This is a followup of 61447.
In January, we used an integrating sphere (calibration information is in alog 61812) to measure the power coming into HAM6. We also measured the power reflected by OM1. Using the data, I did the following.
- Back-propagate the power in HAM6 to the power impinging PRM to calibrate IM4_TRANS_SUM in terms of the power on PRM.
- Intermediate optics reflectivity and transmissivity were provided by Craig (im4_trans_calibration_via_ham6_power_measurement.py).
- Calibrate ASC-AS_C_SUM in terms of the power coming to HAM6, using the power reflected by OM1 and the nominal reflectivity of OM1.
- Use ASC-AS_C_SUM to calibrate OMC QPDs (ASC-OMC_A_SUM and ASC-OMC_B_SUM) in terms of the power coming to HAM6 as well as the power impinging the OMC input coupler.
- Power impinging the OMC input coupler was calculated from HAM6 power, reflectivity of OM1/2/3 and a 45deg reflector on the OMC breadboard to sample the power for OMC QPDs.
- I ignored the transmission and the loss on a small 45deg mirror between OM3 and OMC breadboard.
In each measurement, Betsy blocked and unblocked the IR light upstream of the integrating sphere to subtract the dark offset. Results are below. Though the error bars are seemingly small, remember that these are just std errors of measurements. No attempt was made to make the better-than-1% type measurement (like paying attention to your body location relative to the integrating sphere and the ambient light source).
|
Sensor (Whitening Gain) |
Calibration [W/CT] |
Note |
|
IM4-TRANS_SUM_OUT in terms of power into HAM6 (18dB) |
4.550 +- 0.009 |
Whitening gain 18dB at the time of measurement |
|
ASC-AS_C_SUM_OUT in terms of power into HAM6 (18dB) |
2.972 +- 0.004 |
Scaled to the O3 whitening gain of 18dB. (45dB at the time of measurement.) |
|
ASC-OMC_A_SUM_OUT in terms of power into HAM6 (12dB) |
0.3562 +- 0.0006 |
Scaled to the O3 whitening gain of 12dB. (45dB at the time of measurement.) |
|
ASC-OMC_B_SUM_OUT in terms of power into HAM6 (12dB) |
0.3326 +- 0.0006 |
Scaled to the O3 whitening gain of 12dB. (45dB at the time of measurement.) |
|
ASC-OMC_A_SUM to power impinging OMC input coupler (12dB) |
0.3494 +- 0.0006 |
Scaled to the O3 whitening gain of 12dB. (45dB at the time of measurement.) |
|
ASC-OMC_B_SUM to power impinging OMC input coupler (12dB) |
0.3263 +- 0.0006 |
Scaled to the O3 whitening gain of 12dB. (45dB at the time of measurement.) |
We also measured the OMC_REFL power using integrating sphere while the OMC was not resonant. If I compare the power into HAM6 (obtained from ASC-AS_C) and OMCREFL, after taking out the reflectivity of all intermediate optics, the two should agree within the error bar. These optics are OM1, OM2, OM3, 45deg sampler for OMC QPD, OMC itself calculated from OMC curved and flat mirrors. I ignored the loss on 1" 45 deg mirror between OM3 and OMC breadboard. I also ignored the AR reflection of the input coupler. The result was:
P(HAM6)*reflectivity(OM1,OM2,OM3,45deg sampler for OMC QPDs**2, OMC itself anti-resonant) / P(OMCREFL) = 1.0064 +- 0.0028.
This just means that I'm likely underestimating the loss and transmission of optics somewhere at sub-percent level.
The UTC windows for each of the measurements as well as the numbers I used are in the attached jupyter notebook.
BTW I forgot to set the whitening gain of ASC-AS_C, ASC-OMC_A and ASC-OMC_B back to O3 value. They're OK now but they'll saturate at some point as we increase the power.
At a convenient time, set H1:ASC-AS_C_WHITEN_GAIN back to 18dB. Set H1:ASC-OMC_A_WHITEN_GAIN and H1:ASC-OMC_B_WHITEN_GAIN back to 12dB. Run dark offset script after that.
We reverted the whitening gains for ASC_AS_C and the ASC_OMC_A and ASC_OMC_B, and ran the dark offset script afterward.
Comparing this alog to the calibration we currently (April 1, 2022) have in theIMC_IM4_TRANS_NSUMfilter module, theInpCal_WFM10 is 4.157. This is a 4.550/4.157 = 1.095 difference. Lately, with 50 W input from the PSL according to IMC-PWR_IN, we have been reporting 39.7 W input power on PRM using IM4_TRANS_NSUM as a witness. If we changed to this IM4_TRANS calibration, we would report 43.5 W input power. It is not clear if our calibration of IM4_TRANS_NSUM has always been wrong by 9.5% (including for O3) or if this change is a result of the PSL work. I will not change IM4 TRANS calibration yet, until I have had a chance to look at the other PDs in this path like MC2 TRANS to get a complete picture.
Correction: Most relevant calibration information for the integrating sphere is in alog 61411, not 61812.
Rick Savage has brought to our attention that the integrating sphere calibration Dripta made in the above alogs is for a 1047 nm wavelength beam, not a 1064 nm.
From Rick in our email:
"
I don't know of a PD sheet that lists the quantum efficiency in which the difference between 1047 and 1064 can be easily resolved.
I think the key factor is going to be the 2% that just comes from the number of photons in a 1W beam at 1047, vs 1064 nm, the quantum efficiency not being very different, I think.
See figure 9 in the attached data sheet for the C30665 3 mm InGaAs PDs we use.
"
This will affect Keita's final measured value because the responsivity R = η e λ / h c,
where η is quantum efficiency,
e is electron charge, and
λ is laser wavelength.
So Keita's assumed responsivity R_1047 was smaller than the actual responsivity R_1064 by ~1.6%,
meaning his estimated measured watts in HAM6 should have been higher.
All other things being equal,
this increases our estimated input power by 1.6%
