KAGRA Logbook

DetChar (General)

hirotaka.yuzurihara - 22:41 Sunday 25 June 2023 (25703)

Pastavi issue on reading xml file incorrectly

Kenta-san reported the issue related to reading the transfer function (xml file) for OMC-ASC.

The transfer function was measured by sweeping 10~300Hz. By checking xml file, it says the measured by sweeping 10~300Hz. But, when we read xml file by using the python package dtt2hdf , for this xml file, it says the frequency band of the transfer function is 0~300Hz.
I saw this issue for the first time. I think this is occuring due to the python package (might be bug?).
To solve the issue, I will try to use the latest package (but name is different) https://git.ligo.org/cds/software/dttxml/-/tree/master/

Detailed check

The xml file contents looks correctly containing 10~300Hz transfer function.

% cat BASC_OSTMYAWtoDARM_20230625.xml | grep "Sweep"

....
10 10.9 11.8 12.8 13.9 16.4 17.9 19.4 21.1 27.1 ....

When we read xml file by using the python package of `dtt2hdf` , the frequency band of the transfer function seems 0~300Hz (incorrect).

>>> import dtt2hdf
>>> import math
>>> xml="BASC_OSTMYAWtoDARM_20230625.xml"
>>> trf = dtt2hdf.access.DiagAccess(xml).xfer("K1:CAL-CS_PROC_DARM_DISPLACEMENT_DQ", "K1:OMC-ASC_DOF4_YAW_OUT_DQ")
>>> freq = trf.FHz
>>> freq
array([ 0. , 10.9, 11.8, 12.8, 13.9, 16.4, 17.9, 19.4, 21.1,
27.1, 29.4, 31.9, 34.7, 41. , 44.5, 48.4, 52.5, 57.1,
62. , 67.4, 73.2, 79.6, 86.4, 93.9, 102. , 110.9, 120.5,
130.9, 142.2, 154.5, 167.9, 182.4, 198.1, 215.3, 233.9, 254.1,
276.1, 300. ], dtype=float32)

Comments to this report:

takahiro.yamamoto - 13:42 Monday 26 June 2023 (25719)

It may seem to be caused by saving a XML file before finishing the measurement.

As shown in Fig.1, there is no measurement point at 10Hz.
This fact means that XML files have been saved before finishing the measurement.
Probably save or abort button was pressed during the measurement, I think.

Before starting a measurement, a 0-filled array that number of elements is measurement points are defined.
After then measured value and frequency are filled in the array one-by-one.
So in the case that measurement was aborted, array elements which is not measured yet is still 0.

Anyway, it doesn't seem to be a code bug of dtt2hdf. It's a specification of diaggui.

If you want to use also such a incomplete file, following code is required.
indecs = np.where(trf.FHz != 0) freq = trf.FHz[indecs] transf = trf.xfer[indecs]

Images attached to this comment

hirotaka.yuzurihara - 15:58 Tuesday 27 June 2023 (25749)

Thank your for your comment, YamaT-san!

I fixed the code, based on the comment.
The valid frequency band is shown in the table. (see first figure)
Second figure shows the noise budget by using this xml file. Blue curves is computed by xml file from aborted measurement and it starts from the valid frequency (11~300Hz).

Images attached to this comment