NAOJ GW Elog Logbook 3.2
Michael and Yuhang
To test the available space and compare the performance of KAGRA QPD with thorlabs PSD, we did the test for INPUT.
1. We found all mirrors are not fixed well inside the mirror mounts. For example, we could easily push them out by hand.
2. We found several ND filters were used to reduce oplev light power. We took measurement with all ND filters or none of them.
The comparison is in figure 1. We could see that KAGRA QPD has a better SNR while the same oplev laser power is used. By increasing oplev laser power, an even better SNR could be achieved. The spectrum below 10Hz is quite similar but some new peaks appear at high frequency.
Apart from these, I have some questions.
0. Are we aiming for a better SNR with KAGRA QPD? If so, why don't we just increase oplev laser power?
1. Why we use ND filters to reduce oplev laser power? Is there any limitation for oplev laser power?
2. Are the screws to fix mirror inside mirror mount easily loosen by themselves? If so, we should check for other oplevs and mirrors on bench.
Note: after this test, we have bring system (simulink files and oplev set-up) back to the original situation.
Oplev laser power has been set in order to be compliant with PSD specs. Since the laser diode that we use doesn't have the possibility to control the power, ND filter were added to this purpose.
User manual of the PSD can be found here: http://www2.nao.ac.jp/~gw-elog/osl/uploads/278_20160721110810_pdp90a.pdf
I made a measurement of oplev laser power for PR/BS/INPUT/END.
They are 270uW (PR), 320uW(BS), 70uW(input) and 107uW(end).
According to the datasheet of thorlabs PSD, the maximum power should be around 66uW.
According to the manual, thorlabs PSD (which are installed on the input and on the end mirror) should have a power level so that the SUM output voltage is ≤ 4 V.
Is this the case?
BS and PR are equipped with different PSD (TAMA ones) which can accept higher power.
Yes, I checked this voltage. With current power, it is 2.8V. This is already quite close to 4V, so we don't have much space for improvement.
Anyway, I tried to remove one ND filter to have 4V, but the spectrum is quite similar with the 2.8V one.
PR and BS can accpet more power, but most of the light was lost since the suspended mirror surface is not HR for 635nm. So we couldn't have more power.