40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  PSL, Page 35 of 52  Not logged in ELOG logo
ID Date Author Type Category Subject
  895   Mon Apr 2 22:33:21 2012 taraNotesNoiseBudgetparameters for coating brownian noise calculation

Coating Brownian noise with uncertainty (worst & best case scenarios)

 nb_with_uncertainty.png

I use the parameters found in the literature for coating Brownian noise calculation. 

The upper limit ( high noise level)  has

  • loss (silica) =1.2 e-4,
  • loss (tantala)= 4.6e-4,
  • Young's modulus (Tantala)= 186 GPa

The lower limit (low noise level) has

  • loss (silica) = 0.1 e-4;
  • loss(tantala) =3.6 e-4;
  • Young's modulus (Tantala) = 102 GPa

The rest of the parameters are their nominal values. The max/min values are ~ 18% from the average level. @ 100 Hz, the average noise level is 4.097 mHz/rtHz. The upper limit is 4.815mHz/rtHz, the lower limit is 3.324 mHz/rtHz.

Quote:

I make a list of parameters found in literature. This will be used for estimate the coating Brownian noise level and its error.

 SiO2 loss:

  • 1 +/- 0.2 x10-4  (Crooks et. al. 2006)
  •  0.4 +/- 0.3 x10-4 (Crooks et. al 2004)
  • 0.5 +/- 0.3 x 10-4 ( Penn et al. 2003)

Ta2O5 loss:

  • 3.8 +/- 0.2 x10-4  (Crooks et. al. 2006)
  • 4.2 +/- 0.4 x10-4  (Crooks et. al 2004)
  • 4.4 +/- 0.2 x 10-4  ( Penn et al. 2003)

 * the values reported by Crooks etal in 2006, are supposed to be more accurate than the results in 2004, because of the better estimation of energy stored in the coatings and the corrected thermoelastic contribution. They mention that the Poisson's ratio has small effect on the level of the estimated thermal noise.

The numbers from Penn et al are extracted from the multilayer coating ringdown measurement. Since they measure the ring down of the coating which has both materials. The values depend on Young's moduli of the materials as well. They use Ysio2= 72 GPa/ YTa2O5 = 140GPa. Thermoelastic loss is not taken into account.

 

The values for Young moduli are usually measured directly with nanoindentation technique.

 SiO2, Young modulus (Thin film)

  • 72 GPa (used in Penn et al 2003)

Ta2O5, Young modulus (Thin film)

  • 140+/-14 GPa  (Alcala et al, 2002, nano indentation,stacks.iop.org/Nano/13/451)
  • 140 GPa (Martin PJ et. al.,1993 Mechanical and optical properties of thin films of tantalum oxide deposited
    by ion-assisted deposition Thin Films: Stresses and Mechanical Properties IV, Mater. Res. Soc. Symp. Proc.). This source is not available by caltech connect. This source is cited in the paper by Penn et al, 2003.
  • 144+/- 42 GPa (Crooks et al 2006, this assumes that Young's modulus of SiO2 is 72 GPa)

                      

SiO2, Poisson's ratio:

  • 0.17: (Bamber, et al,2001), but they use SiO2 for calibration and assume that its Young's modulus and Poisson's ratio in thin film form are known.

 Ta2O5, Poisson's ratio

  • 0.23 (assume that it is the same as in bulk, Crooks, 2006)

The uncertainties in Poisson's ratios of the materials have small effect on the coating noise level. For examples, the 10% increase of SiO2's, and Ta2O5 Poisson's ratios, causes the thermal noise to increase by 0.09%, and 0.06%, respectively.  

list of all materials' properties,here.

 

  894   Fri Mar 30 00:50:43 2012 frank, taraDailyProgressSeismicSeismic coupling from curtain frame to beat

The coupling from frame motion to beat signal is now significantly reduced after we fixed the cable wiring .

 

      The plot below shows the acceleration measured on the frame and the corresponding beat signals. I did not calibrate the signal since this is just a qualitative check.  I tried to excite the frame so that the signals from the frame motion were comparable (blue and cyan). After we fixed the cables wiring, the beat signal (pink) are less sensitive to the frame vibration than before (red).

      Note that the set up were similar in both measurements. The input range for PLL was 5kHz, with gain 200.

compare.png

 

I have not measured the beat signal after we fixed the cable yet. It will be interesting to see if the beat has been improved or not.

  893   Thu Mar 29 20:36:59 2012 FrankNotesFSSmodifications

Made a few modifications to the TTFSS servo to get a better performance at low frequencies as the suppressed laser noise is very close to the expected CTN (or even above below 10Hz):

Replaced a few resistors in the input stage with thin film ones (schematic D040105-C) and added some more gain as we usually rail the common gain knob at the moment:

R2 -> 49.9R
R1 -> 49.9R
R7 -> 49.9R
R4 -> 402R
R3 -> 56R (metal film resistor on socket to make this adjustable. will be fixed once we know exactly how much more gain we need)

original gain of the first stage was 3.16, new gain is 7.17. Monitor outputs were limited by high thermal noise of the original values of the resistors R1 & R2 (and so our measurements).

I've also added a switchable boost stage. There is only one opamp in the common path (the first adder stage right after the mixer). As this is also used to measure transfer functions i didn't like to modify that to avoid more complicated calibrations every time we measure something. Instead i've added it to the fast path using U7. Crossover-frequency between fast actuator and PC is ~10kHz, so can't go really high there. Original feedback resistor is 5.6k (R29) , so i've added a 4.7nF cap and added a switch to the TTFSS box (on the side, easy to reach and  to not destroy the front panel). Integrator frequency is switchable at ~6kHz.  Servo was tested and working fine.

Complete TF not measured yet since we are still working on the acoustic and seismic isolation of the table. Full characterization is planned to be done on Friday.
 

Attachment 1: D040105-C.pdf
D040105-C.pdf D040105-C.pdf D040105-C.pdf
  892   Thu Mar 29 18:50:51 2012 taraNotesNoiseBudgetparameters for coating brownian noise calculation

I make a list of parameters found in literature. This will be used for estimate the coating Brownian noise level and its error.

 SiO2 loss:

  • 1 +/- 0.2 x10-4  (Crooks et. al. 2006)
  •  0.4 +/- 0.3 x10-4 (Crooks et. al 2004)
  • 0.5 +/- 0.3 x 10-4 ( Penn et al. 2003)

Ta2O5 loss:

  • 3.8 +/- 0.2 x10-4  (Crooks et. al. 2006)
  • 4.2 +/- 0.4 x10-4  (Crooks et. al 2004)
  • 4.4 +/- 0.2 x 10-4  ( Penn et al. 2003)

 * the values reported by Crooks etal in 2006, are supposed to be more accurate than the results in 2004, because of the better estimation of energy stored in the coatings and the corrected thermoelastic contribution. They mention that the Poisson's ratio has small effect on the level of the estimated thermal noise.

The numbers from Penn et al are extracted from the multilayer coating ringdown measurement. Since they measure the ring down of the coating which has both materials. The values depend on Young's moduli of the materials as well. They use Ysio2= 72 GPa/ YTa2O5 = 140GPa. Thermoelastic loss is not taken into account.

 

The values for Young moduli are usually measured directly with nanoindentation technique.

 SiO2, Young modulus (Thin film)

  • 72 GPa (used in Penn et al 2003)

Ta2O5, Young modulus (Thin film)

  • 140+/-14 GPa  (Alcala et al, 2002, nano indentation,stacks.iop.org/Nano/13/451)
  • 140 GPa (Martin PJ et. al.,1993 Mechanical and optical properties of thin films of tantalum oxide deposited
    by ion-assisted deposition Thin Films: Stresses and Mechanical Properties IV, Mater. Res. Soc. Symp. Proc.). This source is not available by caltech connect. This source is cited in the paper by Penn et al, 2003.
  • 144+/- 42 GPa (Crooks et al 2006, this assumes that Young's modulus of SiO2 is 72 GPa)

                      

SiO2, Poisson's ratio:

  • 0.17: (Bamber, et al,2001), but they use SiO2 for calibration and assume that its Young's modulus and Poisson's ratio in thin film form are known.

 Ta2O5, Poisson's ratio

  • 0.23 (assume that it is the same as in bulk, Crooks, 2006)

The uncertainties in Poisson's ratios of the materials have small effect on the coating noise level. For examples, the 10% increase of SiO2's, and Ta2O5 Poisson's ratios, causes the thermal noise to increase by 0.09%, and 0.06%, respectively.  

list of all materials' properties,here.

  891   Thu Mar 29 00:46:33 2012 frank, taraDailyProgressSeismicSeismic coupling from curtain frame to beat

We rearranged the cables and the setup a little bit to isolate the table from the frame.The work is still in progress.

 

We removed some unused cables that ran from the frame to the table, and rearranged some equipments. They are:

  • PMC RFPD DC out, PMC TRANS,
  • 4 CCD-camera cables, which are now on the table only.
  • Laser cables (Fast feedback, Slow control, laser control, laser power). The laser controller is now on the table.
  • Marconi for 14.75 MHz sideband is now on the curtain frame.
  • 35 W laser and the unused vacuum chamber are removed from the table and rest on the floor.
  • The hose for air spring will be replaced. It is also connects between the air springs and the frame directly.

IMG_0687.JPG          IMG_0688.JPG

  890   Tue Mar 27 23:32:22 2012 frank, taraDailyProgressSeismicSeismic coupling from curtain frame to beat

We are trying to decouple motion from the curtain frame to beat signal. Currently, just by tapping the frame, we can see the effect on beat signal clearly.

The cables running between the table and the frame are very likely to transfer the motion from the frame to the table causing extra acoustic noise in the beat. So we want to see the improvement by removing some of the unused cables. First, we need to check the current coupling from the frame motion to the beat. An accelerometer is installed on the frame (see picture). We tried to drive the table with a PZT, but it failed. The PZT is not strong enough to shake the frame at 5 Hz. As an alternative, we will just tap the table and see the motion on the frame, on the table, and on the beat. Then we can estimate the coupling (at least for the resonant frequency).

 IMG_0681.jpg

  889   Tue Mar 27 17:01:23 2012 taraNotesNoiseBudgetCalibration verification for PLL

I check the calibration factor used in PLL measurement to make sure that we have used the right calibration factor. The result is fine. The calibration we have been using is correct.

     From the beat measurement, the total noise budget is very similar to the beat signal at high frequency with some offset. So we suspect that the calibration for LO frequency noise / beat we have might be wrong. 

     I set the tuning range to be 10kHz(marconi), with gain 200 because this range has frequency noise large enough to be seen on beat signal, and I have the data for this setup from PSL:874.

     The calibration for beat signal with 10kHz input range is 7e3 Hz/V. For 1kHz input range, we have been using 700 Hz/V (rescaling with the tuning range). The direct measurement using a voltage calibrator + frequency counter gives 712 Hz/V, see figure below. The calibration remains the same with 160MHz or 100MHz carrier frequency.

cal_1k_pll.png

    Electronic noise from PLL (straight blue plot with ~f dependent) is also added for completeness. The level measured after SR560 is 40uV flat, the OLG TF of PLL is 4.44e4/ f, the electronic noise from PD can be calculated following the instruction in PSL:816.

it seems that we have a good agreement between the measured beat and the sum of 2 limiting sources at high f.

freq_noise_cal.png

 

Note that the frequency noise from Marconi is measured with 160 MHz carrier, while our beat is currently ~ 180 MHz. However, the carrier does not change the frequency noise that much (cf PSL:874).

To sum up, the calibration we have been using is correct. The mismatch between the beat signal and the total noise budget is probably be the other noise sources we have not taken into account yet.

 

  888   Mon Mar 26 19:51:41 2012 taraNotesNoiseBudgetThermal noise vs spotsize and cavity length

I calculated thermal noise from different mirrors to find out what factor we might gain compared to the current thermal noise level. For the possible design with 0.5m-0.5m mirrors with 1.45" cavity length, we gain a factor of 8.6.

factor_ROC_length.png

 Fig1: The factor gain for Coating noise with different mirrors ROC combination and cavity length. note: the title of the plot should not contain "R=0.5, concave-flat".

It turns out that using flat-concave style will not be the winning solution, we win more with 0.5m-0.5m mirrors with the same cavity length.

 

The code for this calculation can be found on svn

  887   Wed Mar 14 17:31:51 2012 taraSummaryNoiseBudgetThermal Noise calculation using gwincDev

The discrepancies between in GWINC code and our code are  different values of phi_perp/ phi_perpendicular, and the simplification of the formula.

 

The coating thermal noise from Rana's plot is about a factor of sqrt(2) lower from the usual code we have been using. The reasons are that

  1. We use different values of phi_perp and phi_para. Values from GWINC phi_perp/phi_para are 1.3e-4/1.7e-4,  ours are 1.7e-4/3.1e-4. This is the main reason why the results do not agree
  2.  Our code uses a simplified Thermal noise formula, where all the terms with 1- sigma are ~ 1 see [fig.1], but another simplified version[fig.2] that keeps one more term of 1-sigma gives a result that agrees with that from GWINC.  With the same parameters, the better simplified version (with term 1-sigma) gives the result smaller than that of the exact formula by only ~ 1%.
  3.  Also, the temperature we use in the code is 35 C, while the temperature in GWINC is 20 C. This contributes to 5% reduction from our result.

 

Harry_etal_2002.png

fig1: excerpt from Harry etal. Class Quantum Grav 19 (2002) 897-917

Harry_etal_2005_p060072.png

 

 fig2. excerpt from Harry etal. www.ligo.caltech.edu/docs/P/P060072-00.pdf

  886   Wed Mar 14 00:21:42 2012 FrankSummaryNoiseBudgetThermal Noise calculation using gwincDev

i will get a copy from Jamie tomorrow and edit our current noise budget file to replace our simple calculations with the updated gwinc calculations. I have to update the plot for the proposal anyway and want to make it look good (and right). The current version of the proposal contains a placeholder only.

Quote:

  I have updated the plots using the more correct loss angles for the silica and tantala - attached are the plots. The overall Brownian noise has gone up by ~20% since the noise was always dominated by the tantala before.

In order to be honest, I suggest that we plot the Brownian noise using some of these confidence intervals - the yellow region indicates what I think the 2*sigma limits are.

You'll have to run the code yourself if you want to put these traces into your Noise Budget plots.

 

  885   Tue Mar 13 23:44:02 2012 ranaSummaryNoiseBudgetThermal Noise calculation using gwincDev

  I have updated the plots using the more correct loss angles for the silica and tantala - attached are the plots. The overall Brownian noise has gone up by ~20% since the noise was always dominated by the tantala before.

In order to be honest, I suggest that we plot the Brownian noise using some of these confidence intervals - the yellow region indicates what I think the 2*sigma limits are.

You'll have to run the code yourself if you want to put these traces into your Noise Budget plots.

Attachment 1: RefCav_TOnoise.png
RefCav_TOnoise.png
  884   Tue Mar 13 22:06:46 2012 taraNotesNoiseBudgetcalibration verification

The flat noise in beat signal  around 100 Hz now is very likely to be the sum from electronics and in loop noise. Better measurement will be done again to confirm the result.

==motivation==

We are investigating what would be the sources of the flat noise we see in the bet signal. Since it is flat, we suspect that it should be shot noise or electronic noise in our setup.

==setup&procedure==

Before we can say anything about contribution from electronic noise to beat frequency, we need to be sure that the calibration from V noise to Frequency noise is correct. The calibration we need is the slope of error signal which can be measured directly, see PSL: .

Once we know the calibration, we check that it is correct. This is done by locking both cavities, measuring the beat signal. Then reduce the gain in RCAV loop until the feature of the inloop noise shows up in the beat. Then we can compare inloop noise to beat. The result looks fine within 10%.

[add fig1] fig1: Inloop noise convert to absolute frequency using slope of error signal and frquency noise from beat.

 

Now we can use this to convert inloop noise and electronic noise to absolute frequency.

 We also want to see the contribution from shot noise. Lower the power is an option to bring up shot noise. We try:

  1. reduce the power only in RCAV, until the beat signal is flat, check inloop noise & electronic noise & shot noise
  2. reduce the power only in ACAV, until the beat flat, check inloop noise & eelctronic noise & shot noise

Error signal slope from

  • RCAV : 76 kHz/V for 1mW input, and 2.5MHz/V for 30uW input*.
  • ACAV: 157kHz/V for 1mW input, and 4.1 Mhz/V for 38 uW input*.

*I did not measure the slope with lower power, I just scaled down them with power.

beat_rcav_30uW.png

fig2: Beat noise in comparison with shot noise, in loop noise, and electronic noise. RCAV power is reduced to 30uW. Shot noise is corrected for the corresponding input powers.

beat_acav_38uW.png

fig3: Beat noise in comparison with shot noise, in loop noise, and electronic noise. ACAV power is reduced to 38 uW. The peaks in ACAV in loop noise does not show up in the beat. I think it might be harmonic lines in the measurement, and not real.

 

So if I plot the electronic noise and in loop noise from regular setup the result:

 beat_regular.png

 fig4: Beat with regular setting (1mW input to each cavity, modulation depth = 0.14, power to EOM = 0.2dBm). The in loop noise and electronic noise are very likely to be the sources of the limiting flat noise.

 

The electronic noise and the in loop noise are very close to the beat level, by increasing the gain in the detection point, we should be able to suppress this to level. We can try using larger modulation index(current value is only 0.14). Another measurement with better resolution will be done in order to add them in the noise budget.

  883   Tue Mar 13 16:29:43 2012 ranaSummaryNoiseBudgetThermal Noise calculation using gwincDev

I looked around in some of the papers for a better estimate of the Q for the different coating materials. From

http://dx.doi.org/10.1088/0264-9381/21/5/101

I get new numbers (I think the numbers I used in the previous calculation were too modern). The numbers from this 2004 CQG article are

phi_silica  = 0.00004 +/- 0.00003 

phi_tantala = 0.00042 +/- 0.00004

which are lower for silica and higher for tantala than what I used before.

  882   Mon Mar 12 23:31:38 2012 frank, taraDailyProgressBEATbeat with floated table, air spring.

We measured beat signal with the upgraded seismic isolation system. It seems that we are sitting on some scattering noise source at low frequency.

 

      Most of the mirrors mount in beat path, are damped with rubber cones. A lead block is put on the board for beat setup to damp down any resonant peaks. With air springs and floated table, seismic noise is significantly reduced. From the shape of the bump around 70 Hz, we suspect that it is the scattering noise. That will be investigated further.

beat_2012_03_12.png

Fig1:  In grey: beat signal after softer RTV springs were used. In pink: beat signal with air spring, unfloated table. In blue: beat signal with air spring, floated table.

  881   Mon Mar 12 15:28:37 2012 taraHowToNoiseBudgetNoise estimation for delay line technique

I finished the calculation for delay line. The detail is in psl:868. Plus, I edit my plot so that it looks nice ( see PSL:878).

  880   Sun Mar 11 22:49:35 2012 FrankDailyProgressPurchasesindependent air supply

today i bought and installed a small air compressor (3gal, 100psi ~$50)  to be independent from the bottle supply we use so far. We don't have access to a high pressure supply in any of the labs and the normal pressure of 32psi is way not enough (we need about 90psi). As those things make a lot of noise we might want to switch to bottle supply when we take long term measurements to not get interrupted by the noise. I don't know yet how frequently the compressor is running, but from my first tests it's not too often. I hooked it up in parallel to the (currently empty) nitrogen bottle so that we can switch between both. We could also add some acoustic damping. I've put it in the box it came with to test it and this already helps a lot so it might worth thinking about.

P1820863.JPG P1820864.JPG

  879   Sun Mar 11 22:40:33 2012 ranaSummaryNoiseBudgetThermal Noise calculation using gwincDev

 I recalculated the thermal noises using the specific layer structure of silica/tantala that was used to get a 300 ppm reflectivity. This is not based on an REO data sheet, but just based on using lambda/4 layers with a lambda/2 silica cap on top. I used the minimum total number of layers required to get better than 300 ppm. In reality, REO may have used another couple of pairs to get it correct.

 I have used phi_SiO2 = 1e-4 and phi_tantala = 2.3e-4. I've asked Gregg Harry and Matt Abernathy to give us updated numbers.

The noise plot below shows that the thermo-optic noise at 100 Hz - 1 kHz is not insignificant, but also not dominant.

The bad news is that the Brownian noise is lower than we've been using in the usual Noise Budget plots (like Tara's last one). So, if my calculation is correct, we're not as close to the real coating noise as we thought...

Next step is to compare the calculations to find out the difference (this means you Tara, don't let Frank stop you). My code is committed to the iscmodeling CVS at MIT:   iscmodeling/coating/iRefCav/. It requires the whole gwincDev/ tree in order to run.

Attachment 1: RefCav_R.png
RefCav_R.png
Attachment 2: RefCav_TOnoise.png
RefCav_TOnoise.png
  878   Sat Mar 10 18:42:42 2012 taraDailyProgressBEATbeat- with air spring isolation

 Beat measurement with new air springs is measured with unfloated table (the N2 supply is empty). Good improvement can be seen at frequency above 20Hz.

beat_2012_03_10.png

The data is compared with beat signal before the air springs, unfloated table on 2012_03_03.  Once we float the table we should see no seismic/mechanical peaks around 100Hz. Then effects from RFAM/RIN couplings might reveal themselves.

Note about the setup:

  • Power into both cavities are ~ 1mW (I also adjust the power for beams to RFAM measurement PDs to be ~0.8 mW).
  • I replace the plastic legs for RCAV RFPD, so that the height is 3 inch, the reflected beam is dumped properly.
  • The measured Marconi frequency noise from Marconi in this nb is not updated yet. I'll update that soon.
Attachment 1: beat_2012_03_10.png
beat_2012_03_10.png
  877   Thu Mar 8 21:48:46 2012 Frank, KojiDailyProgressSeismicair spring isolation installation finished

got some T-connectors and connected all four springs to the lab air supply. The pressure of 32PSI is barely enough to float the chamber. The side with the horizontal extra flange is almost too heavy to float, but it's OK. The other side is much better. But overall the chamber is pretty good leveled. Tara re-aligned the laser to both cavities and the beam height is ~6.5in as designed.

IMG068.jpg IMG070.jpg

we measured the TF from the table to the vacuum chamber. Sensor is attached to the top of the ion pump in the center (left picture), the other one on the aluminum block next to the chamber (see picture). Resonance frequency is at 5.19Hz and has a transmissibility of ~10. Both numbers are slightly higher than expected from the datasheet.

air_spring_vertical_tf.png

  876   Thu Mar 8 14:56:21 2012 taraHowToNoiseBudgetNote on Noise Hunting

Seem that people have been asking what have we done to reduce the noise in the beat measurement, so I think it is a good idea to summarize it down here as a future reference.

I'm listing the topics for now, the corresponding details will be added in the future.

  •      RFPD for PDH : we had problem with the PDH loop(psl:706). This might be the combination of bad RFPD, small modulation depth, problem with TTFSS.
  •      Modified RFPD (PSL:792, PSL:795), with that the inloop noise is below coating noise level and the loop is not gain limited. (PSL:806)
  •      14.75 MHz resonance EOM: We added a resonance EOM for adding the sideband frequency. The broadband EOM that we used before could not give us enough modulation depth (PSL:745).  Plus, we got rid of the RF summing box which might contribute to weird behaviors of the loop (PSL:711,720). Now the broadband one is only for feedback control.
  •      Lower side band frequency: We lowered the sideband frequency from 35.5MHz to 14.75 MHz, because have better Q at lower frequency.
  •      TTFSS position on the table (PSL:778) : Beofre: lower UGF(33kHz, psl:654)
  •      Polarization of the beam (and RFAM effect)
  •      Scattering light: (dump beam properly, use super polished mirrors, psl:647)
  •      Seismic and Mechanical peaks (seismic, PSL:690), seismic stack(psl:668,837), softer spring(PSL:762,814), float table(psl:696), mechanical peaks (PSL:818,820,827)
  •      Noise in PLL readout:
  •      Put two cavities in the same vacuum chamber: this reduce the thermal drift b/w the two cavities, and allow us to use lower tuning range on PLL, (psl:667), better thermal sheild (PSL:776)
  •      Reduce RFAM: Thermal control EOM(PSL:744), setup(PSL:854)

 

 plan:

bring up coating noise level, (psl:657), using shorter cavity, smaller spotsize.

 

note:

some useful numbers: psl:733

  875   Thu Mar 8 01:15:50 2012 FrankDailyProgressSeismicair spring isolation installed

[Tara, Frank, Zach, Jenne]

i've picked up all parts, tapped a few more holes and assembled the two beams. Zach helped us lifting the (really heavy) vacuum chamber to install the two beams. I had to cut one side of the center insulation (the one with the flange on the side) as it was impossible to install with the chamber resting on the two beams. Even after cutting it it was really difficult to install. There is barely enough space for the other side. The other pieces are easy to install.

Had a hard time to find an adapter for the schrader valve. Have been to 3 bicycle stores, three car supply stores, home depot, osh etc. - the simply don't have an adapter which one can screw on the valve and connect a hose or anything like that. Ended up making one out of some valve extensions. Jenne helped me installing the air supply for one side to check how much pressure we need to lift it. It turns out that the air supply in the lab is sufficient. Will get all missing parts tomorrow from the hardware store to finish installation and measure the TF of the new isolation.

IMG058.jpg IMG056.jpg

IMG063.jpgIMG065.jpg

  874   Wed Mar 7 22:54:06 2012 taraNotesNoiseBudgetIFR2023B frequency noise lock w/ Rb clock

Frequency noise of IFR2023b (locked to Rb clock) at different settings are measured. This data will be used in beat's noise budget. With reference signal from Rb clock, Marconi's noise is lower than coating noise level below 300Hz.

        Noise contribution from IFR2023b/Marconi used in PLL for beat read out has to be updated. The current data in the nb is  from Marconi with no reference signal.  With Rb clock 10MHz reference signal, the frequency noise will be lower at frequency below 1kHz,see ATF:822. Since we are going to reduce the LO frequency noise by reference it with Rb clock,  we need to measure noise level for the nb.

     The settings are:

  • carrier @160MHz, with 3 different input ranges, 10kHz, 1kHz, 100 Hz.
  •  carrier @100MHz,  with 10kHz, 1kHz, 100Hz input range.
  • carrier @ 10MHz, with 10kHz, 1kHz, 100Hz input range.

     The carriers are chosen at different 3 frequencies because we want to check if the noise will be lower or not. As we aim to lower beat frequency down, 100MHz, and 10MHz are our choices.

 ==setup==

IMG_0399.jpg

Both Marconis are locked to the same Rb clock(not shown in the diagram). I use a transformer between the clock and each marconi to prevent ground loop.

From the block diagram, to convert the voltage noise to absolute frequency noise, we simply need to know the calibration factor of the Marconi. This can be obtained by measuring the marconi output frequency vs input tuning voltage.

==result==

freq_noise_ifr2023b_rbclock.png

Data sets from 160MHz carrier are plotted in red tone, 100MHz sets are in green tone, 10Mhz sets are in blue tone.

==Comments==

  • We might gain some reduction in frequency noise for going to lower beat frequency 10MHz, but with 100Hz tuning range.
  • Reference signal from Rb clock does not reduce Marconi noise at 10 kHz input range. It helps only with smaller tuning range (1kHz, 100Hz).
  • For data with 10kHz input range, the shape of 100MHz is different from 160MHz and 10MHz sets. I double check it and it seems real. The UGF is definitely above 20kHz.
  • Peaks in the plot come up when I lock both Maronis to the same Rb clock. They are gone when I don't lock, so the peaks might not be real.
  • Data are available on SVN.
  873   Wed Mar 7 11:39:37 2012 Tara, FrankNotesTempCtrlacav temp tuning - 2nd try

previous measurement is invalid - due to changing in the alignment from floated table to unfloated table over night we don't know what we really measured.

What we can say is that the beat frequency changed from 181.265MHz to 185.9MHz, so by 4.66MHz for a 0.2K increase of the shield using 1.53V heater voltage.

aligned everything properly, locked both cavities, changed tuning input range to 100kHz and turned heater off. We might stop the new measurement at any time as we got the basic information we need

  872   Wed Mar 7 01:39:39 2012 Tara, FrankNotesTempCtrlThermo modulation for ACAV

I turn on the temperature modulation on ACAV. For ACAV loop, I attenuate the signal to marconi with -30dB, reduce the power to AOM to 9dBm (original value is 13dBm) and increase the tuning frequency range to 500 kHz (max @800kHz).

weird shit going on when modulating ACAV temperature with the rad heater. Looks like we have some sort of parasitic cavity. Will redo measurement with smaller VCO input range

On the left: two temp sensors on ACAV shield in degC
bottom right: heater voltage monitor in Volts (temporary stolen channel :-) )
top right: feedback to AOM VCO input in Volts

stepresponse.jpg

  871   Wed Mar 7 00:43:05 2012 FrankHowToNoiseBudgetphase noise and other things

VERY IMPORTANT ! (HAS TO BE MEASURED NEXT BEFORE ANYTHING ELSE):

  • phase noise of Marconi locked to Rb-clock for input ranges 100Hz to 10kHz - current level in NB is too high below 1kHz
  • projection of RF-AM contribution from both EOMs - we know that we are close to that if we are slightly misaligned
  • proper RIN coupling - we know that it does not dominate at the moment as we can change optical power levels without seeing an effect, but we don't know when it does
  • Acoustic coupling estimation - do we need acoustic shields for cavity readout or beat?

seismic is not important at the moment as it will change end of the week anyway

  870   Tue Mar 6 20:48:47 2012 frank, taraDailyProgressBEATbeat is back

After debugging the RCAV loop, we measured the beat signal again.  We still cannot reach coating thermal noise, the noise floor now seems to be mostly seismic related.

 

beat_2012_02_22.png

 Fig1: Beat measurement, grey trace shows the beat measurement from 2012-02-22, floated table. After that we changed the springs on seismic stack. Blue is the beat signal, unfloated table, and Pink is the beat signal with floated table.

 

     The new beat signal has more mechanical peaks around 300 Hz - 1kHz than the old one does. This is probably because I do not properly damp a few mirrors. The input range used in today's measurement is 2kHz. It does not change between 1 or 2kHz input range, so we probably sits on other technical noise sources.

     In conclusion, with the new springs, we can reduce seismic noise in 5Hz- 100Hz bandwidth (see grey and pink traces in the plot). Next step, we will work on damping mechanical peaks in the signal properly, and adding air spring for the vacuum chamber.

    Note: The setup was optimized for unfloated table. The blue trace was measured after we optimized everything, from polarization to beam alignment, both before and after the cavities. When I floated the table, the beam was barely aligned and could not be locked. I realigned the beam with periscopes, and adjusted the beam alignment on beat RFPD before measuring the pink trace.

  869   Tue Mar 6 17:30:27 2012 Tara, FrankDailyProgressNoiseBudgetTTFSS in-loop noise floor

the last couple of days we had observed that our TTFSS in-loop noise floor we can reach is much higher (40nV/rtHz) than it was before (8nV/rtHz), but we couldn't figure out what it was. The general shape was the same. So we kept backtracking what we did the last week and finally ended up with the proper alignment and temp stab of the EOM we use for our PDH sidebands.

It turns out the in-loop noise floor is sensitive of how we align the broadband EOM (PC) we use as a frequency actuator. We can't properly measure RFAM when applying a modulation on it (SNR too low) and are already good aligned. Instead we aligned the EOM properly to the beam and aligned the polarization removing a L/2 waveplate after it. We are using the first PBS of the faraday isolator as our polarization reference (there is almost no space between the parts to put anything else in the beam path). It turns out that there is a slight mismatch in polarization if we do so. So we glued a little cube PBS on a post and used that as a reference instead of the PBS of the isolator.

With the old technique we have about 21uW out of 20mW in transmission of our new reference PBS - if we align the half waveplate with the micrometer screw right in front of our PC we can get as low as 8uW. This tiny difference we can't see with our RFAM detection technique at the moment. Now the noise floor is back to 8nV. My guess is that we create high frequency AM instead of phase modulation which screws up the sensitivity somehow. However we didn't see this on the spectrum analyzer (LF and RF). UGF is back to ~1MHz.

Next we will try stronger modulation using higher voltage and minimize it using RFAM detection.

  868   Tue Mar 6 00:53:02 2012 taraNotesNoiseBudgetNoise estimation for delay line technique

I made a diagram for estimating the noise of the beat signal(or marconi) as measured by delay line technique. This will tell us if this method will be able to reach the coating noise sensitivity or not. The calculation gives the estimated frequency noise of Marconi to be ~ 10 mHz/rtHz, but the measurement has unidentified noise level of 20 mHz/rtHz. We still have not figured out what is the source of the noise.

delay_nb.png

  1. Start with a source, we use a Marconi with signal = 10 dBm,noise = -150 dBm.
  2. An amplifier with Gain = 16dB, NF = 8dB. S becomes 10+16 = 26dBm, N becomes -150 +16+8 = -126 dBm.
  3.  We use 4 ch splitter, thus the power on each channel is smaller than the input by a factor of 4 (-6dBm): S/R =+20/-132 dBm 
  4.  For the signal that goes to LO side of the mixer(13 level), it is attenuated to 13 dB. From here, we will assume that the signal from this side is perfect and does not contribute to the output.
  5. For the delay line side(-9dB), the S/R are +11/-141 dBm. It should be ~10dBm, so 11 dBm is fine.
  6. The mixer has conversion loss ~ 6dB, the signal is attenuated by -6dB, S=11-6 = 5dBm. Noise is also attenuated by -6dB, but the double sided frequency will be folded to one sided f due to the mixer. So the noise power is increase by a factor of 2(+3dB)*. The final S/R are +5 dBm and -144 dBm respectively.

* This is under the assumption that noise from both sides are incoherent. However, both sides are from the same source with 200 ns delay. The coherent time and line width (FWHM) of the signal is related by FWHM = 1/(pi*Td), see, coherence time. For upper limit approximation, let the Maroni's FWHM to be 10kHz (it should be smaller than the tuning range as well, even down to mHz linewidth), then the coherence time will be 10 micro seconds. So with 200 ns delay time, both sides should still correlate. We have to be careful about how to calculate noise at this point.

Note: I use ZX05-1MHW-S+ mixer for this setup.

 For calibration:

I check the slope of the free running signal. I use two Marconis. One for LO side, it is set at 160MHz,13 dBm. Another one for RF side, @160MHz 10dBm. The slope is 1V/ rad. The peak to peak is saturated at 1.2 V pk-pk.   To change to frequency noise, multiply by 2pi*Td*f, where Td is the delay time in the cable (206nS for our setup), so the calibration is  1.26 * f [V/MHz]. 

From step 6, noise power is -144dBm =  14nV/rtHz. with the calibration factor this gives  14e-9 / (1.26 V/MHz) = 11mHz/rtHz.

 ==comparison with measurement==

The setup is similar to the above figure. The amplifier at the end is SR560.

delay.png

  1. Determine the noise floor of the setup: Noise level of the system is measured by terminating the LO side of the mixer with 50 Ohm. It is around 5 mHz/rtHz (black line). This is lower than the expected signal (purple line).
  2. Measure the signal(frequency noise of Marconi). The oscillator is set at 160MHz, with modulation off. The result is ~20mHz/rtHz (in green). This is higher than the expected level (10mHz).
  3. Verify that the unidentified noise is real: Marconi frequency noise with modulation on, 10kHz, is measured by delay line to check if we can see the frequency noise. This setup is chosen because the frequency noise of Marconi at 10kHz input range is well identified from beat measurement and from direct measurement (plotted in red), and it will serve as a reference for us.  The signal measured by delay line is plotted in blue. It turns out that the feature of 10kHz input range noise is barely seen. It seems that the sensitivity of delay line technique is not good enough because of some unexpected noise source.

 

    The sensitivity of delay line technique is too high to measure the frequency noise of Marconi with 10kHz input range. Unless we figure out what causes this flat noise, PLL is probably a better choice for our readout technique.

 

Attachment 2: delay.png
delay.png
  867   Mon Mar 5 20:41:21 2012 FrankNotesDAQVME crate rebooted

crate rebooted after applying some fixes to the channel database - AOM VCO feedback is now back being monitored (uncalibrated for now as we change the range of the marconi from time to time)

  • chamber temp dropped by ~10mK
  • temp control script restarted, heater set to initial value of 1.75
  866   Mon Mar 5 17:15:03 2012 FrankNotesScheduleair spring isolation

machined all steel parts myself this morning and submitted them for welding to Mike. He said we can pick it up at the end of the week, so if we are lucky we can install the stuff Friday afternoon already.

Questions:

  1. Do we want him to paint them? If yes, which color?
  2. How do we pressurize the springs or measure the pressure?
  3. Do we adjust the pressure simply by leveling the beam which might result in different pressure levels? What is more important, equal springs or a leveled chamber?

Parts missing:

  • 3/8-16 x 0.75 socket head screws and washers (4 sets) for mounting air springs to angles
  • new posts for beat breadboard
  • AL-baseplates to mount springs to table - will be ready tomorrow

New beam height will be ~6.5in  ->  5-7/8 + 3/8 for the beam + 1/4 clearance

  865   Mon Mar 5 17:07:04 2012 FrankNotesSeismicSLM-1A air spring datasheet

- for future reference -

SLM_Series.pdf

  864   Mon Mar 5 16:29:50 2012 frank, taraDailyProgressBEATbeat is back

As mentioned in the previous entry that the in loop noise is higher than before. We investigate what might be the causes. We suspect that mode hopping or relaxation oscillation might be the case, but they are not.

The loop gain for TTFSS cannot be increased up to the usual setup, one thing we notice is that the signal from error point is quite large ( the input range selected on SR785 cannot go down below -30 dBVpk, it used to be -48 or -50 dBVpk)and oscillating at 200kHz. So we are also looking for what causes the oscillation at this specific frequency as well.

1)Checking for mode hopping:      Mode hopping might produce unexpected behavior of the laser. We want to make sure that the   usual SLOW_DC value is not close to where mode hop occurs.

 method: scan PMC and look at the signal from reflected power. The dip from transmitted TEM00 should become shallower when  mode hop occurs because some of the power is lost to other TEM00 mode.

 result: we are no where near mode hopping. The SLOW_DC coarse knob can be adjusted between 200 - 700 easily without seeing mode hopping. The usual set value is around 500 +/-20.

 

2) Relaxation oscillation: (cf siegman, laser, ch 25)

   Since we are looking for what cause the oscillation at 200kHz, relaxation oscillation is a possible candidate. We use a PDA10CS to measure the spectrum of the laser power. We use the beam reflected from the first Faraday isolator just after the NPRO. A broad peak at 462 kHz is observed, so this is unlikely to cause the oscillation.

relax.JPG

fig1: relaxation oscillation is measued (green) with the frequncy at 462kHz. The yellow line is the noise floor of the PD.

    To sum up, we are looking into what degrade the performance on RCAV loop. We ruled out two possibilities which are laser mode hopping and relaxation oscillation.

  863   Sun Mar 4 00:45:26 2012 frank, taraDailyProgressBEATbeat is back

Beat is back.

The beat signal( with the new softer springs) is measured and compared with the previous result (unfloated table). 

several notes about this measurement:

  • The resonant peaks from the stack show up as expected at 10, and 35Hz.
  • I have not added the contribution from seismic in the nb yet since TF between beat signal and seismic is not measured. We will measure that once the temperature settles.
  • Beat noise at high frequency  and  PLL readout noise do not match. This is probably because the gain setup on PLL changes and the readout noise changes with the gain, see psl:816 .
  • The table is not floated. If we float the table,seismic noise at frequency above 3-4Hz along with acoustic noise around100- 1kHz should decrease, and we might have a good chance to measure thermal noise around 100 Hz.
  • In loop noise for RCAV is around 50nV/rtHz instead of 10 (we could go down to this level before, see psl:855). The gain on TTFSS loop cannot be increased to what it was before (Common/Fast=920/850)We are checking what are the problems.
  • A quick measurement for ACAV in loop noise gives flat noise floor of ~40nV/rtHz. The slope from ACAV's error signal is 52kHz/V. Thus, the frequency noise floor from ACAV is ~ 2mHz/rtHz. Although it is higher than the designed value  (smaller than shot noise level), it is certainly not the current limiting noise source. 

 

beat_2012_02_03.png

  862   Sat Mar 3 18:07:15 2012 ZachDailyProgressBEATbeat is back

And I stole one of your air dusters :-)

There used to be TWO label makers in the ATF---one of which looked like that one. What ever happened to those?!

Quote:

NO!! no dance party for you because you didn't return our label maker.

 

 

 

  861   Sat Mar 3 17:50:45 2012 taraDailyProgressBEATbeat is back

NO!! no dance party for you because you didn't return our label maker.

Quote:

Quote:

We expect it to be stable by tomorrow afternoon.

 What about my dance party?! 

disco.png

 

  860   Sat Mar 3 14:53:51 2012 ZachDailyProgressBEATbeat is back

Quote:

We expect it to be stable by tomorrow afternoon.

 What about my dance party?! 

disco.png

  859   Sat Mar 3 14:52:40 2012 FrankNotesTempCtrlscript crashed - restarted

the temp control script crashed with an epics exception some time ago - the reason why it took forever to heat the chamber. Restarted the perl script and also changed the hard stop for the heater from 1.7 to 2.0 in the code to get more umpf

  858   Fri Mar 2 23:37:36 2012 taraDailyProgressBEATbeat is back

 Beat signal is back now, but we have not measured the spectrum yet since the temperature is still drifting fast (10kHz/ min). We plan to measure it tomorrow afternoon.

    The optics for RCAV are aligned. The visibility is ~95% without adjusting the lenses for mode-matching. I do not redo the mode match yet because we probably have to move the chamber for installing the air springs very soon. For ACAV, the beam is aligned, with visibility only ~ 80%. I also adjusted the cable length for PDH locking so that the error signal looks symmetric (Frank added the 4-ch splitter for demodulating RFAM, so the phase shift changed a bit.  All reflected beams  are properly blocked with razor blade blocks.

    The beat frequency is ~ 188MHz instead of the nominal value of 160 MHz because the temperature is not settled at the set point yet. It takes longer than before because of the copper shields around the cavities.  We expect it to be stable by tomorrow afternoon.

  857   Fri Mar 2 01:02:50 2012 FrankNotesSeismicair spring isolation design

after talking several times to the machine shop guys and checking stock at central we came up with the following design for the two beams supporting the vacuum chamber including thermal insulation. The main requirement was to not raise the beam height too much to not make everything to unstable. We also can only raise it about 1in before reaching the upper limit of the existing periscope. The current design adds 3/8" of height from the beam plus additional space below (~1/4). The width is designed to accommodate a second thermal insulating box around the existing setup if required for more stability.

vacuum_chamber_dualcav_with_beam.jpg

  856   Thu Mar 1 12:04:00 2012 FrankDailyProgressNoiseBudgetRCAV loop characterization

the way you calculate the noise uses a few wrong assumptions:

  • the impedance is not 2k; this is the effective impedance seen at the DC-output including all gain, not at the RF output
  • the impedance is complex, but only the real part of the impedance contributes with thermal noise + all noise from amplifiers and their circuit inside the PD contribute to your PD noise
  • you forgot the stuff inside the TTFSS box which is part of your noise measurement at the output on the front of TTFSS box
  • the noise level you measured at the EP was with the loop closed, not only the electronic noise floor

Quote:

We locked RCAV, optimized beam alignment, and measured the noise from error point. The measured noise@error point is ~ 10nV/rtHz (equivalent to 0.4 mHz/rtHz ~a factor of 3 below shot noise limit).  This noise will be add into the noise budget.

      

    ==setup==

     The power is adjusted so that the DC output from the RFPD is 2V (about 1mW) when all the laser power is reflected on the RFPD. Common/Fast gain on TTFSS = 920/850. error signal slope is 35.5kHz/V.

   ==comparison with calculation== 

   If the loop gain is high, it can suppress laser free running noise until it reaches flat electronic noise. The electronic noise can be estimated if we know impedance(Z), gain(G) of the RFPD. The mixer(6dB conversion loss) + 50Ohm terminate reduce the level by a factor of 4 (-12dB).  Thus the noise level is ~ sqrt(4kTZ) * G/4.  For RCAV RFPD, Z = 2000,G = 10, so the noise level is ~ 15nV/rtHz. (The measured noise is lower than the approximation, I'll double check this before posting the noise budget).

 

     

 

 

  855   Thu Mar 1 10:18:13 2012 Tara, FrankDailyProgressNoiseBudgetRCAV loop characterization

We locked RCAV, optimized beam alignment, and measured the noise from error point. The measured noise@error point is ~ 10nV/rtHz (equivalent to 0.4 mHz/rtHz ~a factor of 3 below shot noise limit).  This noise will be add into the noise budget.

     1) in loop noise( I make some change as Frank suggests in PSL:856)

    ==setup==

     The power is adjusted so that the DC output from the RFPD is 2V (about 1mW) when all the laser power is reflected on the RFPD. Common/Fast gain on TTFSS = 920/850. error signal slope is 35.5kHz/V.

   ==comparison with calculation== 

   This calculation serves as a sanity check for the signal we measure. The measured in loop noise at error point will contain several noise sources (electronic noise floor, shot noise, free running noise of the laser which is not completely suppressed), but, with the current known information, I can estimated only the electronic noise. The result will at least give us some lower limit of the noise floor.

   If the loop gain is high, it can suppress laser free running noise until it reaches flat electronic noise. The electronic noise can be estimated if we know impedance(Z), gain(G) of the RFPD. The mixer(6dB conversion loss) + 50Ohm terminate reduce the level by a factor of 4 (-12dB).  Thus the noise level is ~ sqrt(4kTZ) * G/4.  For RCAV RFPD, Z = 2000,G = 10, so the noise level is ~ 15nV/rtHz. (The measured noise is lower than the approximation, I'll double check this before posting the noise budget).   

    Since the real part of the impedance of the whole circuit will take sometime to calculate, I'll just look at the measure noise output and use that number. Frank measured the voltage noise of the RFPD (in psl:795.) to be 20nV/rt(Hz). The noise is attenuated through the mixer (-6dB), so the expected thermal noise at the output is 20/4 = 5 nV/rtHz . This is not too bad, the noise floor of SR785 is ~7nV, so the incoherent sum is sqrt(5^2+7^2) = 8.6 nV.

   

 

  2) modulation index

    I measured the transmitted power from carrier and sideband which are 1.3mW and 6.45 uW. Power ratio from sideband/carrier is = 0.05. This corresponds to modulation depth(beta) = 0.14. [from (J1(beta)/J0(beta))^2 = 0.05 where J0 and J1 are Bessels fn of the first kind.

     We use 4ch splitter instead of 1 (see setup here) so the power to EOM is reduced to .2 dBm (measured, and it agrees with the setup with 6dBm attenuator). The EOM has 50 ohm load as specified in the manual, with .2dBm the input voltage is 0.23V. The datasheet says that the modulation efficiency is more than 0.2 rad/V. This gives beta = 0.046, ( from the measured value, the modulation efficiency can be calculated to be ~ 0.6 rad/V), so I think it is ok.

    The current mod index is ok, since the inloop noise is below shot noise, and we don't loose much power to the sideband.

     

 

  854   Wed Feb 29 23:45:34 2012 FrankNotesRFRF system schematic updated

the schematic below reflects the current RF system. The level 7 mixers for the RF-AM detection are currently operated at slightly reduced RF power (we don't have more and i don't have a good amplifier to fix this). The LO is only 3.5dBm instead of 7dBm at the moment. A quick test with only one channel (without the 2-way splitter) showed that there is almost no difference in output signal from 6.6dBm to 3.5dBm LO power. As the RF-AM part is a non-critical subsystem at the moment i think we can use it as-is for now.

RF-schematic_v0.png

  853   Wed Feb 29 21:43:19 2012 Tara, FrankDailyProgressRFAMPD's for both cavities installed

We've installed two pd's, one for each cavity. As the demodulated signal was very small when we optimize the EOM the usual way so we decided to replace one PD with a 14.75MHz resonant PD we've stolen from the TNI to get some more signal. The other one is a 10CS from Thorlabs (which i wanted to use for both of them). It actually turns out that we don't gain very much using the resonant PD, so we might switch back to the other one.

The DC level of the demodulated signal is less than 1mV, the noise level is limited by the readout electronics when optimized (nV level). We currently have one SR560 (DC-coupled) each with a gain of 1000 and sitting on the white noise floor. We will check with a better pre-amp tomorrow.

We temporarily sample the signal taken in front of RCAV with EPICS to monitor it over time. The EOM is not stabilized, only passive isolation is installed. When we misalign the EOM and start heating the EOM we can see the noise spectrum going up and down periodically, as well as the (tiny) dc signal, so everything is as expected, except that we can't measure anything when optimized.

schematic, pictures,diagram, etc  later when we have a final setup.

  852   Wed Feb 29 01:03:03 2012 FrankNotesDAQfixed errors in channel database - daqd restarted

fixed some errors in our channel database for the framebuilder and restarted the daqd process

  851   Tue Feb 28 22:50:11 2012 FrankNotesScheduleupdate

The following tasks will be finished within the next 24h:

  • aligning both cavities
  • align beat
  • install RF-AM PD (start with one to get estimate for noise budget)
  • measure TF from seismic to beat (if possible, vacuum chamber temp might not be stable enough)

we will then focus on (re-)measuring some things we need very soon for a more accurate NB (there are more but the following ones are my favorites for now. We are about a factor of 3 above the CTN @200Hz)

  • RF-AM - not in NB yet, but seems to dominate 100Hz-1kHz region, so we do this first (seen last week while optimizing everything)
  • RIN  - old measurement must be wrong as it is too high
  • seismic  - new stack, so new TF
  • phase noise of Marconi locked to rubidium clock - we know that the noise will be less below 1kHz, above it most likely does not change the phase noise level
  • electronic noise from both PDH loops

 

  850   Tue Feb 28 00:19:41 2012 FrankNotesDAQdaqd restarted

restarted the framebuilder process. Channel list needs to be updated, i didn't clean up the FSS channels so there are still old/unused channel names.

Current list of saved channels.

106 channels total
105 trend channels

chnum   slow    |name                           |rate   |trend  |group  |bps    |bytes  |offset |type   |active
0       0       C3:PSL-BOX_SENS1                                16      1       0       4       64      0       4       1
0       0       C3:PSL-BOX_SENS2                                16      1       0       4       64      64      4       1
0       0       C3:PSL-BOX_SENS3                                16      1       0       4       64      128     4       1
0       0       C3:PSL-BOX_SENS4                                16      1       0       4       64      192     4       1
0       0       C3:PSL-BOX_TEMPAVG                              16      1       0       4       64      256     4       1
0       0       C3:PSL-BOX_HEATER                               16      1       0       4       64      320     4       1
0       0       C3:PSL-BOX_SETPT                                16      1       0       4       64      384     4       1
0       0       C3:PSL-BOX_KP                                   16      1       0       4       64      448     4       1
0       0       C3:PSL-BOX_KI                                   16      1       0       4       64      512     4       1
0       0       C3:PSL-BOX_KD                                   16      1       0       4       64      576     4       1
0       0       C3:PSL-BOX_ENABLE                               16      1       0       4       64      640     4       1
0       0       C3:PSL-BOX_TIMEOUT                              16      1       0       4       64      704     4       1
0       0       C3:PSL-BOX_SCALE                                16      1       0       4       64      768     4       1
0       0       C3:PSL-ACAV_RFPDDC                              16      1       0       4       64      832     4       1
0       0       C3:PSL-ACAV_RCTRANSPD                           16      1       0       4       64      896     4       1
0       0       C3:PSL-ACAV_LOCKEDLEVEL                         16      1       0       4       64      960     4       1
0       0       C3:PSL-RCAV_SENS1                               16      1       0       4       64      1024    4       1
0       0       C3:PSL-RCAV_SENS2                               16      1       0       4       64      1088    4       1
0       0       C3:PSL-RCAV_SENS3                               16      1       0       4       64      1152    4       1
0       0       C3:PSL-RCAV_SENS4                               16      1       0       4       64      1216    4       1
0       0       C3:PSL-RCAV_TEMP                                16      1       0       4       64      1280    4       1
0       0       C3:PSL-RCAV_TEMPAVG                             16      1       0       4       64      1344    4       1
0       0       C3:PSL-RCAV_S1CAL                               16      1       0       4       64      1408    4       1
0       0       C3:PSL-RCAV_S2CAL                               16      1       0       4       64      1472    4       1
0       0       C3:PSL-RCAV_S3CAL                               16      1       0       4       64      1536    4       1
0       0       C3:PSL-RCAV_S4CAL                               16      1       0       4       64      1600    4       1
0       0       C3:PSL-RCAV_SETPT                               16      1       0       4       64      1664    4       1
0       0       C3:PSL-RCAV_KP                                  16      1       0       4       64      1728    4       1
0       0       C3:PSL-RCAV_KI                                  16      1       0       4       64      1792    4       1
0       0       C3:PSL-RCAV_KD                                  16      1       0       4       64      1856    4       1
0       0       C3:PSL-RCAV_ENABLE                              16      1       0       4       64      1920    4       1
0       0       C3:PSL-RCAV_TIMEOUT                             16      1       0       4       64      1984    4       1
0       0       C3:PSL-RCAV_SCALE                               16      1       0       4       64      2048    4       1
0       0       C3:PSL-RCAV_RFPDDC                              16      1       0       4       64      2112    4       1
0       0       C3:PSL-RCAV_RCTRANSPD                           16      1       0       4       64      2176    4       1
0       0       C3:PSL-RCAV_LOCKEDLEVEL                         16      1       0       4       64      2240    4       1
0       0       C3:PSL-FSS_HEATER                               16      1       0       4       64      2304    4       1
0       0       C3:PSL-FSS_FREQCOUNT                            16      1       0       4       64      2368    4       1
0       0       C3:PSL-FSS_VCOMON                               16      1       0       4       64      2432    4       1
0       0       C3:PSL-FSS_VCOMON_CAL                           16      1       0       4       64      2496    4       1
0       0       C3:PSL-FSS_VCOFREQ                              16      1       0       4       64      2560    4       1
0       0       C3:PSL-FSS_RFPDDC                               16      1       0       4       64      2624    4       1
0       0       C3:PSL-FSS_LODET                                16      1       0       4       64      2688    4       1
0       0       C3:PSL-FSS_PCDET                                16      1       0       4       64      2752    4       1
0       0       C3:PSL-FSS_FAST                                 16      1       0       4       64      2816    4       1
0       0       C3:PSL-FSS_PCDRIVE                              16      1       0       4       64      2880    4       1
0       0       C3:PSL-FSS_RCTLL                                16      1       0       4       64      2944    4       1
0       0       C3:PSL-FSS_VCODET                               16      1       0       4       64      3008    4       1
0       0       C3:PSL-FSS_TIDALOUT                             16      1       0       4       64      3072    4       1
0       0       C3:PSL-FSS_MODET                                16      1       0       4       64      3136    4       1
0       0       C3:PSL-FSS_VCODETPWR                            16      1       0       4       64      3200    4       1
0       0       C3:PSL-FSS_MIXERM                               16      1       0       4       64      3264    4       1
0       0       C3:PSL-FSS_SLOWM                                16      1       0       4       64      3328    4       1
0       0       C3:PSL-FSS_VCOM                                 16      1       0       4       64      3392    4       1
0       0       C3:PSL-FSS_TIDALINPUT                           16      1       0       4       64      3456    4       1
0       0       C3:PSL-FSS_SW1                                  16      1       0       4       64      3520    4       1
0       0       C3:PSL-FSS_SW2                                  16      1       0       4       64      3584    4       1
0       0       C3:PSL-FSS_PHFLIP                               16      1       0       4       64      3648    4       1
0       0       C3:PSL-FSS_VCOTESTSW                            16      1       0       4       64      3712    4       1
0       0       C3:PSL-FSS_VCOWIDESW                            16      1       0       4       64      3776    4       1
0       0       C3:PSL-FSS_INOFFSET                             16      1       0       4       64      3840    4       1
0       0       C3:PSL-FSS_MGAIN                                16      1       0       4       64      3904    4       1
0       0       C3:PSL-FSS_FASTGAIN                             16      1       0       4       64      3968    4       1
0       0       C3:PSL-FSS_PHCON                                16      1       0       4       64      4032    4       1
0       0       C3:PSL-FSS_RFADJ                                16      1       0       4       64      4096    4       1
0       0       C3:PSL-FSS_SLOWDC                               16      1       0       4       64      4160    4       1
0       0       C3:PSL-FSS_VCOPWR                               16      1       0       4       64      4224    4       1
0       0       C3:PSL-FSS_VCOMODLEVEL                          16      1       0       4       64      4288    4       1
0       0       C3:PSL-FSS_TIDALSET                             16      1       0       4       64      4352    4       1
0       0       C3:PSL-FSS_LOCK                                 16      1       0       4       64      4416    4       1
0       0       C3:PSL-FSS_SLOWLOOP                             16      1       0       4       64      4480    4       1
0       0       C3:PSL-PMC_PMCTLL                               16      1       0       4       64      4544    4       1
0       0       C3:PSL-PMC_RFPDDC                               16      1       0       4       64      4608    4       1
0       0       C3:PSL-PMC_LODET                                16      1       0       4       64      4672    4       1
0       0       C3:PSL-PMC_PMCTRANSPD                           16      1       0       4       64      4736    4       1
0       0       C3:PSL-PMC_PCDRIVE                              16      1       0       4       64      4800    4       1
0       0       C3:PSL-PMC_PZT                                  16      1       0       4       64      4864    4       1
0       0       C3:PSL-PMC_MODET                                16      1       0       4       64      4928    4       1
0       0       C3:PSL-PMC_PMCERR                               16      1       0       4       64      4992    4       1
0       0       C3:PSL-PMC_SW1                                  16      1       0       4       64      5056    4       1
0       0       C3:PSL-PMC_SW2                                  16      1       0       4       64      5120    4       1
0       0       C3:PSL-PMC_PHFLIP                               16      1       0       4       64      5184    4       1
0       0       C3:PSL-PMC_BLANK                                16      1       0       4       64      5248    4       1
0       0       C3:PSL-PMC_GAIN                                 16      1       0       4       64      5312    4       1
0       0       C3:PSL-PMC_INOFFSET                             16      1       0       4       64      5376    4       1
0       0       C3:PSL-PMC_PHCON                                16      1       0       4       64      5440    4       1
0       0       C3:PSL-PMC_RFADJ                                16      1       0       4       64      5504    4       1
0       0       C3:PSL-PMC_RAMP                                 16      1       0       4       64      5568    4       1
0       0       C3:PSL-PMC_LOCK                                 16      1       0       4       64      5632    4       1
0       0       C3:PSL-PEM_RMTEMP                               16      1       0       4       64      5696    4       1
0       0       C3:PSL-PEM_BOXTEMP                              16      1       0       4       64      5760    4       1
0       0       C3:PSL-FSS_RFAM_RCAV                            16      1       0       4       64      5824    4       1
0       0       C3:PSL-FSS_RFAM_ACAV                            16      1       0       4       64      5888    4       1
0       0       C3:PSL-FSS_EOM_TSET                             16      1       0       4       64      5952    4       1
0       0       C3:PSL-FSS_EOM_TACT                             16      1       0       4       64      6016    4       1
0       0       C3:PSL-FSS_EOM_IMON                             16      1       0       4       64      6080    4       1
0       0       C3:PSL-FSS_EOM_SETTEMP                          16      1       0       4       64      6144    4       1
0       0       C3:PSL-GEN_DAQ1                                 16      1       0       4       64      6208    4       1
0       0       C3:PSL-GEN_DAQ2                                 16      1       0       4       64      6272    4       1
0       0       C3:PSL-GEN_DAQ3                                 16      1       0       4       64      6336    4       1
0       0       C3:PSL-GEN_DAQ4                                 16      1       0       4       64      6400    4       1
0       0       C3:PSL-GEN_DAQ5                                 16      1       0       4       64      6464    4       1
0       0       C3:PSL-GEN_DAQ6                                 16      1       0       4       64      6528    4       1
0       0       C3:PSL-GEN_DAQ7                                 16      1       0       4       64      6592    4       1
0       0       C3:PSL-GEN_DAQ8                                 16      1       0       4       64      6656    4       1
10001   0       C3:FB1-FB_DUMMY                                 16384   0       0       4       65536   6720    4       0

  849   Mon Feb 27 23:57:25 2012 FrankNotesDAQchannel list for framebuilder updated

changed the channel list for epics channels which will be recorded. Modified file according to changes mentioned in previous post.
File contains also channels which physically don't exist at the moment, e.g. RF-AM channels and 2nd. temp stabilized box around chamber.

File: /cvs/cds/caltech/chans/daq/C3PSL_EPICS.ini

  848   Mon Feb 27 21:45:55 2012 FrankNotesRefCavnew beam height for cavities

The beam height changed by 1/8". Current beam height is 5-7/8". The top stack plate is slightly off-centered towards the ACAV side but we don't want to re-open the chamber now to fix this (it's not much, but one can see it). We can do this the next time we open it anyway or if we see weird coupling between different stack modes. When we open it next time we will add some markings at the end faces of the top stack plate to better see if it's centered or not.

We will align the beams along the hole pattern as designed first and move the cavities to the right position. Once we add the air springs (hopefully soon) we have to change the beam height again.

  847   Mon Feb 27 20:52:42 2012 FrankDailyProgressDAQDAQ reconfigured, channels added/removed

cleaned up the VME stuff a little bit - removed all old channels we don't use anymore, mostly FSS stuff, added new channels for the in-vac sensors and renamed a bunch of others to match the current situation. The only thing i didn't touch is all channels from "RCAV", which is now common to both cavities but i didn't want to dig too deep and change all scripts etc. So that's left for the near future.

Current channels connected to 16bit ADC (VMIC 3123):

  • CH0 : C3:PSL-PEM_BOXTEMP (Sensor in 2nd thermal box around vacuum chamber)
  • CH1 : C3:PSL-PEM_RMTEMP (Sensor measuring temp in clean enclosure)
  • CH2 : C3:PSL-RCAV_TEMP (4 Sensors on vacuum chamber minus voltage reference)
  • CH3 : C3:PSL-RCAV_SHIELD1 (In-Vac sensor on RCAV copper shield)
  • CH4 : C3:PSL-ACAV_SHIELD1 (In-Vac sensor on ACAV copper shield)
  • CH5 : C3:PSL-ACAV_SHIELD2 (In-Vac sensor on ACAV copper shield)
  • CH6 : C3:PSL-FSS_FREQCOUNT (Frequency counter output measuring beat)
  • CH7 : C3:PSL-FSS_VCOMON (VCO feedback monitor signal)
  • CH8 : C3:PSL-BOX_SENS1 (Sensor #1 of 2nd thermal box around vacuum chamber)
  • CH9 : C3:PSL-BOX_SENS2 (Sensor #2 of 2nd thermal box around vacuum chamber)
  • CH10 : C3:PSL-BOX_SENS3 (Sensor #3 of 2nd thermal box around vacuum chamber)
  • CH11 : C3:PSL-BOX_SENS4 (Sensor #4 of 2nd thermal box around vacuum chamber)
  • CH12 : C3:PSL-RCAV_SENS1 (Sensor #1 on vacuum chamber)
  • CH13 : C3:PSL-RCAV_SENS2 (Sensor #2 on vacuum chamber)
  • CH14 : C3:PSL-RCAV_SENS3 (Sensor #3 on vacuum chamber)
  • CH15 : C3:PSL-RCAV_SENS4 (Sensor #4 on vacuum chamber)

new/modified channels are available in real-time, but not saved at the moment!

old database-files are located in "20120227"-subfolder on the SUN in the psl folder and on the svn in /software

  846   Mon Feb 27 19:17:57 2012 taraNotesElectronics EquipmentNoise in delay line read out technique

I estimate the noise sensitivity in the delay line technique. The calculation should tell us how good this method can be used to measure frequency noise in beat signal.

 The calculation follows the setup in PSL:828. All relevant datasheets can be found here.

      ==Methods==

      I try to compute how each component in the setup generates noise and shows up at the end of the stream. With the calibration factor, I can convert the noise back to its equivalent frequency noise at the input.

      The only component that introduces noise in the setup is the ZHL-1A mixer. Its noise figure is ~ 8dB. Assuming that the only noise from the input side is thermal noise in 50 ohm, then the noise level after the amplifier is ~ 50nV, see details in the note below.

     For each signal trace, the signal goes through 4-ch splitter (-6dB), cable, then the mixer (~3dB conversion loss). These components give a factor of (1/4) x (1/2) to the signal. I'm not sure how noise in both delay line will sum up at the mixer. So for now I just assume noise coupling from one side of the mixer. The noise level after the low pass should be 50nV/8 = 6.3 nV. The calibration from Voltage to frequency noise at the mixer output is 2.5 MHz/V (from measurement on 2012_02_22,160MHz, svn)  Thus the absolute frequency noise  is  6.2nV x 2.5 MHz/V = 15.5 mHz.  This level is ~ a factor of 1.5 lower than the frequency noise of Marconi (10kHz tuning range) , see psl:834, psl:833. If this is the only limiting source, we should be able to measure coating noise upto ~500Hz (limited bandwidth due to the chosen delay time is not taken into account yet).

     It would be nice to be able to measure the noise and compare it with the calculation. To measure the noise, we need a low noise input source (Marconi with 1kHz tuning range should be ok) with power as specified in the setup. However, the measured noise level is higher than the expected noise and we don't know what the cause is. So we can not verify the calculated noise level yet.

==note==

Noise figure = 10log10( Noise Factor). From the datasheet, noise figure of the amplifier ~ 8dB which corresponds to Noise Factor = 6.3.

Noise Factor = SNR from input / SNR from output. For our setup, the signal from input is 5dBm (0.4 V), with noise ~ 1nV flat (50 Ohm thermal noise). the output signal is 22.8dBm (3.1V). Thus the expected noise at the output = Noise Factor x (signal_out/signal_in) x noise_in = 6.3x (3.1/0.4) x 1nV ~ 50nV.

ELOG V3.1.3-