The optical spectrum of the GHz oscillator: FWHM ~ 4nm

3 phase noise measurements made on the Amplitude GHz oscillator with different PZT configurations :

- black curve: PZT connector is open
- green curve: PZT connector is shorted by 50 ohms
- blue curve: PZT is excited by 100mVrms of white noise coming from a generator.

on the blue curve, one can clearly see a phase noise increase in the region 10kHz - 1MHz but it is not evident the peaks seen with the PZT open or shorted are related to the peaks excited with the noise injected on the PZT.

with a PZT not excited, one can just observe that the phase noise is decreasing a lot around 10kHz to reach the reference oscillator phase noise floor and then increase again exactly when the PZT resonant frequencies appear, between 20kHz and 200kHz.... reaching at the end the phase noise detection floor.

I add below the measurements done on October 20th, the ones done in September which are very similar and on which one can see a peak around 26kHz.

Thorlabs PZT datasheet.
Reference: PC4QR

Installation of the MIKAN.
powermeter (with OD1) just after the oscillator shows 440mW for 4A of the MIKAN pump current.

Installation of a periscope as the oscillator beam position is very close to the table... not easy to put devices at this height (be aware to use proper wavelength range mirrors: Thorlabs BB1-E03)
Installation of a HALF waveplate to align beam polarisation on the isolator axis
Installation of a High power isolator (the one of ThomX): Newport ISO-FRDY-05-1030-N
powermeter (with OD1) just after the isolator shows 427mW

Coupling into a 50-50% fiber coupler using the Thorlabs XYZ table NanoMax TS.
We reached 117mW after the 50% arm which means we coupled about 230mW (the coupling better than 50%).

MIKAN phase noise and RIN measurements:

we have excited the PZT with a swept sinus wave from 1kHz to 10kHz and from 10kHz to 100kHz.
here are the 2 different spectrums:

- the 1kHz-10kHz is a standard spectrum where we see the impedance behavior of the PZT: Zpzt ~ 1/jCw

- the 10kHz-100kHz exhibit several PZT resonances and the 1st one is close to 28kHz.

(without PZT resonances, we should have the same behavior at higher frequencies than in the range 1kHz-10kHz)

To summerize the results of this week :

We made the setup in attachement 5 and align it to have the good mode in the cavity (TEM00).

With the good alignment, we found the lock of the GHz, in attachement 1. We used the motors in the cavity to position very precisely the mirrors to find this lock.

In attachement 2, we can see the error signal obtain thanks to the PDH, which is very clean.

But after that, we saw some modulations at 30 kHz frequency which represent the resonance frequency of the piezo in the GHz (attachement 3). To compensate this, we placed a fibered AOM to cut the high frequencies.

Finally, we observed a big difference in the transmission noise if the motor loop is open or closed (attachement 4). To have the best noise, we have to stop the closed loop and stay in open loop, after adjusting the mirrors with the cosed loop.

We have also measured the finesse (2000) and the coupling (20%), detailed in the previous post.

We did the measurement of the Finess and coupling with the NKT, with 2 EOM, without AOM (so the lock signal is noisy), on the burst line.

For the coupling, we have the same value as with the GHz locked on the burst line too, which is around 20 percent.

it seems this low coupling comes from the geometrical coupling as we observed an incoming beam bigger than the cavity mode on the cameras.
(the same oscillator, locked on the lock line exhibit 50% coupling)

the FSR center value is 879.9MHz, and the FWHM is between 879.67MHz and 880.1MHz.

So, the linewidth is about 430kHz and then the Finess value is roughly the same as before about 2050.

in addition to the standard locking scheme with the GHz laser PZT,

we added an AOM after the PDH modulation EOM and we drove it with an FM modulated signal generator (FMDev = 2.4MHz) seeded by the error signal.
(we didn't put a 50ohm plug to adapt the error signal coming from the PDH box, otherwise, it is too much smaller)

the result is a transmitted signal almost clean for some milliseconds... but we still have regular unlocks that the PZT loop is unable to drive.

the PZT resonant frequency around 30kHz seems much less present in the error signal.
todo list:

- take some data of the error/trans signals to make a post-mortem analysis (a windowed FFT could tell us if the 30kHz is more powerful just before an unlock)

- make an RLC model of the cable+resistor+PZT capacitor, to try to find a way to dump the 30kHz frequency.

after 3hours (1pm --> 4pm) the cavity moved by 12um in total

new positions: +5.1065mm on M1 and M2 (5.1050 before)

the drift is 4um/hour which amounts to about 10kHz drift / hour+the temperature of the optics room also changed unfortunaltey (linked to extinction of air flow at around 11am this morning). the drift in the room temperature was about 0.5K from 1pm to 2pm and 3K since 11am.

maybe this measurement needs to be redone to more precisely correlate the drift to an overall temperature change

the M1 is not at 5.1045mm to put the fundamental on the scanning range of the laser piezo

M2 still at 5.104mm

M3/M4 at -2.9mm still

to find the fundamental mode we moved the cavity mirrors to new positions:

+5.104mm for both planar mirrors in closed loop

spherical unchanged (-2.9mm)

a cavity length measurement gave 880.9Mhz. The m1/M2 have been put to +5.1000mm both M3/M4 srill at -2.9mm

doing the sensitivty measurement of the piezo we observed satellite peaks around 800 to 900 Mhz. Removing the phones from the experience room(in the cupboard in the sas) removed that effect immediately.

A 10V applied on the laser piezo was found to induce a frequecny change of about 5kHz, compatible with expectations from the piezo sensitivity. this was measured by a freqeuncy analysis of the signal produced by the laser itself on a photodiode det10a

following the measuremetn at 879.9 MHZ of the laser theoptical cavity length has been adjusted to that value from 876MHz setting of the 5th of may 21.

new motor positions:

M3/M4 : -2.9mm

M1/M2 : +5.196mm

the ip adress of the smaract driver is now 10.0.53.10:5000

please think about putting such kind of changes in the logbook

With the help of Ronic the cavity was locked in preparation to measure the polarization of the reflection line when the cavity is locked (measurement when it is not locked was done before)

the purpose is to compare the two measurements (locked  vs not locked)

General details:

• when not locked : measurements of the polarization was taken from the point where the photodiode is placed in the picture enclose
• When Locked: we can't measure it from that point as we need the line to split into two, one goes to PDH to maintain the lock of the cavity and the other one we use for our measurement.
  • The reflection line is split at the point where the beam splitter is (BS, behind the photodiode in picture), we intended to take the measurements from this point.

Observation:

before starting the measurement of polarization, we observed

• the power measured for the reflection line (point at end of red arrow in picture, after BS) is really sensitive to the polarization, it shows when rotating the half-wave plate
• but, when measuring the power at the point shown in the picture (before BS, where the photodiode is placed) it is not sensitive to polarization.

This tells us that the dielectric BS placed in the reflection line affects the polarization.

This could affect the stability of the locking of the cavity, as the PDH is sensitive to polarization.

** Further investigation is needed before proceeding **

Footnotes:

• BS: Beam splitter.
• Dielectric component's sensitivity to polarization
• most of the components placed in the transmission line are dielectric.

After data processing, I finally find the polarization in transmissin of the FPC. As you can see below, the polarisation is not linear but elliptical.

The four ellipsies are here due to having not enougth parameters to extract with certainty one ellipse.

We put a diaphragm in the input beam but it doesn't work to suppress the ring light only but it attenuates the whole transmitted beam (cavity mode and ring light) viewed on the camera.

I made many tests about polarization and we can see some points thanks to the following figures :

1- The ellipsicity of the NKT is near to 0 (between -0,2 and 0,2) which means the polarization is rectilinear horizontale (attachement 1).

2- The power of the NKT has no influence on the polarization, verticale or horizontale (attachement 2 and 3).

3- The polarization is not changed by the type of mirror (Ag or Diélec) (attachement 4)

The next step is to measure the polarization after the cavity to know the ellipsicity.

To have the polarisation, we must have 1 value because the laser in entry of the system has 1 polar. Due to that, the graphe of the ellipticity is not true, it is the mean value on the following table which shows the ellipticity and the polarization.

We change the setup a bit to have the same polarization on the photodiode and the PDH (in reflexion of the cavity). CF: schematics (futur).

Thanks to that, the symetry between reflexion and transmission is better now (as you can see on the plot).

We also put more power on PHD to have a better SNR by changing splitters. (We had 100 µW and now we have more than 1 mW, at the expense of the power on the camera).

Thanks to that we obtained a good locking, The coupling is quite good (roughtly 50%, on the green line)

In the following plot, the green line is the reflexion and the yellow lin is the transmission.