Inlinino bin-files

[insalt-glitch]

Hi,
I am working with AC-S instruments. I am wondering, how I can read out the .bin files from Inlinino. There is also the Compass software from Sea Bird which produces binary files that can be parsed by your project pyACS. Do you know to what end the outputs of Inlinino and Compass are compatible? Both programs produce text-files and binary files. Is there a difference in calibration, etc.?

[doizuc]

Hi Nils,
Bin files can be read with pyACS or with prepACS, a software developed by WetLabs. The data format of the bin files is documented in the Appendix A of the User Manual of Compass software.
The bin files store data as output by the ACS, its raw uncalibrated counts. Hence, in case of an incorrect device file or errors during the calibration of the instrument leading to an invalid device file, one can recover the data with a corrected device file. The bin data is also stored in binary format, which is more compact than the CSV format. The CSV format is more convenient for quick analysis, while bin files are better for longer-term archiving.
Best,
Nils

[insalt-glitch]

Thank you for the documentation! I tried to use pyACS on a .bin-file coming from Inlinino; the program fails. In particular, the output file only contains the header information because the frames are detected to be invalid.

[doizuc]

Great catch! Inlinino actually appends the host's timestamp at the end of each ACS frame (see LogBinary), This is to accurately track the time of each frame acquired by the ACS as the ACS internal clock drifts significantly. You need to overload pyACS to parse that timestamp separately. It should be pretty straight forward.

Compass has some weird features that makes the processing of data trickier, please find below an email I wrote back in 2019 explaining some of the issues using Compass 2.1rc2 that lead us to move away from that software.

The .dat files generated by compass scheduler (hourly) provide the instrument timestamp, c spectrums, wavelength interpolated a spectrums as well as the calibration file used in the header. The problem with those files is they do not provide instrument temperature, external temperature, and pressure. We also suspect that there is a wavelength registration issue in the "a" spectrums looking at the 676 peak.

Compass records files in two different ways which produce slightly different outputs:

• normal: using the save button after starting the acquisition

  • outputs only one ascii file: acs_###_yyyyymmddHHMMSS.dat (note the underscore between acs and the serial number)
  • the file timestamp is at the beginning of the acquisition
  • the wavelength are not interpolated, they are good to use

• scheduler (only available with Compass 2.1rc2): using the save hourly option

  • outputs 2 files:
    • acs###_yyyyymmddHHMMSS.dat (note the difference with the absence of the underscore between acs and the serial number)
    • acs###_yyyyymmddHHMMSS.bin
  • the timestamp is at the end of the acquisition
  • the dat file is recorded with a wavelength interpolated on a (might include other processing, not documented)
  • the bin file need to be read with prepacs.exe or pyACS

The good thing is compass scheduler also record the raw binary data of the acs finishing by .bin. This can be extracted to an ascii file with a windows utility named prepacs.

[insalt-glitch]

Thank you for the explanation! If I understand Inlinino's bin-file structure correctly, there is no checksum in the frames, but a pad byte is always written (after the 8-byte timestamp)? This way all frames are guaranteed to have the same length; they can be parsed reliably without any checksum.

In all other aspects the frames should be identical to the frames in a Compass bin-file.