J. Kissel

On Sep 14 and Sep 15 2026 (see LHO:91891 and LHO:91925, respectively), the PSL team swapped out the primary RF EOM in the PSL from "former, damaged by a dust storm in the PSL" to "refurbished spare." After the dust EOM was discovered to cause mode-matching issues in the IFO due to thermal distortions caused by the dust, we re-allocated the power through the PSL, such that there was less power going through the EOM. This was done by throttling the power out of the PMC. After replacing the EOM, they restored the "nominal" power allocation through the PSL, re-increasing the power output of the PMC.

That power increase out of the PMC means the the ALS / SQZ / SPI pick-off path now has more power, and thus the SPI. According to various optical power monitor channels, there's 35% more power in the SPI system.

see first attachment 1-month trend, where the vertical bar time cursor highlights the EOM swap / powr increase.

Assuming nothing else changed, that changes our power budget to as follows (just taking the values from LHO:90627 and increasing them by 35%):
                                  Former [mW]    New Power [mW]
PSL IN port of SPI prep 	    178.0            240.75
MEAS Fiber output at HAM3 D4-1J1     64.0             86.56
REF Fiber output at HAM3 D4-1J2      18.3             24.75
Output of MEAS Fiber Collimator      64.0             86.56
Output of REF Fiber                  15.0             20.29
In front of FBR_PWRIN_MEAS            1.0              1.35
In front of FBR_PWRIN_REF             1.1              1.49
Launched to HAM2/QPDA                42.0             56.81   (putting the power on QPDA behind 50R:50T M_C1 at 28.4 [mW])
Returned from HAM2                   21.4             28.94   (putting the power on QPDB behind 50R:50T M_B4 at 14.5 [mW])
(c.f. designed power budget E2400441, and optical layout D2400108).

Thankfully, no ADC channels of photodiodes are saturating at this increased power level; see proof in short timeseries of IFO PDs, QPDs, and FBR_PWRIN PDs.

We should review these new power levels along the signal chain and consider throttling the SPI pick-off path input. My biggest point of concern is power through the long optical fiber run between the PSL and the SPI prep chassis. The second biggest concern is the total power in the QPDs.