Displaying report 1-1 of 1.
Reports until 11:44, Tuesday 10 August 2021
H1 SQZ
georgia.mansell@LIGO.ORG - posted 11:44, Tuesday 10 August 2021 - last comment - 17:24, Friday 13 August 2021(59582)
VOPO work, week of August 9

M2 PZT low-voltage weirdness

We noticed that while sweeping the M2 PZT at lower voltages (~0-40V?) the scan looks weird - it seemed like the PZT had fewer um/V at low voltage as there were (with 0V on M4 PZT) no fundamental resonaces in this region. M4 PZT behaves as expected over its entire sweep. Adding an offset to the M4 PZT while scanning the M2 PZT brought some resonances to the low-voltage region of the M2 sweep, but these also didn't make sense - it looked like two broad TEM00's very close together. Keita took some pictures which illustrated this more clearly.

Pump polarization check

We checked the pump polarization with a fresh precision photonics PBS (a so-called 532 PBS on the table had a very low extinction ratio, it might be mislabeled). We first looked at the polarization on reflection of G:POL1 as a sanity check:

Power on reflection of the polarizer
Pin 590 uW
Ptrans 1.9 uW
Prefl 600 uW
 
Here Ptrans/Prefl = 0.32%.
 
We then put the PBS after G:L3:
 
Power on reflection of the OPO
Pin 0.61 mW
Prefl 0.6 mW
Ptrans 10.22 uW 
 
Ptrans/Prefl = 1.7%, so the light is 98.3% s-pol.

B path beamsplitter

We put the wincam on the B path (after B:M4) to look for clipping on the B:L2 lens baffle, as suggested by Adam on Friday's meeting. The beam looked clean and round except for a weak vertical streak of light through the middle, which we failed to track down. I saved a screenshot but forgot to get it off the wincam computer. We noticed that the beam was grossly miscentered on the B:BS1 and decided to move the beamsplitter to recenter the beam, using an iris off the table as a reference. We fixed the centering on B:BS1, and then moved B:M4 to fix this centering too. We remeasured the distances between B:M3, B:BS1, and B:M4, and also remeasured the tranmission of B:BS1 to be 1.07%:

 

B:BS1 splitting ratio

  Power [W] P dark [W] V ref [V] V ref dark [V] Normalized power P/Pin
P trans 6.96E-06 6.90E-08 1.054 0.034 6.76E-06 1.06E-02
P refl 6.34E-04 6.10E-08 1.034 0.034 6.34E-04 9.93E-01
P in 6.10E-04 5.59E-08 0.989 0.034 6.39E-04 1.00E+00
Comments related to this report
keita.kawabe@LIGO.ORG - 15:35, Wednesday 11 August 2021 (59594)

M2 PZT polarity is good.

Locked OPO using IR light and feeding back to M2 PZT. Scanned the voltage on M4 PZT. Measured the voltage input of the HV amplifier (i.e. scaled down feedback signal). In the first attachment, the cyan trace is the actual voltage across M4 PZT (520mVpp) and the purple is the input of the HV amp with the gain of 15 (voltage across M2 PZT would be 100mV*15=1.5Vpp). 

You can see that they're in opposite phase to each other, meaning that M2 PZT is pulling when M4 PZT is pushing. Therefore we see that the polarity of the M2 PZT is the same as M4 PZT.

The reason why M2 PZT feedback voltage appears to be negative is because the bulk of the DC offset on M2 PZT is taken care of by the offset knob on the HV amp.

At the time of the measurement, DC offset on M2 PZT was about +38V, where the sensitivity of the PZT is much smaller than at the higher voltage (see below). That probably explains why the ratio of M2 voltage to M4 voltage was about 3.

M2 PZT doesn't move much under 20V.

If you look at the data of alog 59557, it seems as if M2 PZT doesn't move much for the voltage under 20V. This is a follow-up of that alog.

The idea is to scan M2 PZT while putting various bias voltage on M4 PZT so I get 00 peak when the voltage on M2 PZT is low.

M2 PZT voltage was ~130Vpp triangular wave at 0.55Hz. The voltage was measured by oscilloscope with a resistive divider this time (as opposed to measuring the input voltage to the HV amplifier like in alog 59557). (R1=125.1e3, R2=0.997e6, dividing ratio of ~8.97.) IR transmission was recorded at the same time.

When M4 PZT offset was zero, it looked as if the M2 PZT was not moving much until the voltage reached 30V (2nd attachment).  This is basically the same as alog 59557, the only difference is the wavelength.

Giving M4 PZT a small offset, we started seeing some peaks at lower voltage across M2 PZT, but 00 mode peaks looked split there (3rd attachment). We confirmed that this is indeed 00 by stopping the scan and manually adjusting the M2 PZT offset without locking to get close to this peak and viewing the transmitted beam on the sensor card.

After scanning M2 PZT with exact same waveform with 6 different offsets on M4, it seems as if something is wrong with M2. Apparently, when the voltage goes down from positive-high to positive-low, at ~10V or so it seems as if the PZT spontaneously reverses its direction (4th attachment, yellow vertical marker). Eventually the voltage reaches its minimum (2.7V measured) and start to increase again (pink vertical marker) so naturally the PZT reverses its direction. Then at ~20V or so (gray vertical marker) it seems as if the direction is reversed again.

(Seemingly spontaneous reversal of M2 PZT direction at ~10V and ~20V sounds as if the material is depoled repeatedly, but the voltage across the PZT stayed positive all the time.)

To help myself understand things qualitatively, I made the round-trip displacement VS voltage of M2 PZT resonstructed from the position of 00 mode peaks (5th attachment). Circular points are 00 mode positions when up-scanning (i.e. the voltage was going down), square points are those when down-scanning. Eyeballing of 5 free parameters (M4 PZT's displacement relative to zero-bias on M4) was used to do this though you should be able to do a better job by math. I also scribbled the voltage-displacement "curve". But these are for qualitative purpose only.

Connection of PZT cables were checked and nothing was wrong (no shortcircuit to the ground nor cross-cabling). I'll measure the capacitance of both PZTs.

(Update: Measured the capacitance of both using Agilent handheld meter, first with DB9-BNC cable, and nothing was strange there. M2 PZT was ~418nF, M4 ~399nF, both at 100Hz. At 1kHz M2 was ~412nF, didn't bother to measure M4 at 1kHz. Later I measured the cable alone and it was about 0.16nF for both M2 and M4. From noliac catalog, it seems like these are both nac2123 with the nominal capacitance of 380nF +- 15% and the outer diameter of 12mm.)

The maximum displace ment of full 3 FSR for IR (~1.5um single-trip) with ~130V seems to be less than 3.3um with 200V (http://www.noliac.com/products/actuators/ring-actuators/show/nac2121/) though it's not crazy. This is consistent with Sheila's assessment (3 FSR with 150V, could be 4, see alog 57466).

In conclusion, it's absolutely usable as far as your voltage is larger than 40V or so, though it's puzzling why this PZT behaves very differently from M4 PZT (59557).

Images attached to this comment
Non-image files attached to this comment
camilla.compton@LIGO.ORG - 17:24, Friday 13 August 2021 (59631)
Throughput Measurements with New Thorlabs S130C Power Meter (PM testing in alog 59603 showed < +/-0.5% precision with beam position/angle)
13th August 2021 with set up 12th August. Keita, Camilla

For easier simultaneous data viewing we downloaded Thorlabs Optical Power Monitor software (link) and Ophir Starlab v3.62 software (link) on Dell Windows 10 machined normally used for Wincam Beam profiler. Keita also upgraded Ophir firmware.

Power into in-air fiber = 40.0mW. SFI1 temperature control only: TEC@ 10.29kOhm with 128mA applied. Using Ophir Power Meter after 90/10 BS off platform as a power reference to take into account laser power fluctuations. Both PM settings: 3s average, 1064nm, filter out; locked down reference meter and will not touch for duration of measurement.

Data stored on this googledoc. Keita will comment with more analysis.

Displaying report 1-1 of 1.