C3:PSL-NPRO_PWRMON channel, and npro.db file are added

for monitoring power output of the NPRO (reflected beam from Faraday Isolator)

I haven't reset the crate yet, so the channel may not appear in DAQ yet.

It's connected to VMIVME-3113 at #C0 S60.

The photo diode is Thorlab PDA55 with RG1000 filter.

Calibration for power is 1.81 mW/V, now it reads 5.56V.

I checked PMC circuit to make sure that the schematic matches the actual board.

This is important because we want to make sure that it works exactly the way it should.

This is also useful when we want to modify it.

I check R and C on the board, everything seems fine except R2.

It says 9091 F which matches the schematic (9.09 k), but a multimeter reads 4.5 K, I might measure it wrong.

I also measured the TF between TP2 and TP4, I'll attach the result.

When I measured TF between TP3 and TP8,  I did not fully push the card into the socket, and this killed PA85 when I applied high voltage to the card.

Koji found me an unused Mach-Zender card at 40m. I replaced its PA85 for PMC card.

I tried it. This time, I made sure that the card was properly in its place, I screwed it down to the crate.

Then I connected PD cable, LO cable, HVout then HVin.

However, it sitll does not work , no high Voltage coming out. Only low voltage ~0.5 V which changes with the value on the PMC DC control slider.

I don't know if I accidentally killed it during the transplant operation.

It's a very bad day to be PA85.

I fitted the TF of PMC servo card simulated from LISO, and found poles at 2.01 Hz, 59.7 Hz, 13.4 MHz, and zero at 479.4 Hz.

 From last week, I used LISO to model the TF but it did not provide the values of poles and zeroes.

Thanks Koji who taught me how to use LISO fitting feature. Now I can find poles and zeroes of the simulation.

The frequency range spans from 1 Hz -  1 MHz, covering out region of interest (~1 Hz up to a few hundred kHz).

  For higher frequency, I couldn't make it work yet.

pole 2.0139158941  ### fitted (name = pole0)
zero 479.4339553158  ### fitted (name = zero0)
pole 59.7091341574k  ### fitted (name = pole1)

factor 9.5703858437

pole 13.4116105660M 99.7575256877m  

The plot shows simulated data and fit data, we have a nice fit that we can hardly see red and green plots.

 The code for simulating TF is pmc.fil, which can be found on previous entry, the code for fitting, tffit.fil, is attached below.

I measured PMC's TF between MIXER OUT and PCMON, the measurement agrees well with LISO model.

The source is 10mV swept sine. It is sent through FP2. MIXER OUT and PCMON are connected to chA and chB respectively.

The measurement was done twice, with loop closed (PMC is locked to the laser )and loop opened (PMC's enable switch is off.)

The data from LISO model is added by 30dB to match the gain slider.

The result from closed loop looks weird at low frequency, I'm not sure why so I'm reading Application note 243 to find out about coherence.

I did this to make sure that the schematic and the board are matching, all parts in the schematic are labeled correctly.

Please ignore fig1, I did not have phase plot in it.

Hoo! So beautiful fit!

 I checked the PSL setup today, and the PMC gain setup has to be changed from 30 dB (maximum on gain slider) to 22.5 for best transmission signal.

From the previous  setup, see quote below, the maximum gain of 30 dB on the PMC gain slider was not high enough,

This means that even though the gain is set to maximum, the signal from the transmitted light does not oscillate.

But today it did oscillate, and I had to reduce the gain to 22.5 dB. When I checked the PMC gain, I turned off the FSS loop

to make sure that the FSS loop won't actuate on the NPRO, and the signal are purely from PMC loop.


There are no other changes of parameters. The power input is 30.7 mW, RF power, phase shift are the same

as before.


After adjusting the PMC's gain, I also roughly adjust FSS's loop gain. I haven't optimized it yet, just determined it by looking

at C3:PSL-RCAV_RCTRANSPD signal on the oscilloscope.

common gain is changed from 5 to 8 dB

Fast gain is changed from 8.5 to 13 dB.


I saved change the values for PMC gain, common gain and fast gain in STARTUP.CMD file.

I don't know what causes the gain changes here, I will check the TF of PMC loop again. If it is real, it means

we might not have to modify the PDH servo card.

NOTE: the slowDC for locking both cavities is ~ -0.103 V

I measured the beat note signal from two different setup (f modulation) and plot the result below

We want to see where our beat signal is, and compare it to the noise budget, and improve the sensitivity.

I'm using the same noise budget for now, because

my noise model from RIN has not been finished yet. I'll try to finish it soon.

For beat measurement, I measured the feedback signal of the PLL loop,

since the UGF of the loop, which is 53 kHz, covers the region of our interest.

I used two frequency span on marconi, 20 kHz and 100 kHz.

I checked the calibration for each frequency spans which are 14.31 kHz/V and 70.8 kHz/V respectively.

The results are the same at low f to ~ 200 Hz, at higher f, 20 kHz span(brown line) has better sensitivity.

*b100 and b20 in the data.mat files are the result of beat at fmod 100kHz and 20 kHz in the format of

frequency, V/rtHz , f / rtHz

--------------

Setting

PMC

power input: 16mW

V_RF: 5.7 V

Phase ADj: 2.23 V

Gain: 13 dB

RCAV

power input: 2.86 mW

V_RF: 10

Phase ADJ: 4.2 V

Common gain: 18.9 dB

Fast Gain: 20 dB

ACAV

power input: 2.0 mW

 I realigned the beam to PMC, ACAV, RCAV, optimized gain, and find a significant coherence between ACAV_RCTRASPD and RCAV_TRANSPD.

I haven't re-aligned the beam to each cavities for awhile, and the alignment was quite bad.

PMC_RCTRANSPD: 9.4 - > 10.9 V

RCAV_RCTRANSPD: 1.7 ->2.07 V

ACAV_RCTRANSPD: 0.8 -> 1.3 V

So I need to optimize the gain setup again, see detail below. 

I measured beat signal and coherence before and after realigning. 

For coherence, I see nothing significant except <beat | ACAV>, see fig1, so I did not save the rest of the measurement.

After I realigned the beam, there is a big coherence between ACAV andRCAV see fig 2.

the coherence between PMC and RCAV follow the same trait as that of PMC and ACAV, but slightly less, so I show only PMC and ACAV.

However,  the beat note before and after realigning the beam are still the same, see fig3.

I'll add RIN from  RCAV/ACAV/PMC .

-----------------

gain setup

--------------

PMC: Gain 16

         RF_ADJ 5.7 V

FSS: Common gain 21.5

         FAST gain 21.2

        RF_ADJ  10

These values allow maximum Transmission power and stable lock (I checked this by unlock and lock the cavity and see if the signal is stable after loss lock)

We need to see a plot of the coherence between the PMC trans and Rcav trans and the Acav trans. Also the RIN of all 3 plotted on the same plot.

Then, inject a small line into the current adjust at 100 Hz and measure the height of the peak on all three trans and the beat signal. Use this peak to scale them all and put them on the same plot in physical units.

Then measure the TF from the current adjust actuator to the PMC trans and close an AC coupled feedback loop on it.

I modulate the laser power @ 100Hz and measure RIN from PMC,RCAV,ACAV trans and beat note at 100 Hz, t

and then find The conversion dp/df for RIN to frequency noise to be 1.01 MHz/W. I also measure the TF from the power adj to PMC TRANSPD.

High coherence between RCAV and ACAV trans seems suspicious, and it  changes with FSS gain setting.

My gain setup might not be optimized. I'll check it tomorrow.

I use a function generator to send sine wave signal at 100 Hz, 0.2 Vpk-pk to modulate the laser power.

The peaks at100 Hz for RIN from PMC/RCAV/ACAV  and fbeat are

                Vrm/rtHz at 100 Hz [V]          DC [V]             RIN 100Hz (Vrms/DC) [1/rt Hz]

 PMC          12.49 e-3                             1.5                  8.327 e-3

ACAV          3.27 e-3                              0.313            10.45 e-3

RCAV        11.57 e-3                              0.59              19.61 e-3 

------------------------------------------------------------------------------------------

 fbeat         984 e-3      with 70.8 kHz/V covertion ->     70Hz/rtHz 

The coherence between beat note and RCAV/ACAV rin is ~1. So if we assume that the mechanism

that converts RIN to noise in beat note are the same in both cavities, we have

fnoise ^2 = (RIN_AC x Pin_AC x conversion)^2 + (RIN_RC x Pin_RC x conversion)^2 .

(This might not be accurate, since the peak from f beat is not that relatively high)

Pin for Acav and Rcav are 2.3 and 3.3 mW.

This gives conversion = 1.01 MHZ / V @ 100HZ

This conversion will be applied to RIN level to see how much it would be convert to fbeat noise.

 (assuming the conversion is flat at all freq, we will find this conversion at different frequencies later)

 TF from  power mod input to PMC trans readout is plotted below. The magnitude at 100 Hz is -17dB

 The source from SR785 is split by a T, one goes to ref channel A, another is sent to laser power modulation.

Response B is

-------------------------------

NOTE: (from the yesterday)

 DC level for

ACAV_RCTRANSPD  = 285 mV 

RCAV_RCTRANSPD = 612 mV

PMC_RCTRANSPD = 1.48 V

Today we improved some frequency noise by reducing the scattering light at the ACAV's foam cap.

We are searching for the cause of noise above the noise budget. 

By shaking the table and the foam box, the frequency noise seems to increase a lot at 100Hz to 500 Hz.

We are looking at both beat signal and VCO signal.

First, we added an ND filter just behind the EOM before the beam is split in to two paths to see if we can

reduce the power thus, reduce the scattering light. There is no significant change.

Next, by tapping the optics to ACAV path the noise increases significantly, so we checked and found

that the beam into the ACAV is clipped on the opening of the foam cap and causes some scattering, see [elog entry].

We tear the aluminum tape on the cap of a bit and tried to make the hole bigger, until we could not see the scattering.

The noise is lower by 1 order of magnitude around 100 Hz from beat and VCO measurements,

see fig 1 and fig 2.  We check the RCAV, which has smaller opening, but there seems to be no clipping.

We will check the scattering problem on RCAV again, by slightly remove the cap and see if there is improvement.

Fig 3, shows the comparison between the beat note freq and the VCO signal. A lot of mechanical peaks appears

on the beat note, but not on the VCO signal. We will want to fix the optics for beat note measurement, maybe bringing

the beam back to 3" height. Also, the beam splitter that add both beams together are very susceptible to seismic.

Aside from the scattering at the opening, we haven't found any major component contributing to the noise yet.

Unit on fig 1  (beat) is converted to Hz/ rt Hz.

Unit on fig 2 (VCO) is not converted yet, I have to look up the df/dV calibration for VCO.

• Even after the removal of the Al tip, the scattered light noise looks still exist.
• Particularly I still could see apparent fringe wrapping when I shook the table or touch the foam construction.
   
• I still could not reject the possibility of the backscattering towards the PMC.
• We had the fringe wrapping up to ~200Hz. This corresponds to the motion of the scattered body or the optical
  path for the scattered light by ~100um. Is that possible?
• The VCO feedback and the beat note PLL feedback seemed to have the same information so far.
• The lollipops at the trans mission ports are terrible. They are mechanically incorrect.
• What is the correct conversion between the VCO feedback and the beat note PLL feedback??
  Both are VCO feedback signals but the slope looks different. Need precise investigation.

The Voltage vs Frequency conversion for VCO is here.

It's not linear over the range. I have to check what is the average value of the VCO during our measurement.

Then, if it's small enough, linear approximation can be used to convert the data from VCO to Hz/rt Hz.

I think It's better to use linear fit because the psd of the voltage does not contain information about the sign of the signal.

What about the frequency response. Why did we see the different shapes between the spectra even with the coherence ~1.
You can measure the transfer function between those two VCO feedback and should try to explain it.

Is there any transfer function in the AOM VCO, for example?
If I can believe D980401-B.pdf, the VCO freq control path seems to have pole-zero pair at 1.5Hz and 41Hz.

I measured the calibration df/dV for VCO to be ~48 kHz/V. Then I convert the Vnoise fro VCO to frequency noise.

I injected the signal at 90 Hz and 1 kHz and measure the corresponding peaks from beat noise and VCO.

Then I rescale both data to match the peaks to get the conversion factor.

                        f=90 Hz, pk/floor               f =1kHz pk/floor

VCO                   1.43mV/20uV                   2.95 mV/20 uV

BEAT                 2.18mV/ 25uV                     4.11 mV / 25uV

conversion fac  

 beat =  VCO x K  =  1.53                                   1.37

let's use 1.4 for average

Since df/dV for beat is 71kHz/ V, df/dV for VCO is 71/1.4 kHz/V =  51kHz/ V.

The plot below shows RC noise as measured by VCO and beat note.

TRANSFER FUNCTION!

Here is the TF of LIGO's VCO. I measured the TF at 3 different RF output voltage levels (C3:PSL-FSS_VCOMODLEVEL) and plotted them together.

To measure the TF of the VCO box, I reduced the VCO RF before disconnecting any cables.

Source out from SR785 is split by a splitter. One goes to Ref ch A on SR785, another goes to VCO wide band input.

RF out to AOM is connected to Response ch B on SR785. 

The TF looks bad when I turned the RF output V to 5 which is normally used during the lock. I'm not sure if there will be

reflection of signal or not, so I decide to lower the RF output V.

The magnitude from each measurements on the plot are offset for comparison purpose.

As we can see, VCO does not have flat freq response. This should be able to explain the different shape between

Beat note freq and VCO feedback signal.

I switched the post to V-block for Faraday Isolator mount, for better stability, and adjusted the Faraday isolator to minimize back reflection to the laser.

I also measure the TF from ACAV path,  The UGF is ~65 kHz.

  The faraday isolator was installed on a standard pillar post, so I use a V block to mount it instead.

After adjusting the FI, I remeasured the beat note frequency, and the signal did not change from yesterday measurement.

(no differences between red and green plots)

blue: beat signal before fixing the ACAV opening

red: beat signal after fixing the ACAV opening

green : beat signal after re installing the FI

ACAV TF: I connect the signal output after the PDH servo box to SR560 A and SR785 B (resp)

                 Then source out from SR785 is connected to SR560 B

                 The output of SR560 is connected to SR785 B (ref) and to the VCO

The setup for SR560 is DC coupling for A and B, select A - B, gain 1, no filter.

I measured the TF between VCO feedback and PLL feedback. The result agrees with the TF of the VCO.

  The frequency noise measured from PLL feedback and VCO feedback do not have the same shape, even though

the coherence is ~1. There might be some devices that do not have flat frequency response and change the shape

of the frequency noise. 

    So I measure the TF between the feedback signal to VCO (ref, chA) and the feedback signal to Marconi (resp, chB).

The excitation is sent to In test 2 ch on FSS servo input which modulates the frequency of the laser to RCAV.

Then I measure the TF of the LIGO's VCO box.

The shape of the transfer functions agree well except a resonance peak around 100 Hz and a pole at 7 kHz.

 This TF can be used to calibrated the feedback signal to the VCO to real frequency noise between two cavities.

Then calibrate the feedback signals correctly. The plot in this entry is totally unacceptable to show to anyone.

I measured the RIN of PMC/ACAV/RCAV and noise in beat note when 250mV white noise was injected in the laser current actuator.

The result is note quite reasonable because the noise grows with f^1.5 instad of 1/f  .

The intensity servo is working now and reduce the noise from beat measurement a bit.

I did this to find out how RIN couples to frequency noise in beat note measurement, thus I can refine my noise budget.

I used SR785 as a source for white noise  @100mV and sent it to OUTPUT ADJ CH on the laser driver. It is the current acutator

which adjusts both intensity and frequency of the laser.

RIN from three cavities change almost at all frequency, but quite alot around 100 Hz -10 kHz, see fig1.

The noise from beat measurement also increases around that frequency range as well, see fig2.

and the coherence between RIN from each cav and the beat measurement are similar, so I plot only the coherence between RCAV_RCTRANSPD and beat note (fig3).

The sum of noise is
 

S_beat^2 = S_rin^2 + S_other^2     ----(A) , so

S_beat_with noise^2 = S_rin with noise^2 + s_other^2    -----(B)

 (B) - (A)  will give

 (extra frequency  noise due to RIN)^2 = (RIN)^2 * (conversion factor)

should tell us how RIN couples into S_beat

 The residue from white noise injection is plotted on fig4.

The input power to ACAV is 2.2 mW, RCAV is 2.5 mW.

I match the frequency noise residue with RIN by   quadrature sum of RIN * Pin * 2e3 * f^1.5, see fig 5.

This is very unreasonable. We expect noise due to RIN to be 1/f. 

 Meanwhile, suppressing intensity noise is still a good idea. I used SR560 to stabilize laser intensity.

The PD used for monitor the laser power is AC coupled to SR560, and the output is fed back

to the laser driver OUTPUT ADJ ch,

Gain setting 2 INV with 30dB attenuator at the output.

Low pass, with a pole at 100 kHz

 To my surprise, the noise of the beat note is reduced a bit, despite verey small coherence  between RIN and fbeat

we observed before. I compare with the data from this morning and I measure the beat frequency noise again

right after I finished with the data with intensity servo (2am), see fig 6.

I'll verify the data again by comparing RIN from each cavities, with and with out intensity servo.

You have to post here the TF between the laser current actuator and the PMC trans PD. Then, you should think about how to make an effective servo using a single SR560 (i.e. what poles).

Then when you hook it up, measure the CLG of the whole thing and see if the PMC trans PD noise spectrum goes down as expected.

setting up the gain for intensity servo

1) TF of current actuator and PMC trans PD:

    The TF goes down at 40 dB per decade(1/f²). It seems to be a good idea to have the UGF around 10³ where the phase margin will be compensated a bit and the slope will go with 1/f.

     - We want to build a servo that looks like a band pass. Our servos can be separated into three parts

      TF (whole servo) = TF of current actuator + TF of SR560 + TF of our intensity servo box, or
     H(s)   = A(s) * B(s) * C(s), we can write H(s) in polynomial terms. TF of current actuator can be measured and fitted to determine its Laplace Transform.

 TF of SR560 is a band pass. Then we can calculate the TF of our intensity servo and bulid it.

The fit and its parameters of current actuator is plotted below.

The poles and zeroes are

pole 1, simple:  14.7 kHz;   1/(s + f_pole *2pi)

pole 2, complex: 6.39 kHz, Q = 0.227;    w² / (s² + w*s/Q + w²)  ;w  = 2 pi *f_pole

pole 3, complex: 7.25 kHz, Q = 0.642;

pole 4, complex: 89.14kHz ,Q = 1.143; (This one is less important, since the expected UGF is ~ 30 kHz)

zero:  390.4 Hz, Q=0.433, (I'm going to treat it as a simple zero)  s - f_zero *2pi

The calculation details are in the matlab code, file bandpass.m.

The TF of the servo will be fitted by LISO.

2) CLG of the whole thing

3) PMC trans PD noise spectrum, (before and after the intensity servo)

I modified Dmass' intensity servo, but the servo is not working yet.

We want an intensity stabilizing servo, such that the UGF of the whole OLG TF is ~ 30kHz. So I need to build a servo to meet this requirement.

The OLG  TF of the current actuator to PMC trans PD was plotted in the previous entry. The slope around 30kHz goes like 1/f^3

so I need a servo with a slope of f^2 around 30 kHz to compensate, and get the total slope of 1/f for stability.

The designed servo consists of double stages of simple one zero- one pole TFs, where the corners frequency are 20 kHz (zero) and 70 kHz (pole) for both stages. 

See fig 1 for (a) simulated TF of the designed ISS servo, (b)TF from current actuator and the servo, and (c) sum of (a) and (b).

If we increase the gain so that the UGF is 25 kHz, PM is 40 degree. and if UGF is 30 kHz, PM is 27 degree.

PM is too small. I might need to modify it later.

The ISS filter I need is fairly simple, see fig 2.

The schematic for the ISS I got from Dmass is here. I modified the top part of the schematic for my use.

After the modification, I measrued the TF of the intensity servo part. The TF looks nonsense. I have to check what did I do wrong.

I measured the beat signal and VCO feedback signal to see a peak from cavities' mode of vibration, there are 4 peaks from 1kHz to 20kHz.

We expect to see  the cavities' resonance around 10k Hz with high Q, say 10^6. I measured the beat signal and VCO feed back to double check my results.

There are 4 peaks, 5.98 kHz, 10.28 kHz, 12.8 kHz, and 13.85 kHz, that coincide. However, the peaks do not look sharp (Q is too low), and the FWHM

of peaks at 5.98kHz and 13.85kHz might be too broad to be the cavities' mode. So 10.28 and 12.8 seem to be good candidates.

The FEA page indicates that the first longitudinal mode is at 13.9 kHz. I suggest to zoom in on that one once you get the frequency noise signals going to the DAQ.

The apparent width that you now see is probably just the FFT window. But, we can use the peak's RMS as a cross check on the calibration since we know that it has k*T/2 of vibrational energy.

I measured the TF between the current actuator of the laser and PZT on NPRO to see how much the current actuator changes the frequency.

The result, if the current act on the laser mostly adj Freq or Intensity, is yet to be determined

The current actuator on the laser driver changes both frequency and intensity of the out going beam.

This experiment aims to learn how much the current actuator drives the frequency of the laser compared to the intensity.

1) TF between current actuator (Vact) and NPRO PZT (Vpzt)

The source is split and sent to 1)the current actuator, 2) ref ch A on SR785

The response is picked up at fast mon on FSS loop, the voltage between Vmon is the actual voltage sent to the PZT, see the schematic.

The calibration for NPRO PZT is 3.07 MHz/V

Thus response/ref is chB/chA = Vpzt/Vact. I correct the unit to be Hz/Vact by multiplying the Vpzt by the calibration, 3.07MHz/V. or add the original result (in dB) by 20log(3.07 e6) dB.

 2)The TF between the current actuator (ref) and PMC trans PD (resp) was plotted in this entry.

The magnitude on the Y axis is Vpmc_transPD / Vactuator. To correct it to RIN, divide Vpmc by DC value of PMC_trans PD [1.37 V] or subtract 20log[1.37] from the result in (dB).

3) TF between  pmc trans PD (ref chA) and PZT (resp chB), see fig 2.

   The unit after the measurement is [Vpzt / Vpmc_pd]. To correct it to Hz/RIN, multiply Vpzt by the calibration and divided Vpmc_pd by its DC level,

or add 20log(3.07e6) + 20log(1.37) 

 To sum up,

1) TF between current act and PZT tells us how much  frequency changes when we modulate through current actuator,

   2) TF between current act and PMC trans PD tells us how much RIN changes when we modulate through current actuator,

3) TF between PMC trans and PZT tells us what changes more after current act is modulated.

This will be compared with frequency change (calculation) due to RIN-> thermo optic.

Then we can decide if current adj mainly change frequency or intensity.

I measured the beat signal from 3 different power input levels, the signal goes down with the power mostly at higher frequency.

When HEPA filters above the table are off, the noise also goes down, this might suggest that we still have scattering light problem.

 We are hunting for the noise source for our experiment. We want to see if our setup is limited by intensity noise or not

so, we try to change the power level and see the beat signal.

 I used 3 levels of power (measured after PMC) to be 10 mW (original value), 20 mW , 2mW, (the settings are listed below.)

The reason that the level at 10 mW decreases from previous beat msmt is that the gain for FSS and AOM loops are optimized

PDH box for AOM loop is also modified (R12, 1k -> 10k), the power into both cavities are adjusted to be about the same.

 The noise level at 10 and 20 mW are almost the same except a hill at 60 kHz on 20mW, and

  the noise from 2mW setup is lower than other settings at high frequency, starting ~ 20kHz.

 At 1 mW, I can't lock the PMC (the transmitted beam on the PD is so faint), so I increase to 2 mW to make locking easier.

At 2mW power, the range for feedback signal for PLL loop becomes smaller.

If it exceeds approximately +/- 50 mV [3.5kHz], the lock loses, but I still keep the frequency range at 100kHz.

I tried to increase the gain on SR560 which is a servo for PLL loop, but

the beat signal was worse at high frequency, starting at 2kHz.  Fortunately, the temperature drift at night (~10mV/min)is much smaller than

the drift during the day(~5mV/sec), so I can wait long enough to measure the beat at low frequency, with the gain level of 1, as usual.

At 2mW setting, I have the HEPA filter on and off in comparison. The noise level decreases a bit when the fan off, this might suggest that

scattering light is still a problem.

Even though the noise from beat measurement change with the power, the amplitude does not go with the power change.

i.e, the noise in the beat note does not go up with the same factor of the power change.

So the setup is not limited by intensity noise.  

setting for the measurement ["-" means the same as previous value]

                                      10 mW              20 mW             2mW

PMC: common gain         16 dB                 -                     30dB                       

PMC:RF                            5.7V                 -                       -

FSS:FAST gain               18.5 dB              -                      -

FSS:common                  22 dB               18 dB             20dB   

FSS:RF                           10 V                  -                      -

PDH for AOM:GAIN        4.45 (knob)       3.2                 7.0

Pin RCAV/ACAV            2.1/2.1 mW        4.8/3.9 mW    0.4/0.4 mW

I added a Faraday isolator after PMC and 35.5 MHz broadband EOM, now the noise becomes less susceptible to the

input power to the cavity.

The isolator is installed between the 35.5 MHz EOM and a lens just before the beam splitter that splits the beam into

ACAV and RCAV path. I measured the beat signal at 10mW (measured in front of the BS, as before).

The beams going into both cavities have the same power level ~0.4 V, see blue plot.

  Then I realized that I had not re-aligned the beams into both cavities. The isolator significantly

alters the beam paths, I aligned the input beam, then measured the beat noise again, at 4mW and 2 mW, see green and red plots.

*I'm very surprised to see no significant difference between before and after alignment.

All new three results are very comparable to 2 mW measurement before the isolator was installed, see pink plot.

However, if we compare 10mW from today and yesterday measurements, there is a significant broad band effect at high frequency.

By adding the isolator, we reduce the back reflected power to the PMC, and the laser. The power dependent noise we saw before

might come from this back reflected beam to the laser.  Tomorrow, I'll try adding an EAOM.

The current PD for the 160 MHz beat signal is 120MHz. We use a 2 GHz PD to compare the results between two PDS,

and there is no difference. 120 MHz PD seems to be working fine for us. However, the beat signals at freq above ~5 kHz we have seen so far are not real signal from cavities' noise.

We have checked several parts on the PSL setup to search for excess noise, we have not checked the PD for beat signal, so

we try this measurement.

We use a 2 GHz PD to see the beat signal from another port of the BS.  The attached figure has

two traces of the beat signal, the one on top from 120MHz PD and the one on the background from 2GHz PD . 

The results are similar up to ~10 kHz.

The difference at high f comes from different bandwidth and gain setup for PLL loop, because

it changes with gain setup on SR560. So, the beat noise results shown so far are valid only up to 10kHz.

At higher f, it just the PLL loop.

some additional information:

the beat noise was measured as the feedback signal to the VCO of the PLL, so the calibration factor does not change with changing optical power, alignment, mixers etc.
It's a convenient way to change individual things in the setup and be able do directly compare the measurements without lots of calibration.
We checked the following things:

• power level on beat PD
• different PD with much more bandwidth
• different mixers

The feedback signal is only valid until about 10k. Tara will measure the UGF again on Monday but the signal above 10k exactly scales with power on PD or gain settings while below it stays constant which is exactly what we expect when having enough loop gain in the PLL loop.

Looking more into detail in the spectrum we looked from some tens of Hz to 10k and then tried to excite the spectrum. We could clearly identify the individual resonance peaks from e.g. the beam splitter mount of the beat setup. We know that the way it is set up is very bad but nevertheless we expect mainly lots of resonance peaks but not this hump shaped spectrum.
Now the interesting part is that if we excite the surface of the optical table by touching it softly with e.g. a balldriver we can excite those resonances. What i found very interesting is if you excite the bottom of the table we excite the hump very broadband. You only have to touch it barely like tipping with your fingertips and the whole hump increases. So my guess is that we have a scatter source somewhere which would also explain the shape (at least from my experience).

So the plan for Monday is to have a closer look on that, then checking all the wholes in the foam insulation and probably make them little bigger (right now they are 1/2 inch). Other things are reducing the power from the laser (something we wanna do anyways in the long term), replacing the PMC which scatters a lot of light (you actually don't need an IR viewer to see that, detector card is enough) by a good one and getting a symmetric layout with two periscopes (simple ones like we have now on the other side of the cavity, we casn replace them later by real good ones) for the beat at a lower beam height  to reduce all resonances to better see where/what the underlying noise floor looks like.

Today we 1)replaced out PMC with DMASS' PMC and get better transmission efficiency, 2) added EAOM to modulate the laser intensity

3) measured TF to see how RIN couples into laser frequency shift, it is small and not the current limiting source for now.

Our PMC is not very clean and get the transmission only ~60-70%. The PMC we got from DMASS is much better, now the transmission is up to ~80-90%, we have not

align it carefully yet.  After replacing the PMC, the beat noise did not change.

The Faraday isolator was re-installed and optimized, I used the wrong side before and dumped the beam inside the isolator, instead of outside.

A PBS and a 1/2 wave plate were installed after the PMC to adjust the power without changing the power input of the PMC, the beat noise gets higher

a bit.  I mounted the beam splitter for beat signal on a more rigid post, and aligned the two beams, then measure the beat noise.

See plot.

 Intensity modulation set was installed. The set consists of a 1/2 WP, an EAOM, a PBS. Then we amplitude modulated by sending a sine wave to

the EAOM, and measured the TF between (The excitation is sent to the EAOM)

1) intensity modulation and ACAV_trans_PD

2)intenisity and RCAV_trans_PD

3) intensity and beat noise

4) ACAV_trans_PD and VCO feed back to AOM

extra: we measure the TF between PMC_trans_PD and ACAV/RCAV_trans_PD, since the line width of PMC is much larger than that of RCAV/ACAV,

we can measure the TF of their poles.

 [The plot will be posted soon]

Today we

1) checked if LO phase noise dominates at high f. It turned out that it is dominate at frequency above 1kHz.

2) found out that  RFPD for current RCAV has no 35.5 resonsnce,The calibration for input range at 10kHz is 7.55 kHz/ V

3) tried to float the table, but the pressure is not enough. The gauge reads 2.5 bars.

As we are hunting down various noise sources in our setup, we have to check every components of the setup.

1) As LO phase noise for the current setup (carrier:160MHz, input range:100kHz )has never been measured,

   we tried to see if it actually the limiting source. By  reducing input range of the Marconi, we might see the change of the beat spectrum.

     The regular setting for input range is 100kHz, which was verified to produce real beat noise upto 10kHz.

we changed to 10kHz input range, and measured the beat noise at gain x5, x10, and x20. 

 Since we checked that at gain 10 and gain 20, the noise spectrum, upto 3kHz, does not change with the gain setup.

We can be sure that, at gain x20, the signal we measure up to 3 kHz represents the real beat noise, see plot 1,

the red circle shows where the noise spectrum at gain x20, still similar to that of x10..

The calibration for 10kHz input range is 7.55 kHz/V.  We see that at high f from 500Hz and above, LO phase noise is dominating.

[Wed Dec 08 13:39:24 2010 ]

I measured the beat noise spectrum from 1 - 3 kHz and plotted it on top of the usual noise budget. see fig 3.

2) We measured the TF of RFPD for PMC, RCAV, and ACAV. 

 By modulating the intensity of the laser with the EAOM, we can measure the RF response at the RFPD for each cavity.

The source out is split into two paths, one to EAOM, another one is used for reference.

PMC-RFPD has a resonance peak around 18.2 MHz.

ACAV_RFPD has a resonance around 36.3 MHz,

RCAV-RFPD does not have a peak at 35.5 MHz as it should have.

We will try to switch RCAV RFPD with that of ACAV to get some extra gain. This should give FSS loop some extra gains.

I'll bring it to 40m and use Jenne's laser to test it again and measure a proper TF. Besides, EAOM might not work well at that high frequency.

2.1) Just for fun, we measure the beat noise spectrum without the PMC installed (we removed it when measuring the TF of ACAV/RCAV RFPD), there is no significant change between PMC and no PMC. see plot 2

3) we tried to float the table. Alas, the 2.5 bar pressure from the valve is not enough. Plus one of the table leg (the one that is close to the fume hood) seems to be broken.

The air suddenly leaks from that leg after we turned on the pressure for 30 mins.

 I switched the RFPD between RCAV and ACAV, now the gain for FSS loop is set to 26 dB, but the beat signal does not change.

From previous elog entry, RCAV's RFPD does not have a peak at 35.5 MHz, and ACAV's RFPD has a peak ~36 MHz. And the FSS loop did not have enough gain,

so I switched RFPDs between both cavities. Attached pictures below show the error signal from RFPD when the cavities are scanned, before and after I switched them.

The power into the cavities are ~ 1mW for both cavities. 

The phase is adjusted to produce the best error signal.

Phase adj were 4.16 then changed to 5.67 V

I also inverted the phase on PDH box for ACAV. 

FSS gain can be reduced from 30(max) to 26 dB. 

Then I measured the beat noise, with 100 kHz input range, gain 5. There are no change compared to before.

I widen the beam holes on the outer foam box.  Now all the holes are ~ 0.75" in diameter.

We are concerned about the scattering on the insulation, so we decided to increase the holes' size.

I used a heat gun to heat up a steel rod to melt the foam panel. 

As the insulation box was opened, I checked for scattering light inside the box. I did not see any scattering ligth except

on the ACAV inner insulation opening.

Also, there is a little scattering light at RCAV's insulation cap, so I measured the beat noise when I opened RCAV's inner cap.

There is no significant change, see fig 1. The trace with higher noise was taken when I opened the cap, probably because of too few average.

The scattering on RCAV cap might be negligible for now.

Today we increased the size of the beam path on inner insulations for ACAV and RCAV, cleaned vacuum chambers' windows, and optics on the table.

We have suspected that scattering light dominates the noise  spectrum at low frequency, so we checked all possible scattering sources from the vacuum chamber.

We increased the size of the opening for the beam paths on ACAV/RCAV insulation caps.

We removed the cap by cutting the AL tape around, after that the caps are just placed back. We haven't taped them yet.

 ACAV's chamber windows are especially dirty, full of scratches, so we cleaned them with methanol.

We also found a reflection coming back from ACAV. The spot size is quite large, ~order of cm. The position is quite far from the incoming beam.

The source is still undetermined.

Optics on the setup are cleaned. They had dust, finger prints on them.

ACAV's periscope was readjusted, so that the beam hits near the center of the mirrors, not close to the edge of the mirror.

 RCAV temperature sensors are fixed and not noisy anymore.

Plots and pictures on the board outside the lab are updated.

As we are waiting for temperature to settle, we clean the lab.

After we fixed the insulation, the noise spectrum at low frequency became better. 

I plotted the result with LO noise. The LO phase noise is now divided by sqrt of 2 for beating two LOs together.

* the legend in the plot is wrong, it should be for "100kHz input range" not 10kHz

 The measurement was done between IFR2023A and B, they probably have

different noise level. so the curves do not exactly match.

For the red plot, the data was taken before I cleaned the lens behind the 2nd faraday isolator, and re-alinged the beam to RCAV.

After that I took the data for blue and green plots. I'm not sure why they are not the same, probably because Frank and Peter were walking around the lab. 

Peaks at 3.75Hz, 8 Hz  are cavities' pendulum resonance frequency. When I push the table, the peaks get higher.

I think dirty optics might still cause extra noise. I'll try to clean ones that are still dirty, and see if there will be any improvement.

Jan and I installed one of his seismometers on the PSL table to get some seismic data for the noise projection.
We ran three cables (x,y,z) to his lab trough the door connecting both labs. Jan is taking some data.
Once the device is in equilibrium (thermally) we can also perform some correlation measurements.

I designed the layout for optics behind the cavities for beat measurement, and calculated the mode matching for the beam.

Since the current optics height for beat is quite high (7 inches), we want to lower it to 3 inches, make it more symmetric, and more compact.

The PD's diameter is 300 mm, so the beam spot on it will be ~50um.

All the lenses I need are prepared.

Beat measurement optics' height is changed to 3". I cleaned all optics already, but I couldn't really clean 1/2 and 1/4 wave plates, one of the f =200 mm lens is quite hard to clean.

I'll wait and ask someone before trying to clean again. I cannot lock both cavities at the same time, once I can, I'll align the beam on the PD.

Also ACAV's PD for ACAV_trans_PD is broken. It gives out 11 V regardless of the beam falling on the PD, so I replace it with a PD that is used for NPRO_PWRMON.

Both cavities are locked at the same time. The temperature setting are, RCAV = 34.95, ACAV = 37.2.

I realigned the beam onto the PD to get maximum contrast. I'll readjust the setting back to the original value

and see if the beat noise is improved.

I just notice that one of the beam on the mirror on ACAV's path behind the cavity is almost clipped. I'll readjust it tomorrow.

I measured the beat noise after I realigned all optics behind the cavities. The power has not been reduced to 1 mW yet.

This is just a quick measurement to see where we stand (red curve). The noise gets worse compared to the best measurement (green) before the optics behind the

cavities are rearranged, but the mechanical peaks around 1kHz are suppressed significantly.

I modulated laser power to see how it causes frequency noise in the cavity, no sign of RIN induced frequency noise has been seen yet.

We have been thinking about how RIN will cause random absorption on the mirror surface and turn into thermo elastic/ thermo refractive noise in the cavity.

So we try to observe it, by modulating the laser power with an EAOM. I used white noise, generated by SR785 at 2.5V, as a source to modulate the intensity.

  The power input to both cavities are 3mW. This corresponds to ~15mW total power before the BS (3 to RCAV, 12 to ACAV path).

1)The intensity changes as we can see from ACAV_trans_PD (fig1).

The lower curve is the power fluctuation of the laser without the modulation. The upper curve is the power fluctuation with the added white noise.

So we are sure that intensity does change from our modulation.

2)Then I measured the feedback signal to PZT and to AOM, with and without external white noise.  These signals represent the frequency change of the laser.

However no change observed in both signals. I plotted only the feedback signal to VCO(to AOM), since this signal should be more sensitive to change than feedback to PZT, as the

beam to ACAV path is stabilized to the RCAV already.

3) There is no observed effect from RIN on beat measurement yet, see fig 3.

The laser power output is ~54 mW, so the maximum power we can get to each cavity equally is ~10 mW.

I will try to increase more power to the cavity, as the effect should be proportional to the input power

Actually, we don't need both cavities to have the same power in order to measure the feedback signal. I can try

1) lock the laser to RCAV only with power of 10mW or upto 40mW and measure the feedback signal to PZT

2) lock both cavities, have maximum power to ACAV (could be upto 12,13 mW, the efficiency is ~25%) and measure the feedback to VCO

To get a sensitive measurement, you should measure a coherent transfer function, not just noise. We expect the thermo-elastic effect to be linear. Try making a swept-sine measurement and fiddle with the integration time to get a measurement. Even it ends up being a very small signal, you can use the measurement parameters and the HP Application Note #243 to set an upper limit on the coupling.

I'm measuring the TF between ACAV (ref) and  feedback to VCO (response).

The TF looks ok, but I still need to verify by increasing the laser power and see if the TF goes up with the same factor.

We are determining the coupling coefficient from intensity to frequency change via thermo-optic effect.

Using a swept sine measurement with 4000 integration cycles, from 10Hz to 100kHz, to measure the TF 

between A) ACAV RCTRANSPD (ref) and B) feedback to VCO(response), we can see some reasonable results.

I tried 2000 integration cycles with feedback to PZT, but it seems the laser is too noisy. I might try longer integration time later.

To make sure that the measured TF is real, we must check that when the power to the cavity is increased, the magnitude of

the TF should increase by the same factor. 

Once we have the TF between ACAV_RCTRANSPD and VCO feedback, we can calculate how RIN would couple in to frequency noise by

Thermo optic noise (RIN) [Hz] = VCO calibration [Hz/V] x TF between VCO feedback and ACAV_RCTRANSPD [V/V] x

                                                    VCO TF [V/V]    x RIN x DC level of ACAV_RCTRANSPD

Note the TF of the VCO box (VCO TF) is measured and fit. One pole is at 1.25 Hz, and one zero is at 38 Hz. see fig 1.

 VCO's calibration for V to Hz can be found here, The temperature is quite stable and VCO_MON has been around -0.7 to -0.3 V, we can approximate to be 1.4 MHz/V for now.

we replaced the photodiode in the RFPF fro the reference cavity (RCAV). The old one looked like shit. Below pictures of the old and new photodiode.

 old photodiode - picture 1

old photodiode - closeup view

new photodiode

The TF of 2 35.5MHz RFPD and 21.5 MHz RFPD are measured by the Jenne laser. RCAV's RFPD has the better response, while ACAV's RFPD peak is slightly shift to 37 MHz. But the peak is low Q so we lose only a few dB. PMC's RFPD is now tune to have a resonance peak at 21.5 MHz and a notch at 43 MHz as it should be.

I adjusted the temperature on ACAV and RCAV so that both cavities can be locked simultaneously.

I measured the beat noise and see improvement at low frequency after we fixed the RFPD.

After we reinstalling the RFPDs on the table, we need to wait for the temperature to settle before both cavities can be locked.

(The temperature readout jumps by ~0.05 degree if we work on the table. The cause is not known yet. we tried pulling the cables for T readout,

but this does nothing to the T readout.  This problem will be investigated soon.)

I check the beat noise and see some improvements at low frequency . It does not look a lot for now, but I think it might go lower

if I re align the beam to the cavities and the beam for beat measurement. It might drift away a bit during the break, and I haven't checked that yet.

The input power is ~ 1 mW for both cavities.

RCAV gain = 25

ACAV gain= 8.0 on the knob.

I used PLL feedback to measure beat noise from dc to 3 Hz and dc to 1.5 Hz to compare with the data taken yesterday by FC.

There is no peak around 3 Hz (probably a peak from cavity suspension)in FC data, I'm not quite sure if we what we see is real or not.

changed the frequency noise readout to the cable-delay version to see how it works.
As the loss of the cable is very large and the signal from the photodetector not very strong i had to add some amplifiers for signal conditioning (see figure below).
Data is acquired with channel C3:PSL-FSS_FREQCOUNT in replacement for the frequency counter. Started taking data around 11/2/4 4:16:30 UTC.
Will do final calibration tomorrow. First test gave signal from mixer changes from peak-to-peak for about 600kHz (+/-10%) in frequency change.