I measured OMC DC PD dark noise in the following two conditions.
1. 1 stage whitening and 100 ohms TIA (reference).
2. 1 stage whitening and 400 ohms TIA (current values).
Spectra looks good for both cases.
I compared them to neither theoretical value nor previous value, so it is very helpful if someone will check it.
Measured data is stored at /users/Commissioning/data/OMC/2024/0622/OMC_PD_darknoise_20240622.xml.
We realized that OMC DC PD cables were disconnected during the work.
So, we need to measure the spectra again after reconnecting the cables.
I measured OMC dark noise again.
Figure 1 and 2 show the result (condition is same as the original post).
Data is stored at /users/Commissioning/data/OMC/2024/0628/OMC_PD_darknoise_20240628.xml.
I compared these dark noises (green and blue) to the previous best one (brown) measured just after the improvement in klog:25965.
The shot noise with the OMC trans power of 15 mW is shown by the red line.
The dark noise with the trans-impedance of 400 Ω is much smaller than the shot noise.
While the noise below 30 Hz is worse than the previous best one, this result is good enough because other noises will limit the sensitivity at low frequencies.
We can use the 400 Ω even with smaller OMC trans power.
Considering the consistency between the previous best one and the simulation of the circuit noise, the dark noise of the 400 Ω also seems reasonable (smaller by a factor of 4), at least at the noise floor.
Do you know the exact measurement time of reference (brown) spectrum?
I tried to check the situation of the whitening filter at the measurement time, but I couldn't identify the measurement time because it was imported from text format data.
Measurement time is set as the GPS epoch time (1980-01-06T00:00:00Z) for the data imported from text format.
If we can know the exact measurement time, the situation of the whitening filter at that time can be easily checked by "FmScan" in the MEDM execute menu (see also klog#29617).
[TYamamoto, Komori]
Abstract:
I looked into the condition of the whitening filter with the help of Yamamoto-san, and found that the number of stages was the same (one) in both the recent measurement by Ushiba-san (blue and green) and the past measurement by Aso-san (brown and red).
The difference in the filter bank settings is whether FM2, FM3, and FM4 are turned on or not.
Details:
The brown curve was generated from an exported txt file.
I'm sorry but I cannot find the exact measurement time of that data.
Therefore, I searched for the measurement time when similar dark noise was observed and found it at 07:50:00 UTC on July 12th, 2023 (red line).
By checking FmScan of the DC PD filter bank, I confirmed that FM6 was turned off at that time, which means the number of stages in the whitening filter was just one in both measurements.
However, FM2 (-12dB), FM3 (-6dB), and FM4 (-3dB) were turned on in the past measurement, while they were turned off in the recent measurement.
These filters compensate for the amplifier in the whitening filter. This is probably the reason why the recent dark noise is worse at low frequencies.
We might need additional measurements with these filters turned on.
On the other hand, we might not be able to turn all of them on due to the large RMS of the signal during observation.
