KAGRA Logbook

VIS (IX)

masahide.tamaki - 21:59 Friday 03 March 2023 (24249)

DAC noise in ITMX TM actuators

[Ushiba, Tamaki]

Related log: klog24157

We measured DAC noise by using ADC and whitening filters, which has already been on IXV rack.
We disconnected the cable between DAC to coil driver and connected DAC output directly to whitening filter input.
Then, whitening filter output is connected to ADC and measured the spectra with the same whitening configurations as in IY (FM1+FM2 ON: +36dB in all frequency band).

Fig1-4 shows the results.
We measured with several situations (suspension was always SAFE, but we tried to measure with master swith open, or set offset 3 cnts, 800 cnts while taing care not saturates).
In all cases we can say measured DAC noise level is 3e-4 cnt/rtHz, which is corresponding to 9.15e-8 V/rtHz (3e-4 cnt/rtHz ✕ 3.05e-4 V/cnt = 9.15e-8 V/rtHz).
This value seems to be small compared to the theoretical limit of DAC noise (~1uV/rtHz ?) as in IY (?)
HW: calculate theoretical limit of DAC noise (16 bit, 64 kHz sampling)

Images attached to this report

Comments to this report:

takahiro.yamamoto - 17:56 Saturday 04 March 2023 (24261)

I wonder we should use 610uV/cnt instead of 305uV/cnt in order to convert a DAC noise from count to Volt if "3e-4 cnt/rtHz" means a raw value of an ADC readout.

3e-4cnt@ADC/rtHz is equivalent to 1.83e-7V/rtHz as ADC input (=3e-4cnt/rtHz * 610uV/cnt).
This is equivalent to 6e-4 cnt@DAC/rtHz (=1.83e-7V/rtHz / 305uV/cnt).

If "3e-4 cnt/rtHz" is a value already converted as cnt@DAC, your calculation using "305uV/cnt" is correct.

masahide.tamaki - 10:55 Monday 06 March 2023 (24273)

its just as you said.
i made an honest mistake, sorry.