max Frames | Max Scan array | mib file size (GB) | Running condition |
1 | 1000 | >1M | 1024 x 1024 | headless + RAW | |
1 | 500 | >1M | 1024 x 1024 | headless + RAW | |
1 | 250 | >1M | 1024 x 1024 | headless + RAW | |
6 | 1000 | 331562 | 512 x 512 | headless + RAW | |
6 | 800 | 110000 | 256 x 256 | headless + RAW | |
6 | 700 | >65536 | 256 x 256 | 16.4 | headless + RAW |
6 | 650 | 58920 | 256 x 230 | headless + RAW | |
6 | 600 | 60245 | 256 x 235 | headless + RAW | |
6 | 500 | 6954 | 64 x 64 | headless + RAW | |
12 | 3000 | 160163 | 256 x 256 | headless + RAW | |
12 | 2000 | 85410 | 256x256 | headless + RAW | |
12 | 1000 | 64705 | 256x252 | headless + RAW |
When you acquire data with RAW format selected, time-stamping DOES NOT work. If time-stamping is selected it would create the folder with date and time but will save the data in the parent folder. The data in this format requires re-arranging (see image below). Please seek help from ePSIC staff if you need to run in RAW format. |
If 4DSTEM data is collected manually (from Merlin software), follow the following steps to convert to hdf5:
python batch_convert_cluster.py e02 <year> <session_ID> |
This should start multiple cluster jobs, converting mib files into hdf5. The output appears in /dls/e02/data/year/session_ID/processing/Merlin.
Related articles appear here based on the labels you select. Click to edit the macro and add or change labels.
|