Detailed operation of PR3.

*Tweaking PR3 to hit the green beam at the center of gate valve between IXA and X-arm.
Since the beam position is far from yesterday's position even though the oplev values are almost zeros, we tweeked the PR3 by using optic align at IM stage.
I added the offset of -11000 counts and 250 counts at K1:VIS-PR3_IM_OPTICALIGN for pitch and yaw, respectively.
After tweaking, the oplev values of both pitch and yaw are almost out of range (fig1).

*Offload IM pitch offset
During the above tweaking, offset values for IM pitch became less than -10000 counts, I offloaded it by using picomotor on IM stage.
I rotated the PR3_IM_PITCH picomotor from -1090 to -11479 for offloading.

*Serch the green beam at X-end by using PD at EXA.
After the offloading, Akutsu-san once centerred the OpLev QPDs.
Then, we sweeped the PR3 pitch and yaw for looking the green beam at X-end.
We aded 0.05-Hz 20-count and 0.06-Hz 5000-count sine signals for pitch and yaw, respectively (suspension motion is +-200urad for pitch and +-250 urad for yaw).
Excittion points are K1:VIS-PR3_IM_OPTICALIGN_PIT_EXC and K1:VIS-PR3_IM_OPTICALIGN_YAW_EXC.
Then, we can find the peak in the PD signal (see fig2).

*Fine alignment and centerred QPDs as a reference
We checked the pitch and yaw vales when PD detected the green beam and move the PR3 inclination at those points.
Then, we can clearly see the peak (maximum value is about 3000).
After that, we centered the QPDs and now green beam can be seen at EXA when QPD values of both pitch and yaw of PR3 are zeros (fig3).

*What we should do next
To check the value of 3000 counts is reasonable or not by checking the gain of PDA, ITMX transmissivity, incident beam power, and so on.

Note:
During this work, we roughly calibrate PR3 IM pitch picomotor and optic align offset.
PR3 optic align PIT: 0.04 urad/count
PR3 optic align yaw: 1 urad/count
PR3 IM pitch picomotor: 0.05 urad/step (FWD direction move the PR3 along +PIT direction)

Also, we found the calibration factor of TILT OPLEV yaw has negative value and + direction of horizontal is opposite from the +yaw direction.
This should be fixed at some point.

The gain of PD in EXA chamber is 50dB.

T. Yamamoto, Akutsu

So let's calculate something to confirm the green beam power found at the EXA-PD.

• According to the Ushiba-kun's klog in the last night, the PD at EXA showed peaks of around 3000 counts for the transmission green.
• Thanks to the Chen Dan-kun's report, the Thorlabs PDA100A (or A2???) gain is set at 50dB.
• In the PD manual (see here), 
  • At 532 nm, 0.34 W/A, maybe.
  • With the 50dB gain, and assuming the load resistance of high-z (I believe our AA is like this), 4.75e5 V/A
• The usual conversion factor: 40V end-to-end diff input / 16bit counts = 40/2^16 = 610 uV/cnt

Considering all the above, 3000 * 40/2^16 /4.75e5 /0.34 = 11 uW.

At the same time, what power should be transferred?

• According to Tanioka-kun, the GreenX power at the Faraday output is about 10mW.
• Acoording to the wiki,
  • ITM, ETM transmission: 0.06, 0.06 each
  • PR2 transmission: 0.9?
  • PR3 reflection: 0.9?
• GV windows transmission: 0.9 each??  Please have a look here. We have three closed GV windows as of today.
• The active area of the PD is 10mm sq. (10x10 = 100mm^2)
• Gaussian beam has a spot radius of w=35/sqrt(2) mm at the PD
• According to my calculation, a Gaussian beam of power P0 and spot radius w would have a power of P within the radius r as follows:
  • P/P0 = 1-exp(-2 (r/w)^2); for simplicity, let the PD area as a 10mm radius circular; substituting r=10mm, we got P/P0 = 0.55.

Then, 10mW * 0.55 * 0.06^2 * 0.9^5 = 11.7 uW

Hmmm... too much matching. Over consistent. Maybe I can suspect myself that I misunderstood something.

Now a calibrated signal in Watt can be available as K1:AOS-TRX_PDA1_DC_OUT.*
Same filters were already prepared for K1:AOS-TRY_PDA1_DC

Ok ok, as I expected, I was wrong in the following points at least.

• This green beam has not passed ETM, so we should omit one 0.06.
• Altough I wrote 1-exp(-2 (r/w)^2), what I calculate was 1-exp(-2 (r/w)).
• r should not be 10 mm but 5mm!

Taking these all,

• 10mW * 0.0787* 0.06* 0.9^5 = 28 uW

Ok, maybe the next round would be to check the measurement...

As I mentioned at klog#17187, I implemented OLDAMP in the TM stage and OLDAMP with DC control in the IM stage.
In Fig.1, left shows the yaw motion and right shows the pitch motion which are read by the TM_TILT_OPLEV. Blank lines are the each motions when the control was off but offset value was inputed in the IM stage to keep the OpLev beam in the QPD center.
Red and blue lines are the each motions when the OLDAMP control was on. Please note that while the black is the out-of-loop signal, red and blue are the in-loop signal because OLDAMP use the OpLev.
Mirror motion of the PR3 is reduced below 0.2 urad according to the in-loop OpLev signal. I think this performance is enough to align the GR light to X end.