Back in March, I ran a modulation depth test, with several goals in mind related to providing useful calibration information for modeling. That alog is still unfortunately sitting in my drafts. However, I was able to use the results to make a side-by-side comparison with a modulation depth test Sheila ran last Thursday, after the power outage. We are still trying to understand what the overall effect of the power outage was on the IFO. Namely, we have lost 1% of optical gain, 86964.
Some background: the modulation depth test aims to measure the fraction of carrier, 9 MHz and 45 MHz power at each port. This is done by measuring the powers at the nominal settings, and then individiually stepping the 9 and 45 MHz modulation up or down by a known value. Using a measured calibration of V/dBm and rad/V (see alog 62883), and using the bessel functions, the power fraction of each field can be measured at each diode based on how much the total diode power changes at each step. (note: I still have a to do list item to better calibrate the modulation depth in radians using OMC scan data).
Procedure: I stepped both down and up in modulation depth, resulting in 5 different measurements (nominal, 9 down, 9 up, 45 down, 45 up). In March, I stepped down by 3 dBm and up by 2 dBm, but in September we were able to step both down and up by 3 dBm. I measured for 3 minutes at each step in March, and 1 minute at each step in September.
EDIT: I realized I made an error, and the first results I report below are actually from Feb 2025. The significant difference here is that in my February measurement I only stepped down by 3 dBm for 9 and 45 MHz each, so there is less data to fit. In my March measurement, I stepped up 2 dBm and down 3 dBm, getting 5 total different measurement times. There should be very little difference in the interferometer between February and March 2025, so the differences in the results I believe are due to the fact that more points (5 versus 3) gives you a much better fit to the data. I would compare March and now for a more accurate understanding of the differences.
February results, well before power outage, only fit from 3 data points:
| Field |
Input |
POP |
REFL |
AS |
| carrier |
0.9771 |
0.9842 |
0.9419 |
0.3386 |
| 9 MHz |
0.01278 |
0.01527 |
0.02950 |
0.1787 |
| 45 MHz |
0.01003 |
0.000474 |
0.02860 |
0.4827 |
March results, before power outage, fit from 5 data points:
| Field |
Input |
POP |
REFL |
AS |
| carrier |
0.9779 |
0.9831 |
0.9164 |
0.2134 |
| 9 MHz |
0.01212 |
0.01332 |
0.04059 |
0.2761 |
| 45 MHz |
0.009687 |
0.002302 |
0.04431 |
0.5413 |
September results, after power outage (and reduction of PSL power, attenuation at IMC REFL), fit from 5 data points:
| Field |
Input |
POP |
REFL |
AS |
| carrier |
0.9783 |
0.9833 |
0.9333 |
0.3556 |
| 9 MHz |
0.01195 |
0.01325 |
0.02533 |
0.1852 |
| 45 MHz |
0.009486 |
0.001909 |
0.04199 |
0.4641 |
EDIT: Including the March results (and trusting them more than the February results) changes the conclusion. The input ratios are very similar between all three measurements. This is also true for POP where carrier and 9 ratios are concerned. 45 MHz is the hardest field to measure because the 9 and carrier are so strong at POP. There may be less 45 MHz at POP, or this is just the measurement uncertainty. At REFL, there may be an increase in carrier light now after the power outage, and there may be half as much 9 MHz as in March.
The most dramatic differences are at the AS port. Just comparing the February and March measurements, there may be considerable uncertainty in how much of each field is at AS in general. However, if we choose to believe the March results, this would suggest a significant increase in carrier at AS, and a significant decrease in 9 and 45 MHz. However, comparing February and now, the 9 MHz and carrier are nearly the same, and the 45 seems to have decreased.
