40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  PSL, Page 40 of 52  Not logged in ELOG logo
ID Date Author Type Category Subject
  645   Thu Jul 28 16:42:26 2011 RaphaelDailyProgressFSSfitted LISO models for EOM and PZT

I fitted the measured transfer functions for the EOM and PZT using LISO. Here are the results:

pzt_measured_vs_fitted.jpg

And here is the LISO file that fitted the data to the measured data: measuredPZT_TF_fit.fil

Likewise, for the PZT actuator

pzt_act_measured_vs_fitted.jpg

measuredPZT_Actuator_TF_fit.fil

Likewise for the EOM

eom_measured_vs_fitted.jpg

measuredEOM_TF_fit.fil

I would really appreciate help in improving these models. I particularly need help improving the EOM fit, as LISO didn't even perform the fit. I just eyeballed it since LISO would do something crazy whenever I told it to fit the parameters. I think the chosen quality factors need to be improved before I can get a decent fit.

  644   Thu Jul 28 01:54:21 2011 frank, taraDailyProgressopticcleaning opto mechanical parts

  Today we started working on the layout. There is one mistake in the layout, the mirror behind AOM for REFCAV is too close to the insulation box, so we have to fix the layout.

 

 The oil on the table actually comes from holes on the table. About 5-6 screw holes had lot of oil, so I flushed them with methanol a few times.

IMG_1869.JPG

The HEPA filter is plugged in and turned on. I unplugged one of the monitors and used the outlet for the filter.

 

 The window is measured to be ~ 6 inches in diameter. Thus, the assumption in the design that the centers between two cavities are 3 inches is ok. If necessary, it can go up to 4 or 5 inches.

IMG_1875.JPG

IMG_1874.JPG

The layout is updated. The spot size in both AOMs are adjusted to 220 um.

2011_07_28.png

  643   Thu Jul 28 01:24:40 2011 ranaDailyProgressElectronics EquipmentPhotodiode Progress

This is sort of OK, but check out the 40m elog for the right way to characterize RFPDs. You really need more resolution to resolve the notch and the y-axis should be in units of Ohms.

Also, the test input is a fine rough guide, but its not accurate to get the high frequency characteristics right or the 2f notch with any accuracy.

  642   Wed Jul 27 00:23:12 2011 RaphaelDailyProgressElectronics EquipmentPhotodiode Progress

I took the transfer function of the photodiode (D980454-00.pdf) from the test input to rf out and found that the resonance and notch where at the same frequency and that the biggest difference is in the gain. Thus, I conclude that it is safe to use the test input to assist someone in readjusting the notch and resonance.

Test_Input_vs_Optical.jpg

We also want to adjust the dc output gain so that a photocurrent of 5mA will produce 10V at DC output. Before the buffer amplifier, the 5mA produces a voltage of 0.1V. So we would like the last amplifier to amplify this voltage by 100. Since it is a non-inverting amplifier, we would like that R16/R11=99. However, I can't think of a combination of resistors in our possession that will produce this. I therefore chose R11=10ohms and R16=1kohms.The board has been modified accordingly.

  641   Tue Jul 26 19:44:31 2011 taraNotesFSSTF of EOM

 I measured the slope of the error signal which will be used for EOM TF, the number is 0.467 [ MHz/V].

      The procedure for measuring the error signal slope is written here. The result changes from 0.275 [MHz/V] to 0.467 [MHz/V] might come from the fact that we modified the RF summing box.

 

slopefit.png

figure 1: error signal as seen on an oscilloscope, the fit gives 2106 [V/s]. With 35.5 MHz sideband as a marker, 36.1 ms apart (not shown here), the calibration for time to Hz is [35.5MHz/36.1ms] ~ 1e9 [Hz/s]. The slope is then 2106 [V/s] / 1e9 [Hz/s] = 2.1 e-6 [V/Hz] ~ 0.5 MHz/V

 

With this calibration factor, we can complete the EOM TF from this entry and have the unit of Hz/V in the transfer function.

 

   

 

  640   Tue Jul 26 18:42:16 2011 raphael, taraDailyProgressopticcleaning opto mechanical parts

Today we removed the optics behind the PMC, ACAV, and cleaned the table.

  •   Optic mounts and posts are cleaned by pressurized air, and kept in a plastic box.
  •   Lens and mirrors are kept in optic cases.
  •   ACAV is moved to the end of the table, ion pump is unplugged.
  • Table is cleaned with methanol, but some grease( under acav) is still on it.

NOTE: I just realized that the HEPA filter above the table close to the entrance ( the one that has the laser) is unplugged.

I could not find any available outlet to plug it back yet. We should turn it on soon.

IMG_1867.JPG

 

 

IMG_1866.JPG

  639   Tue Jul 26 03:18:42 2011 JenneHowToLaserSpelling

Quote:

Jenny Laser      

Jenne Laser  (thanks!)

  638   Mon Jul 25 23:59:56 2011 RaphaelDailyProgressElectronics EquipmentPhotodiode Progress

Hey, so I readjusted the resonant frequency of the photodiode so that resonance would be at ~35.5MHz frequency. Here is the photodiode circuit: D980454-00.pdf. I chose L2 to be 100uH. Here is the optical transfer function:Photodiode_Optical_TF.jpg

Here is the Jenny Laser setup to measure the optical transfer function:

 jennyLaserSetup.png

 

The power coming into the photodiode is 0.41mW. I recalculated the theoretical voltage that we are supposed to get from the quantum efficiency to be 12.1mV. I measured 7.09mV. I calculated the quantum efficiency from the equation QE= responsivity *hc/(lambda*e). From there, I multiplied power by QE to get the photocurrent. The voltage coming into the first detector would just be V=I*R, where R is 20 ohms. The next amplifier multiplies the voltage by about 1.9.

One problem that I know of is that the spot size of the laser hitting the  photodiode isn't the smallest it could be. The spot size hitting the diode could be smaller if I moved the photodiode back. However, I ran into problems trying to do this. After I positioned the photodiode to the point where the spot size was smallest, the beam ended up clipping off the lens whenever I tried to readjust the reflector. It took me forever to adjust the Jenny laser so that the beam was collimated and the laser hit decent parts of the 50/50 beam splitter and reflector. So I think in the near future, I will try to raise the lens by some amount so it doesn't clip.

  637   Mon Jul 25 21:02:25 2011 frank, raphael, taraNotesFSSTF of EOM

 We measured EOM's transfer function again. The result is unexpected, it is flat from ~ 100k to 1MHz, instead of going up with frequency.

      ==setup==

     The setup is shown in this entry , but this time we used an RF summing box modified by Raphael. It does not have 50 ohms or resonant at 35.5MHz, but it's not worse than what we used before.

     The TF is measured from three points,see fig 1 below.

  1. out1/ref
  2.  out2 / ref
  3.  Mixer out/ ref:

 schematic.pdf

figure 1: schematic for TF measurement. The excitation is split by a power spiltter, one is sent to EOM via RF summing box, another one is used for reference. The response is chosen from three points, OUT1, OUT2 and MIXER out. Since we have no direct access to the mixer out in the circuit, we split the signal and use an external mixer for mixer out measurement. This external mixer allows us to measure the response without seeing effects (extra poles, bandwidth) from other electronic components in the circuit.

 

    ==RESULTS==

 eomtf.png

  We tried measuring the TF from three different points because we want to check which part of the circuit causes extra phase lag in the TF.  The first time we took the measurements,  the TFs from each points had different phase lag ( up to 52 degree), but when we remeasured again, the phases from three points actually did not vary that much, except the 180 degree phase flip. We have not yet determined what cause this problem, so we have to keep this in mind just in case.

 NOTE:

  • The data has not yet been converted to [frequency/Volts] yet, we need to measure the slope of the error signal.
  • The  reason that the EOM TF is proportional to frequency up to ~10kHz and starts to go flat up to ~ 1MHz is not yet understood.

The data will be fit, and used in Simulink model.

After the TF is measured, we will continue to work on the new layout and start removing all the optics and clean the table.

 The data and code for plotting are attached below.

Attachment 3: code_2011_07_25.m.zip
Attachment 4: data_2011_07_25.mat.zip
  636   Fri Jul 22 20:51:13 2011 frank, taraDailyProgressopticcleaning opto mechanical parts

As we removed some optics on the table, we use pressurized air to blow away dust/dirt on the mechanical parts (mount/ post/ lens holder) Optics have not been cleaned yet. We will clean it before we put everything back on the table. The cleaned parts are kept in a plastic box.

 

IMG_1862.JPGIMG_1863.JPG

  635   Fri Jul 22 16:31:12 2011 taraDailyProgressAOMAOM adaptor plate

 

 As Frank suggested, I edited the drawing, so that the adaptor plate can accommodate both types of AOM.

 

 

Attachment 1: aom_base.PDF
aom_base.PDF
Attachment 2: eom_block1.PDF
eom_block1.PDF
Attachment 3: eom_block2.PDF
eom_block2.PDF
  634   Fri Jul 22 00:17:03 2011 RaphaelDailyProgressFSSPhotodiode progress

Today I spent the whole day wondering why I obtained a bad optical transfer function from the photodiode. It turned out that the power supply to the reference diode in the Jenny laser setup was unplugged. Seriously, a certain person told me that the power supply to the reference was already setup, and another certain person told me that the power supply looked like it was on when I asked for help. I was also getting a readout from the DC output. I still don't understand why I got a DC readout if the power supply was turned off!

Frank also replaced the photodiode in the circuit (perhaps unnecessarily).

The transfer function seems to look good. The notch frequency is at 71.5MHz, which is where we want it to be. I need to modify the board so that the resonance is at 35.5MHz. Currently it is around 11MHz. 

I also need to double check my calculations for the quantum efficiency with Frank. We both suspect that the dc voltage out is too low given the responsivity and the measured input power of 0.39mW

Tomorrow, I plan on finishing modifying the photodiode board to be resonant at 35.5MHz and to optimize the dc output for a low powered laser. I also plan on modifying the summing box.

-Raphael

  633   Thu Jul 21 18:03:20 2011 taraDailyProgressAOMAOM adaptor plate,

I made drawings for aom block and adaptor plate. The assembly is for 3" beam height.

 

The assembly is consisted of two parts. The bottom aluminum part is for mounting new focus 9071 4 axis stage on the table.

The top plastic part, is for mounting an AOM to the 4-axis stage.

 

The bottom part actually is designed for a standard EOM, so the height is  3". With a plastic adaptor plate, it can be used for an AOM as well, so I'll order a few of the alumnium parts.

 

There are 2 designs for AOM adaptor parts, because we have two AOM models. They have different screw size and mounting positions, I 'll order a couple for each design. 

 

Attachment 1: aom_base1.PDF
aom_base1.PDF
Attachment 2: aom_base2.PDF
aom_base2.PDF
Attachment 3: eom_block1.PDF
eom_block1.PDF
Attachment 4: eom_block2.PDF
eom_block2.PDF
  632   Wed Jul 20 23:00:30 2011 FrankDailyProgressopticoptic layout for new fss setup

the lens and mirror are in the ATF, a second VCO is in the left cabinet.

 

 

  631   Wed Jul 20 15:54:21 2011 taraDailyProgressopticoptic layout for new fss setup

  I edited the layout so that the spots in both AOMs are 200 um. I'll list what optics we might have to buy.

 

Most of the optics are already used on the table. I need to find:

  •  a lens with f = 343.6 m (plcx R =154.5mm)
  •  one more curve mirror with R = 0.3m for the second AOM.
  • aom adaptor plate (need to submit this to the work shop to have it done
  • periscope sets for both ACAV and RCAV (we need 4 in total, but we have only 2 sets)
  • second VCO

The optics on ACAV path have been removed, I left the optics on RCAV path for now because Raphael might want to remeasure EOM TF.

Once the measurement is done, all optics will be removed. We will clean the table, clean the optics before put them back on the table.

 

 

 

2011_07_20_layout.png

  630   Mon Jul 18 19:58:05 2011 RaphaelDailyProgressFSSPhotodiode modification progress

Today I learned how to solder an inductor onto a board.

We measured the transfer function of the photodiode board from the test input to the RF output. I changed parts on the board to have resonance at 35.5 MHz and a notch at 71 MHz. We changed L2 to 330nH and and L8 to 330nH. Frank also changed parts on the board, but I'm not exactly sure what he did.

Here is the schematic of the board along with the recommended values for a 35.5MHz resonant frequency.

D980454-00.pdf

Even when we changed L2 to 330nH instead of the recommended 81nH, resonance is at around 40MHz. However, the response still looks good and we do indeed see a notch and a resonance. Frank suspects that there is something wrong with the capacitance of the photodiode (might be broken)

  629   Mon Jul 18 11:41:22 2011 ranaDailyProgressopticoptic layout for new fss setup

There's no need to use such a large spot size on either the AOM or the EOM.

When using high power this could be an issue, but you can use a beam radius of more like 100-200 microns for the AOM to get fast response time.

  628   Mon Jul 18 02:20:01 2011 taraDailyProgressopticoptic layout for new fss setup

  The new mode matching for optics in front of the cavities is done.  The rest (for beat measurement) will be finished soon.

A few changes in this layout are:

1) spotsize for AOM is 500 um, as specified by the datasheet.

2) Mirrors behind the AOMs will be changed to R= 2.0 m instead of 0.3 m.

3) Spot size in the 35.5 MHz EOM is ~300um which is good for the model.

4) More mirrors (for steering the beam) for the AOMs are added.

 

I'm a bit worried about using f=57.4mm lenses because they are quite sensitive when we have to move the lenses around, but the space is very limited this time.

I'll let Raphael double check my calculation so he can learn how to do mode matching.

2aom_fss_2011-07-15.png

 

 

  627   Fri Jul 15 18:37:51 2011 RaphaelDailyProgressFSSPhotodiode modification progress

Currently, I am working on modifying a photodiode so that will be used to lock the laser to the reference cavity. We want to modify it so that it resonates at 35.5MHz and so that it is optimized for a low powered laser (~100microWatts). Here is the schematic for the photodiode circuit.

D980454-00.pdf

Using LISO, I modeled the RF part of the circuit using the values they suggested in the table to make it resonant at 35.5MHz. Here is the file that I used to plot the RF part

pd.fil

And here is the obtained transfer function.

PDRF.jpg

Frank is unsatisfied with the resonance at 35.5 MHz.

We then try to optimize the circuit for a smaller laser power. We want to adjust the resistors connected to U1 to change the gain at that stage. I plotted on LISO the transfer function of that single op amp for various amplifications to see if 35.5MHz would fall in the nonlinear region of the op amp if we increased the gain.

PDOpAmpGainStability.jpg

I found that we can't really increase the gain much without it becoming nonlinear. Frank said it wasn't worth trying.

I plan to adjust the DC path for low power on Monday.

So the plan is to test the functionality of the PD at it's current state, which is to resonate at 54MHz. Once we determine that the PD is working properly, we will adjust it to resonate at 35.5MHz, and see how else we can optimize it from there.

Today, we just did a smoke test of the PD and everything seems ok so far.

P.S. I soldered for the first time today. I made my first cable today, and it only took my an entire afternoon! :p

 

  626   Thu Jul 14 16:14:09 2011 RaphaelDailyProgressFSSSumming Box Calculations

Tara, Frank, and I were unsure of the values that we read off the summing box board in the PSL lab.

IMG_1771.JPG

We took the transfer of the current box from the servo input to the PC output and then used LISO to modify the values that we read off the board. I plotted the physical transfer function versus the LISO plot using the parameters I think we should have.

ServoInputToPCOutputSummingBoxTF.jpg

We see a notch at 21.5MHz, which is problematic because we are using a 35.5MHz local oscillator RF input. Thus, the RF input will not be attenuated by the filter in front of the Servo input, so the RF signal will go into the servo. This is not what we want, so we need to redesign the box so that the notch is at 35.5MHz

So I believe the current summing box looks something like this:

IMG_0089.JPG

The LISO file is attached: sumbox1.fil

I then plotted the transfer function from RF in to PC out and from RF in to HV in using LISO:

CurrentSummingBoxRFtoPCTF.jpg

We see resonance at around 21.5MHz

CurrentSummingBoxRFtoHVTF.jpg

We see attenuation at around 21.5MHz. I'm not exactly sure why the box is designed to have resonance at a slightly high frequency.

I also plotted the impedance as seen from the RF input, which also shows resonance to be at 21.5 MHz.

CurrentSummingBoxRFImpedance.jpg

Perhaps measuring these transfer functions would let me be more sure of the values I picked for the LISO model.

We must redesign the box so that the filter in front of the servo input has a notch at 35.5 MHz and the path from RF in to PC out resonates at 35.5MHz. I propose the following schematic:

IMG_0090.JPG

The LISO model for this schematic is attached: sumbox2.fil

I plotted the transfer function going from the servo input to PC out. The notch frequency is at 35.8MHz.

NewSummingBoxHVtoPCTF.jpg

I also plotted the transfer function from RF in to PC out. We have resonance at about 35.8MHz.

NewSummingBoxRFtoPCTF.jpg

 We also plotted the transfer function going from RF in to servo in.

NewSummingBoxRFtoHVTF.jpg

And then I plotted the impedance as seen from RF in. Unfortunately, it is not impedance matched to 50 ohms at 35.5MHz. I couldn't figure out a way to do it and Frank suggested that I just move on and that anything that I do would be better than what we currently have.

NewSummingBoxRFImpedance.jpg

  625   Wed Jul 13 23:12:22 2011 taraNotesNoiseBudgetThermal noise vs spotsize

I did some calculation to check how small the spot size on the cavity's mirrors can be. I assumed we use the current mirrors with  Radius of curvature=0.5m .

For 5 cm cavity length, the spotsize on the mirrors are 200 um. This only increases the coating noise by a factor of 1.5, not very significant.

 

The current cavity is not optimized for thermal noise measurement. We want to make the thermal noise higher, so that it is easier to reach. To make coating noise higher, spot size on the mirror should be smaller.

To decrease spot size, we can change the cavity length or the radius of curvature of the mirror. Since it is easier to change the cavity length, I kept the mirrors' radius of curvature constant (0.5m).  I assumed that the shortest cavity length that is reasonable is 5 cm. This gives spotsize = 200 um.

The current spot size with 20cm cavity is 291 um. So the coating noise is 291/200 ~ 1.5 larger.

spotsize.png

Note that the smallest spotsize we can use to measure the thermal noise correctly is given in Liu and Thorne. I calculated it to be 35 um x sqrt(1Hz / f). So the spotsize with 5 cm cavity length(200um) is still usable.

 

Attachment 2: spotsize.m.zip
  624   Mon Jul 11 15:52:36 2011 RaphaelDailyProgressFSSneed to modify RF summing box

Last Friday, we attempted to measure the transfer function of the EOM and the EOM actuator gain. However, we were unsuccessful due to problems with impedence matching.

What we did notice was that the RF summing box, which takes the TTFFS: EOM HV signal and the 35.5MHz RF signal as inputs and outputs to the EOM, that we are using for the current setup is different from the RF summing box that was designed to work with our setup.

Here is the current RF summing box:

119.JPG

Here is the RF summing box schematic that we are supposed to be using for the current setup:

D040469-B.pdf

If we modeled the signal going from the HV Actuator input to PC out on the RF summing box designed for this experiment to be:

IMG_0086.JPG

we expect the impedance to be:

RF_Summing_Box_Notch_Impedence.jpg

where the notch is designed to prevent signals with frequencies near 35.5MHz from going into the HV actuator input and interfering with the TTFSS.

We measured the transfer function of the summing box from the HV Actuator input to PC out:

Measured_RF_Summing.jpg

As you can see, the notch is at 21.5MHz, whereas the RF signal is at 35.5MHz. This is problematic because the RF signal can go into the HV actuator input and interfere with the TTFSS. Such a phenomenon may explain why the whole FSS hasn't performed as expected. We now need to modify the RF summing box so that the notch will be at 34.4MHz. We need to also be careful with impedance matching.

Attachment 3: Measured_RF_Summing.jpg
Measured_RF_Summing.jpg
  623   Tue Jul 5 14:30:35 2011 Raphael CervantesDailyProgressFSSneed help improving LISO models

Hey, I tried to improve the models by Koji's suggestions, but I still can't quite get them right. Can someone give me some suggestions? The files are attached.

Attachment 1: eomfit.fil
zero 21n #zero0
zero 84.5m 1M #zero1
pole 1k #pole0
pole 124.8 #pole1
pole 141.3 #pole2
pole 319.8 #pole3
pole 1.9k #pole4
zero 28k #zero2
zero 218k #zero3
zero 218k #zero4
... 37 more lines ...
Attachment 2: D040105-C.pdf
D040105-C.pdf D040105-C.pdf D040105-C.pdf
Attachment 3: eom.fil
op u4 ad797 gnd n3 n4
c c23 1u n1 n2
r r22 1.1k n2 n3
r r19 24.9k n4 n3
c c15 3.3n n4 n3
c c24 1u n4 n5
r r24 1.3k n5 n6
op u5 ad797 gnd n6 n7
r r20 47k n7 n6
c c18 33p n7 n6
... 19 more lines ...
Attachment 4: d0901893-a.pdf
d0901893-a.pdf d0901893-a.pdf d0901893-a.pdf
Attachment 5: eomHV.fil
#bottom half of schematic
c c21 1n n1 n2
c c22 220n n1 n2
r r42 3.3k n2 n3
op u6 pa85 gnd n3 n4
r r47 100k n4 n3
c c25 47p n4 n5
r r45 1k n5 n3

#top half of schematic
... 16 more lines ...
Attachment 6: eomHVfit.fil
zero 215.7n #zero0
pole 218.4 #pole0
pole 29.1k #pole1
pole 30M #pole2
zero 2.7M #zero1
delay 32.4n 
zero 15M 1k #zero2
pole 168.5M 1M #pole2
factor -139n

... 12 more lines ...
Attachment 7: pzt.fil
#stage 1
op u8 op27 n3 n2 n4
r r35 560 n1 n2
r r38 499 gnd n3
r r30 4.87k n4 n2
c c36 3.3n n4 n2

#stage 2
r r36 750 n4 n5
op u9 ad797 gnd n5 n6
... 29 more lines ...
Attachment 8: pztfit.fil
pole 10k #pole0
pole 32.5k #pole1
pole 240.6k 3 #pole 2
zero 239.1k 20 #zero0
pole 14.2M #pole3
pole 5.4M 1.2 #pole4
pole 14.2M #pole5
pole 14.2M #pole6
delay 6.8n
factor 200
... 15 more lines ...
Attachment 9: EOM.jpg
EOM.jpg
Attachment 10: EOM_HV.jpg
EOM_HV.jpg
Attachment 11: PZT.jpg
PZT.jpg
  622   Thu Jun 30 03:25:13 2011 KojiDailyProgressPMCneed help improving LISO models

For EOM HV:

You need a single zero (or double zero) in order to make the bend of the magnitude curve at high freq.
However, this adds the phase advance at around the cut off freq although you actually have the phase delay.
This means that you need additional double pole just above the measurement freq. This pole may have relatively
high Q so that it can cancel the phase advance by the zero.

 

For EOM circuit:

A zero is missing at around 10^5 Hz. Also a pole or several poles are necessary to realize the magnitude curve and the phase delay at the high freq edge.

 

Filter1:

Limit the freq range from 10^5 instead of rediculous 10^-9.

 

PZT:

Replace one or two single poles by a double pole.

  621   Wed Jun 29 18:01:30 2011 Raphael CervantesDailyProgressFSSSuppressed Free Running Noise

The suppressed NPRO free running noise is: suppressed NPRO free running noise=NPRO free running noise / (1+FSS Open Loop Gain)

I explained the measurement procedure in a previous elog. I made a mistake when I converted the FSS Open Loop Gain from dB to absolute units.

Frank mentioned that he believed the FSS Open Loop Gain is too low.

Attachment 1: FSS_Open_Loop_Gain.jpg
FSS_Open_Loop_Gain.jpg
Attachment 2: NPRO_noise.jpg
NPRO_noise.jpg
Attachment 3: Suppressed_NPRO_noise.jpg
Suppressed_NPRO_noise.jpg
  620   Wed Jun 29 15:59:51 2011 JenneDailyProgressPMCLISO models

Quote:

Just a thought....

If you're going to post LISO models (the .fil files), it might be handy to also include a sketch of the circuit, or a link to the circuit.

 Already done. Sorry about that.

  619   Wed Jun 29 15:47:40 2011 JenneDailyProgressPMCLISO models

Just a thought....

If you're going to post LISO models (the .fil files), it might be handy to also include a sketch of the circuit, or a link to the circuit.

  618   Wed Jun 29 15:35:06 2011 Raphael CervantesDailyProgressFSSneed help improving LISO models

I've tried making LISO models for the PZT path, EOM path, the EOM HV path, and the notch filter right next to TP2 on the RF path. The files for these are entitled pzt.fil, eom.fil, eomHV.fil and filter1.fil. From there, I tried to fit poles and zeroes to the LISO models. These files are named pztfit.fil, eomfit.fil, eomHVfit.fil, and filter1fit.fil. The results of these fits are displayed on the attached graphs. As you can you, I've had trouble fitting poles and zeroes at higher frequencies. If someone can help me with that, that would be fantastic.

The purpose of these fits is so that I can obtain all the poles and zeroes necessary to put into a simulink file simulating the entire FSS setup. We plan to compare the entire FSS setup to the measured transfer function to see how they are different.

Attachment 1: eom.fil
op u4 ad797 gnd n3 n4
c c23 1u n1 n2
r r22 1.1k n2 n3
r r19 24.9k n4 n3
c c15 3.3n n4 n3
c c24 1u n4 n5
r r24 1.3k n5 n6
op u5 ad797 gnd n6 n7
r r20 47k n7 n6
c c18 33p n7 n6
... 19 more lines ...
Attachment 2: eomfit.fil
zero 921n #zero0
zero 84.5m 1k #zero1
pole 700 #pole0
pole 399 #pole1
pole 112 #pole2
pole 146 #pole3
pole 1.9k #pole4
zero 191k #zero2
#zero 101k #zero3
#zero 506k #zero4
... 14 more lines ...
Attachment 3: eomHV.fil
#bottom half of schematic
c c21 1n n1 n2
c c22 220n n1 n2
r r42 3.3k n2 n3
op u6 pa85 gnd n3 n4
r r47 100k n4 n3
c c25 47p n4 n5
r r45 1k n5 n3

#top half of schematic
... 16 more lines ...
Attachment 4: eomHVfit.fil
zero 60.2n
pole 218.7
pole 29.3k
pole 2.1M
#pole 7.04M
zero 1.4M
delay 5.5n
factor -139n

param factor 1f 1M
... 9 more lines ...
Attachment 5: filter1.fil
l l3 1.2u n1 n2
c c12 47p n1 n2
c c13 220p n1 gnd
c c14 220p gnd n2
uinput n1
uoutput n2
freq log 1n 1000M 10000
Attachment 6: filter1fit.fil
pole 8.35M 1k
zero 21.14M 1k
factor 1

param factor 1n 1M
#param pole0:f 1M 20M
#param zero0:f 1M 20M
fit filter1.out dbdeg rel

Attachment 7: pzt.fil
#stage 1
op u8 op27 n3 n2 n4
r r35 560 n1 n2
r r38 499 gnd n3
r r30 4.87k n4 n2
c c36 3.3n n4 n2

#stage 2
r r36 750 n4 n5
op u9 ad797 gnd n5 n6
... 29 more lines ...
Attachment 8: pztfit.fil
pole 9.6k #pole0
pole 33k #pole1
pole 222k 3 #pole 2
zero 233k 50 #zero0
pole 4M #pole3
pole 8M
pole 8M
pole 7M
pole 5M
factor 1
... 14 more lines ...
Attachment 9: EOM_HV_Magnitude_Circuit_Model_vs_Pole_Zero_Model.jpg
EOM_HV_Magnitude_Circuit_Model_vs_Pole_Zero_Model.jpg
Attachment 10: EOM_HV_Phase_Circuit_Model_vs_Pole_Zero_Model.jpg
EOM_HV_Phase_Circuit_Model_vs_Pole_Zero_Model.jpg
Attachment 11: EOM_Magnitude_Circuit_Model_vs_Pole_Zero_Model.jpg
EOM_Magnitude_Circuit_Model_vs_Pole_Zero_Model.jpg
Attachment 12: EOM_Phase_Circuit_Model_vs_Pole_Zero_Model.jpg
EOM_Phase_Circuit_Model_vs_Pole_Zero_Model.jpg
Attachment 13: Filter1_Magnitude_Circuit_Model_vs_Pole_Zero_Model.jpg
Filter1_Magnitude_Circuit_Model_vs_Pole_Zero_Model.jpg
Attachment 14: Filter1_Phase_Circuit_Model_vs_Pole_Zero_Model.jpg
Filter1_Phase_Circuit_Model_vs_Pole_Zero_Model.jpg
Attachment 15: PZT_Magnitude_Circuit_Model_vs_Pole_Zero_Model.jpg
PZT_Magnitude_Circuit_Model_vs_Pole_Zero_Model.jpg
Attachment 16: PZT_Phase_Circuit_Model_vs_Pole_Zero_Model.jpg
PZT_Phase_Circuit_Model_vs_Pole_Zero_Model.jpg
Attachment 17: D040105-C.pdf
D040105-C.pdf D040105-C.pdf D040105-C.pdf
Attachment 18: d0901893-a.pdf
d0901893-a.pdf d0901893-a.pdf d0901893-a.pdf
Attachment 19: D0901894-A.pdf
D0901894-A.pdf
  617   Mon Jun 27 23:55:43 2011 Raphael CervantesDailyProgressFSSFSS measurements progress

Update: Numbers are wrong. Will talk to Tara tonight. Open Loop transfer function is too low and I probably made a mistake with the suppressed noise plot.

 

Common Gain Knob and Fast Gain Knob Calibrations

Objective: The knobs on the FSS servo boards adjust the gain of the AD602JR variable gain amplifiers between Out 2 Common and Out 1 Fast. However, the knobs are simply numbered between 1 and 1000 and do not tell us what the actual gain of these amplifiers are. We aim to measure calibrate these knob numbers to actual gain numbers in dB.

Procedure: We measure the open loop gain from Out 2 Common to Out 1 Fast. The path includes two variable gain amplifiers, AD602JR, one of which is controlled by the Fast Gain knob and the other which is controlled by the Common Gain knob.

  1. We start off with both knobs at 0.
  2. We use a spectrum analyzer to inject a signal into RF test. We use Out 2 Common as the reference and Out 1 Fast as the response.
  3. We measure the transfer function of this path. We pick a frequency to monitor it's change as we vary the Fast and Common Gain. In our case we picked 1.05 kHz and measured it's gain to be 12.4dB when both knobs are 0.
  4. We  vary the common gain while we keep the Fast Gain constant and measure the subsequent gain at 1.05kHz. We than fit a line through the measured points to obtain an equation for the response of the transfer function as a function of the Common Knob number.
  5. We repeat the same measurement by setting the Common Gain knob to 0 and measuring the gain at 1.05kHz as we vary the Fast Gain Knob.

Results:

These graphs show the measured transfer function from Out 2 Common to Out 1 Fast.

Gain_Knobs_Calibration_Magnitude_Transfer_Function.jpgGain_Knobs_Calibration_Phase_Transfer_Function.jpg

We fit a line into both measurements and obtain the following:

Total Gain (dB) = 0.04* Common Knob # + 0.04*Fast Knob # +12.2

Graph for this fitting will come later if necessary.

 Measuring the Suppressed NPRO Free Running Noise

 Objective:

Procedure: We recognize that the suppressed NPRO free running noise is calculated by the expression: NPRO suppressed free running noise = NPRO free running noise/(1 + FSS Open Loop Gain). We noite that the FSS Open Loop gain is in absolute units and not in dB.

Measuring the Open Loop Gain:

To get the FSS Open Loop Gain, we use a spectrum analyzer to measure the open loop gain from Common Out 2 to Common Out 1 and the open loop gain from Common Out 1 to Common Out 2. Since the transfer function is in dB, calculating the entire FSS Open Loop Gain is just a matter of adding the two transfer functions together. To measure the transfer function of from Common Out 2 to Common Out 1, we use the spectrum analyzer to inject a signal into the Common Excitation and take Common Out 2 as the reference and Common Out 1 as the response. To measure the transfer function from Common Out 1 to Common Out 2, we inject a signal into RF Test and take Common Out 1 as the reference and Common Out 2 as the response.

Results:

We add both transfer functions to get the FSS Open Loop Gain and obtain the following result:

FSS_Open_Loop_Gain.jpg

NPRO Free Running Noise:

Frank gave me his data for his NPRO free running noise and I fit the data into a line. I am justified in excluding certain points because of the line harmonics in the lower frequency and the equipment noise floor in the higher frequency. log(free running noise) = -1.08 * log(frequency) +4.53

NPRO_noise.jpg

Suppressed Free Running Noise:

We simply calculate NPRO suppressed free running noise = NPRO free running noise/(1 + FSS Open Loop Gain) where the FSS Open Loop Gain is in absolute units insted of dB to get the following result.

Suppressed_NPRO_noise.jpg

 

Attachment 1: Gain_Knobs_Calibration_Magnitude_Transfer_Function.jpg
Gain_Knobs_Calibration_Magnitude_Transfer_Function.jpg
Attachment 2: Gain_Knobs_Calibration_Phase_Transfer_Function.jpg
Gain_Knobs_Calibration_Phase_Transfer_Function.jpg
Attachment 3: Suppressed_NPRO_noise.jpg
Suppressed_NPRO_noise.jpg
Attachment 4: FSS_Open_Loop_Gain.jpg
FSS_Open_Loop_Gain.jpg
Attachment 5: NPRO_noise.jpg
NPRO_noise.jpg
  615   Thu Jun 16 00:03:11 2011 taraDailyProgressFSSTF measurement on FAST path for TTFSS

I measured TF of Fast path from TTFSS and compared with LISO model. The results agree well.

      == Motivation ==

     This work is a part of loop characterization. The goal is to understand why the UGF of the open loop gain is only ~200kHz instead of ~500 kHz. So I check if the TF of the loop match the design or not. FAST path is working as designed, see Bode plot of the TF below. I'll measure EOM path next.

      == setup ==

       I used SR785 to measure the TF. The source from RFout was split by a T, one for Ref (chA), another one for excitation at Test in. The Response was taken from Out2 to ChB. I used an oscilloscope to monitor Out2 to make sure that the signal is not saturated. The gain setup (common/fast) of TTFSS was set to 350/100. The power input of the laser was 1mW.

TF_fast.pdf

FIGURE1: Setup for fast path measurement. The ports are shown on the corresponding schematic by green arrows.


     

fastpath.png

FIGURE2: Comparison between LISO model and TF measurement of Fast path. The model is offset to compensate for common gain/ fast gain setup in the servo.

 

Next is to measure EOM path and compared with LISO, then the whole OLG TF of the setup. 

  614   Tue Jun 14 23:07:03 2011 taraDailyProgressopticoptic layout for new fss setup

The mode matching for new FSS is calculated. The plan is shown below.

 

Note for the setup:

1) the spotsize in the AOM is 200um, the specsheet says 550 um (I might have to correct this).

2) Two AOMs are of the same model.

3) For mode matching to the AOM in acav path, I used only a single lens.

4) focal lengths of the lenses are in mm, We have to order the one with * (f = 57.4 mm)

5) Both cavities are 1" apart (3" from center to center)

6) Mistake in the drawing: the x2 QWPs just before the beams enter the vacuum chamber should be placed before the periscopes, not after.

2aom_fss.png

  613   Wed Jun 8 18:12:46 2011 taraNotesFSSTF of EOM

 

   I measured the TF of the 35.5 MHz EOM. The TF will be fitted by LISO and used in SIMULINK model later.

     ==setup==

   eomtf.pdf

     ==measurement==

0) the laser power used is 1mW, common gain/fast gain = 150/100. I reduced the gain as much as possible so that the signal from mixer out reflects the real frequency shift due to the excitation.

1) The TF between the excitation to EOM(ref) and mixer out is measured.

2) Measure the slope of the error signal at mixer out to get a calibration factor from V to Hz (0.22 MHz/V)

3) Convert the TF from (1) which is in [Volt(Mixer out) / Volt (excitation)] to [Hz / Volt(excitation)] and plot.

EOMTF.png

 

The magnitude part of the bode plot is in Hz/Volt. 

4) The TF will be fit by LISO, and used in simulink model.

  612   Thu Jun 2 21:22:02 2011 taraNotesFSScharacterizing fss loop

I listed what I need to do to fully characterize the fss loop.

 

1) Measure the whole TF of the loop. This will be compared with the simulation.

2) Simulate the whole TF of the loop. This requires:

      2.1) Use LISO to simulate the TFs of FAST path and PC path

      2.2) measure the TF of the pzt (actuator for fast path)

      2.3)measure the TF of the EOM (actuator for PC path)

3) Make a SIMULINK model

  611   Wed Jun 1 23:49:57 2011 taraNotesFSScharacterizing fss loop

I'm checking the TTFSS loop to understand why the UGF is so low (~100k instead of ~ 500k). I started from comparing the EOM path between that of the old FSS and the current one. Their shapes are not similar. The current TTFSS does not roll off at high frequency as the old fss does.

 

   The previous fss (blue) has UGF around 500 kHz, while the current TTFSS has UGF around 100k, see this entry. So I check what's the designed TFs for EOM paths between the two servos look like. These TFs, simlated from LISO model, are taken between the end of the common path and just before the signal is amplified by a high voltage opamp. See the schematic below for the details. 

    LISO codes for simulation can be found in the attached file section. File eom_old.fil for old fss note that I used ideal opamp for ad847 in old fss (I don't have it in my lib). File 11g.fil is for current TTFSS

 new_old_eom.png

fig1: TF from EOM path. From current TTFSS(red) and old FSS(blue).

 

EOM_fss.png

 fig 2, EOM path in old fss, where I used LISO to simulate the TF. Note that I used ideal opamp for ad847.

EOM_ttfss.png

 

 fig3, EOM path in current TTFSS.

 

I'm working on measuring TF of the whole loop and compared it with the simulation. The result will be posted later.

Attachment 2: eom_old.fil
#EOM path
#d980536-e-c
# Tara C, 2011_06_01


c C13 .047u  nin n1
r R17  1.1k  n1 n2
r  R14  24.9k  n2 n3
c  C18   3300p n2 n3

... 40 more lines ...
Attachment 3: 11g.fil
#EOM path 
#TTFSS D040105
# Tara C, 2011_06_01


c  C11 330p  nin n3
r  R14 3.01k  n3 n4
r  R15 100    n4  n5
l  L1  220u   n4 n6
c  C13 30p    n6 gnd
... 41 more lines ...
Attachment 5: EOM_ttfss.png
EOM_ttfss.png
  610   Wed Jun 1 21:00:05 2011 FrankNotesopticpurchases

we have to design our own. The 40m one has 2" mirrors (too large, we don't have the space), wrong height for incoming/outgoing beam and is clamped to the table, which i think is bad in terms of stability.
The design principle does not look much different compared to the original refcav periscope design, except for the mirror holder itself. That was bad designed for the old one.

 

Quote:

They are found in DCC. Some comments

- You can not steer the beam. The beam should be steered before or after the periscope.

- The side plates were too thick. It can be 1/2" thickness to reduce the total weight.

D1001446-v1 40m Vertex Green Locking Periscope A Base Daisuke Tatsumi et al. Auxiliary Optics
Basic R&D
15 Jul 2010
D1001447-v1 40m Vertex Green Locking Periscope A Sidebar Daisuke Tatsumi et al. Auxiliary Optics
Basic R&D
15 Jul 2010
D1001448-v1 40m Vertex Green Locking Periscope A Mirror Holder Daisuke Tatsumi et al. Auxiliary Optics
Basic R&D
15 Jul 2010
D1001613-v1 40m Vertex Green Locking Periscope A PTFE Post Koji Arai Auxiliary Optics
Basic R&D
15 Jul 2010

Quote:

I looked up 40m elog and found Daisuke's design for periscope. I'll make a sketch FSS' periscopes.

The design for 40m pericopes by Daisuke can be found here .

Quote:

The periscopes for the refcav ought to be made custom. None of the store bought type are stiff enough. Koji has a design from the 40m green that Daisuke made.

 

 

 

 

  609   Wed Jun 1 01:50:35 2011 KojiNotesopticpurchases

They are found in DCC. Some comments

- You can not steer the beam. The beam should be steered before or after the periscope.

- The side plates were too thick. It can be 1/2" thickness to reduce the total weight.

D1001446-v1 40m Vertex Green Locking Periscope A Base Daisuke Tatsumi et al. Auxiliary Optics
Basic R&D
15 Jul 2010
D1001447-v1 40m Vertex Green Locking Periscope A Sidebar Daisuke Tatsumi et al. Auxiliary Optics
Basic R&D
15 Jul 2010
D1001448-v1 40m Vertex Green Locking Periscope A Mirror Holder Daisuke Tatsumi et al. Auxiliary Optics
Basic R&D
15 Jul 2010
D1001613-v1 40m Vertex Green Locking Periscope A PTFE Post Koji Arai Auxiliary Optics
Basic R&D
15 Jul 2010

Quote:

I looked up 40m elog and found Daisuke's design for periscope. I'll make a sketch FSS' periscopes.

The design for 40m pericopes by Daisuke can be found here .

Quote:

The periscopes for the refcav ought to be made custom. None of the store bought type are stiff enough. Koji has a design from the 40m green that Daisuke made.

 

 

 

  608   Tue May 31 19:01:12 2011 taraNotes Plan for FSS

 I listed some issues for FSS experiment that we should discuss within this week and other small details.

              ==important issues==

  • ***Design Suspension of both cavities in the same vacuum chamber ***
  • ** Decide if we want to active control the cavities' temperature or not, so we can choose

           1) heater/insulation for the cavities, If we don't want to actively control the temperature, insulation inside the chamber might not be necessary.

           2) the appropriate AOM type, and

          3) Local oscillator for beat measurement

  •   Experiment's Goal, do we want to develop a new TNI? if so, should we redesign the cavity so that the spot size is smaller, and make the cavity shorter in order to bring up the thermal noise level and become less sensitive to seismic noise?

             

              ==smaller topics==

  •  Design for RFPD, do we want to change the sideband frequency?
  • Servo for ACAV path, where do we get one, will we use the previous FSS?
  • New periscope, do we want fixed or adjustable mirror mounts?
  •  Beam splitter for the beat path, is it ok to use a cube bs with standard bs mount?

 

 

  607   Tue May 31 11:31:18 2011 taraNotesopticpurchases

I looked up 40m elog and found Daisuke's design for periscope. I'll make a sketch FSS' periscopes.

The design for 40m pericopes by Daisuke can be found here .

Quote:

The periscopes for the refcav ought to be made custom. None of the store bought type are stiff enough. Koji has a design from the 40m green that Daisuke made.

 

 

  606   Mon May 30 15:23:53 2011 ranaNotesopticpurchases

The periscopes for the refcav ought to be made custom. None of the store bought type are stiff enough. Koji has a design from the 40m green that Daisuke made.

 

  605   Wed May 25 18:29:09 2011 taraNotesopticpurchases

I ordered a few opto mechanical components to replace the current shaky periscopes.

The new  periscope is shown here, elog:574. Currently we have only one set, so I ordered a post clamp to complete another set. 

I also ordered 4 mirror mounts that can be mounted on 45 degree mounting adapters. The thickness of these mounts are thinner than a regular mirror mount, so it can be fit on the adapter. I plan to use these in Crackle experiment as well. 

 

 

  604   Thu May 19 23:16:44 2011 Tara ChalermsongsakNotesRefCavHOM, from carrier and both sidebands

 I made a plot for HOM from 24.5 MHz sidebands.  The 17th and 20th order are quite close to the main frequency and its sidebands.

      We were thinking about using 24.5 MHz sidebands, but it seems that 35.5 MHz is safer for the current setup.

 

 RA: plot deleted - put some details into the ELOG to explain your plots to prevent deletion.

 


HOM frequency shift for RefCav is plotted below.

Y axis is the frequency shift due to Gouy phase in MHz.

X axis is the (n+m)th order of the Hermite Gauss mode

The waist of the beam inside the cavity is 261 um* (symmetric cavity, R =0.5m, L = 0.2032 m.) 
Thus, the frequency shift between n+m+1 and n+m mode is 219.763 MHz. (see Lasers, p 762 for details)

Blue line represents the 0th order of the carrier's frequency (thus y axis =0) The purple and the brown lines, at y= 24.5 and -24.5 MHz, are the 0th of + and - sidebands respectively.

The color dots represent the frequency shift from Higher order mode which is specified on x-axis, blue for HOM from the carrier's frequency, purple and brown for HOM from +/- sidebands.

Attachment 1: HOM_2011_05_19.png
HOM_2011_05_19.png
  603   Thu May 19 11:32:46 2011 taraDailyProgressopticoptic layout for new fss setup

I added more details on the layout, and necessary half wave plates in the beam path.

fss_2011_05_18.png

  602   Wed May 18 20:51:35 2011 taraDailyProgressopticoptic layout for new fss setup

I planned the layout for new fss setup.

The new setup has 1) both cavities placed in the same vacuum chamber, 2) two AOMs used in both RCAV and ACAV paths, 3) more compact beat path.

 In the layout, I assumed that

  • Two cavities in the chamber are 3 inches apart.
  • Two AOMs are of the same model, have the same setup
  • There is no change of plan for the layout between PMC and the laser

This is just a plan, no mode matching has been calculated yet.

I am concerned  that the mode matching lens might block the beam in ACAV path where the incoming beam and reflected beam cross, but this can be adjust later.

The outer foam box will be smaller, but it should have enough space to keep some electronics inside like we have now.

I should find two similar sets of beam splitters/ mirrors for beams in the beat path behind the cavity. So the pick up beams from two cavities can have same power.

Right now the power going into two PDs for RCTRANSPD are not the same because the splitter are not the similar.

Note that we might install a platform  behind the cavities so that we don't need the periscopes to lower the beam, and get rid of their associated mechanical peaks.

fss_layout_2011_05_18.png

  601   Wed May 18 15:30:12 2011 taraNotesEnvironmentLIGO properties check

Rod sent a guy to check on equipments that have LIGO property tags on them.

I showed him around the lab to scan the equipments with the tags.

  600   Mon May 16 22:59:45 2011 frank, taraDailyProgressNoiseBudgetRIN coupling coefficient

I tried to measure coherence TF between RCTRANSPD and FASTMON, but I could not. After I adjusted gain setting, RFAM adjustment,electronic offset , I could not observe frequency shift due to amplitude modulation.

      From this entry, I measure the power spectrum of transmitted power(RCTRANSPD) and laser frequency change(FASTMON). But it is only one peak at 10Hz. This does not give much information about how much the frequency shift comes from RFAM or the real RIN coupling effect.  So, I want to measure the TF between FASTMON and RCTRANSPD, then misaligned the polarization of the laser into the EOM to increase the RFAM effect ,and see how much TF will change.

     Since I already aligned the EOM to minimize the RFAM effect, and adjusted the electronic offset. I checked if I could see the peak in FASTMON due to amplitude modulation or not. If there is no peak, it means I cannot measure TF. It turned out that I could not see the peak at amplitude modulating frequency anymore.

 I tried:

  • changing power input to be 1mW, 5mW and 10mW and Vmod was varied from 2 Vpkpk to 10 Vpkpk.
  • changing amplitude modulating frequency to 5Hz, 10Hz, 20Hz
  • changing modulating function from sine wave to square wave, with various frequencies and amplitudes.

     compare.png

      The above plot shows the PSD of FASTMON (top) and RCTRANSPD (bottom) when the system is amplitude modulated, and no amp modulated. The power input is 10mW, the modulating function is sine wave, 10Vpkpk, the linewidth is 125mHz. Common/Fast gain = 300/430. There is no sign of any peak at 10Hz in FASTMON at all.

    

     I think the peak I saw before came from :

1) electronic offset: it was not optimized before. To adjust it, block the beam on the RFPD, monitor mixer out (via "common out2" channel), and adjust the offset so  that mixerout becomes zero.

2) RFAM: this is explained in here.

3) gain setting: I made a mistake before, as Frank pointed out, when I optimized the gain. I only looked at noise level from mixer out at low frequency without      checking the oscillation at high frequency. Now I adjusted it back so there is no oscillation, and the system is more stable.

       So, to sum up, the coherent TF measurement could not be done for now, because there is no peak in FASTMON. It means that the laser noise is noisier than the RIN coupling effect and any possible residual RFAM effect.

I'll work on the TF of the servo before continuing on the RIN coupling coefficient, but the next test might be trying larger modulation depth.

  599   Wed May 4 01:20:00 2011 taraDailyProgressopticminimizing RFAM/ aligning 35.5 MHz EOM

I minimized the RFAM by aligning the 35.5 MHz EOM and remeasured the RIN coupling coefficient.

The upper limit is 5 [Hz/uW (fluctuation of input power = RIN x Pin) ]@ 10 Hz

(This entry is approved by Kiwamu and is written in his style)Tue May 10 19:20:22 2011

  As pointed out in the LIGO-X meeting that my setup might suffer a lot from RFAM, so I came back to:

  • 1. minimize the RFAM by aligning the 35.5 MHz EOM,
  • 2. determine how much Vmod I can apply to amplitude modulation without exciting the RFAM noise above the background, and
  • 3. remeasure RIN coupling coefficient again with the allowed maximum Vmod.

 


[Setup]

The power input was 1mW as usual.

The frequency of Vmod is 10Hz. The amplitude of Vmod to EAOM for amplitude modulation was varied from 2 to 10 Vpkpk. Common/Fast gain was 500/900. I had to reduce it so the signal is not too large. I measured the spectrum of FASTMON and tried to observe the peak at 10Hz  with 12.5 mHz linewidth. The background level was ~10mV.

I do this to determine what is the maximum driving voltage where the effect from RFAM is still small compared to the background.

____________________________________

Drive Vpkpk     FASTMON peak(Vrms/rtHz)

10                   74.7

8                     48.17

5                     29.6

3                    23.02

2                  ~comparable to BG level ~ 10mV/rtHz

____________________________________

[ 1. aligning EOM ]

I picked up the beam after EOM on RCAV path and sent it to a PD (Thorlabs PDA10A.)  There were 35.5 MHz pick up on the table, so I had to choose where the peak from pickup was minimum. Then I adjusted the half wave plate before the EOM and EOM's pitch/yaw position to minimize the peak.

[ 2. determine max Vmod ]

 Although we want to modulate the power as small as possible to have a good linear approximation, we also need the signal to be large enough to be able to see the effect. However, the alignment of the EOM is not perfect, there will be RFAM effect adds into the signal. If the modulation is too large, the RFAM will mask the real signal.  I need to determine what is the maximum Vmod I can use without having the RFAM effect excited above the background.

     To see the effect of RFAM, I kept the setup similar to what I did with RIN coupling coefficient measurement, but without locking the cavity, and the laser frequency off from the resonance. This will tell us how much "fake signal" is produced by RFAM.

     When the cavity is not locked, all the carrier and sidebands will be incident on the RFPD. The signal should be flat (beat between the carrier and both sidebands cancel each another,) and after it is demodulated by 35.5 MHz from LO, the level should be zero. However, if the amplitude is modulated at 35.5 MHz due to misalignment of the EOM, this will appear as DC signal at the error point. Hence, any power modulation at f0 (for this case, 10 Hz) will multiply up the error signal and cause offset fluctuation and slope change at f0. Slope change is not a problem, but the offset is. It will change the point where the laser will be locked, as the error signal moves up and down. Thus the system will interpret it as frequency noise of the laser and try to fight against it. This will appear as a peak in the FASTOUT spectrum at the modulation frequency, f0.

     I measured the spectrum of FASTOUT (MIXER OUT is another option) to see the effect of RFAM 

[ 3. remeasure RIN coupling coefficient ]

So I used 2Vpkpk drive, locked the cavity, and measured FASTMON again to see if I can measured the RIN coupling or not. The gain was set back to optimum value (common = 970 fast= 900.)  However, there was no observable peak at 10Hz from FASTMON signal. It was quite flat ~100 uVrms/rtHz.

 I made sure that the amplitude was really modulated by checking RCTRANSPD. It had a 5.37 Vrms/rtHz peak at 10Hz with 200mV DC level. Therefore, the laser noise is higher than the thermo-optic effect at this modulation level. I cannot increase modulation depth because the RFAM will mask the signal. 

If I use this number to calculate the coupling coefficient, (flat level of FASTOUT, and peak from RCTRANSPD)

it will be ~ 8 [Hz/ uW of fluctuation of the input power into the cavity] still larger than 1[Hz/uW] as measured at 40m, but it's getting smaller than the last entry (60 [Hz/uW] of input power)

       I still can change the power input, but I think the RFAM will scale up by the same amount and mask the signal again. I'll try that later.

 

 Let's check what does this value give us in the noise budget @10Hz. The input power is 1mW, RIN = 10^-4. Frequency noise will be

8 [Hz/uW] x 1000 [uW] x 10^-4 [RIN] = 0.8 [Hz/rtHz] which is higher than coating noise (10 [mHz/rtHz]@10Hz) So we still cannot ignore the effect.


[Take II]

 I tried 16 mW input power, there was signal from RFAM when I measured FASTOUT with unlocked cavity, the peak was 46 m[Vrms/rtHz] above 10m[Vrms/rtHz] background. Vmod = 1Vpkpk.

When I lock the cavity and measure the coupling:

FASTMON peak = 278.6 uVrms/rtHz

RCTRANSPD peak = 30.82 mV, DC level = 2.57 V, Pin = 16mW. Linewidth = 12.5mHz.

Common/Fast gain = 480/906

Use the calculation from here.

The upper limit for coupling coeffiicient is ~ 5 [Hz/uW]. It is only the upper limit because RFAM effect is still present.

  598   Tue May 3 11:59:55 2011 taraDailyProgressNoiseBudgetRIN coupling to frequency noise

The coupling coefficient I measured before changed with input power/ alignment of the beam. What I measured before might be the RF-AM effect. [figures and more details will be added]

 

I read Frank's entry, and tried to see the effect of other parameters on coupling coefficient. I tried

      1) misaligning  the  beam into the cavity, and

        The power level is still 1mW input. I turned the knob of the periscope to misaligned the beam. RCTRANSPD DC value dropped from 250 mV to 207 mV, the peak at 10Hz in FASTMON disappeared. 

      [add fig]

 

 

      2) changing power input to the cavity. I tried 0.5 mW, 1 mW, 2mW, 5mW

 As I changed the power, I adjusted the gain so that the noise from MIXEROUT remained the same, but at certain power I could not see the peak in FASTMON signal.

[add fig]

 

 

 

  597   Mon May 2 10:32:58 2011 FrankDailyProgressNoiseBudgetRIN coupling to frequency noise

did you try to optimie/minimze the coupling before the measurement? If not you should check how you can make it worse/better (e.g. alignment to cavity, RFPD, polarization, RF-AM etc) and then measure it. It might be that you are sitting on kind of a maximum right now, who knows.

How about measuring a complete TF? Your signal looks large enough to do that.

Quote:

I measured RIN -> frequency noise coupling coefficient at 10Hz from RCAV. The result is 0.02 Hz/uW (frequency shift/power fluctuation built up in the cavity*)

*by power fluctuation built up in the cavity, I mean [RIN x power input as measured at the input periscope x Finesse/ pi ]

 

Following the list I wrote down here

1) TF between FASTOUT/FASTMON is a lowpass with a pole at 10Hz. FASTOUT is fed to the laser for controlling its frequency. FASTMON is a channel for monitoring. This is as expected from the RC low pass filter at FASTOUT.

TF.png

 2) I modulated the beam's polarization and used a PBS so the polarized transmitted beam is amplitude modulated. I used sine wave out from SR785, 2Vpp @ 10Hz connected to the EAOM. I measured the RCTRANSPD signal with an oscilloscope to check the signal is 211 mV +/- 7mV. The power is modualted by ~ 4%.

Then measure the peak at the modulating frequency (10Hz) from RCTRANSPD and FASTMON spectral density with SR785. The linewidth is 0.976 mHz. FASTMON was connected to chA, RCTRANSPD to chB. The data were averaged over 4 samples.

  The DC level of RCTRANSPD was 220 mV.

 PSD.png

 The peak from FASTMON and RCTRANSPD are 6.75 mV/rtHz and 55.2 mV/rtHz. To convert this to the coupling coefficient we have to:

1) convert the voltage output from FASTMON to FASTOUT, the TF is measured and shown above. 

    V_fastout = V_fastmon x sqrt(  1 /   1 + (f/10Hz)^2  ), so  for 10Hz,

    V_fastout = V_fastmon x sqrt (1/2)

2) convert the Voltage out to frequency change by 3.09 MHz/V factor.

3) convert RCTRANSPD to RIN by dividing the PSD by DC level (220mV)

4) convert RIN to power fluctuation by x power input (1mW) x Finesse (9710) / pi

 

or

Freq/Hz   =  Vmon   x TF  x 3.09 [MHz/V]   

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

                 V_RCtranspd / rctranspd_DC     x Pin  x Finesse/pi

 

  For 10Hz, 1mW power input, rctranspd_dc = 220 mV, we get

  RIN coupling coeff at 10Hz  =        0.0188 Hz/uW   = 18.8 mHz/uW.

 

I repeated the measurement with:

1)half modulating power, 1Vpp. the coupling coefficient decreases to  0.015 Hz/uW, so it seems that the modulation range I chose might be too large so the effect was not linear.  The coefficient should remain constant!!

2) 20Hz modulating power, 2Vpp (but the line width was 7.8 mHz, not 1mHz for quick measurement)

, the coupling coefficient is  9.35  mHz/uW. The result is smaller than that of 10Hz, which is expected from 1/f effect, but I think I should have used the same line width and look at a few more frequencies.

 

I'll have to check the coupling coefficient from ACAV, and check the beat signal to see if they are canceled or not.

 


Comparison to the calculation I did

The calculation I did here gives the noise at 10Hz from 10mW input, RIN = 10^-4, Finesse 10^4, to be 1.4 [mHz/rt Hz] which is already lower than the coating noise (10mHz/rtHz at 10Hz)

I can convert it to the estimated coupling coefficient to compare with what I measured.

1.4 mHz/rtHz       =   coupling coeff [Hz/W] x  Pin [W] x RIN [1/rtHz] xFinesse/pi,  or

coupling [Hz/W]    =   1.4 mHz/rtHz  / Pin [W]  / RIN [1/rtHz] / (Finesse/pi)

                            =   0.44 Hz/W

This is 5 order of magnitude lower than what I measured !!! and the calculated noise is only one order of magnitude lower than the coating. It means that the calculation is not correct and we cannot ignore the effect. we will run into it before reaching coating thermal noise if the effects on both cavities aren't canceled.

 

 

  596   Sun May 1 22:39:56 2011 taraDailyProgressNoiseBudgetRIN coupling to frequency noise

I measured RIN -> frequency noise coupling coefficient at 10Hz from RCAV. The result is 0.02 Hz/uW (frequency shift/power fluctuation built up in the cavity*)

*by power fluctuation built up in the cavity, I mean [RIN x power input as measured at the input periscope x Finesse/ pi ]

 

Following the list I wrote down here

1) TF between FASTOUT/FASTMON is a lowpass with a pole at 10Hz. FASTOUT is fed to the laser for controlling its frequency. FASTMON is a channel for monitoring. This is as expected from the RC low pass filter at FASTOUT.

TF.png

 2) I modulated the beam's polarization and used a PBS so the polarized transmitted beam is amplitude modulated. I used sine wave out from SR785, 2Vpp @ 10Hz connected to the EAOM. I measured the RCTRANSPD signal with an oscilloscope to check the signal is 211 mV +/- 7mV. The power is modualted by ~ 4%.

Then measure the peak at the modulating frequency (10Hz) from RCTRANSPD and FASTMON spectral density with SR785. The linewidth is 0.976 mHz. FASTMON was connected to chA, RCTRANSPD to chB. The data were averaged over 4 samples.

  The DC level of RCTRANSPD was 220 mV.

 PSD.png

 The peak from FASTMON and RCTRANSPD are 6.75 mV/rtHz and 55.2 mV/rtHz. To convert this to the coupling coefficient we have to:

1) convert the voltage output from FASTMON to FASTOUT, the TF is measured and shown above. 

    V_fastout = V_fastmon x sqrt(  1 /   1 + (f/10Hz)^2  ), so  for 10Hz,

    V_fastout = V_fastmon x sqrt (1/2)

2) convert the Voltage out to frequency change by 3.09 MHz/V factor.

3) convert RCTRANSPD to RIN by dividing the PSD by DC level (220mV)

4) convert RIN to power fluctuation by x power input (1mW) x Finesse (9710) / pi

 

or

Freq/Hz   =  Vmon   x TF  x 3.09 [MHz/V]   

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

                 V_RCtranspd / rctranspd_DC     x Pin  x Finesse/pi

 

  For 10Hz, 1mW power input, rctranspd_dc = 220 mV, we get

  RIN coupling coeff at 10Hz  =        0.0188 Hz/uW   = 18.8 mHz/uW.

 

I repeated the measurement with:

1)half modulating power, 1Vpp. the coupling coefficient decreases to  0.015 Hz/uW, so it seems that the modulation range I chose might be too large so the effect was not linear.  The coefficient should remain constant!!

2) 20Hz modulating power, 2Vpp (but the line width was 7.8 mHz, not 1mHz for quick measurement)

, the coupling coefficient is  9.35  mHz/uW. The result is smaller than that of 10Hz, which is expected from 1/f effect, but I think I should have used the same line width and look at a few more frequencies.

 

I'll have to check the coupling coefficient from ACAV, and check the beat signal to see if they are canceled or not.

 


Comparison to the calculation I did

The calculation I did here gives the noise at 10Hz from 10mW input, RIN = 10^-4, Finesse 10^4, to be 1.4 [mHz/rt Hz] which is already lower than the coating noise (10mHz/rtHz at 10Hz)

I can convert it to the estimated coupling coefficient to compare with what I measured.

1.4 mHz/rtHz       =   coupling coeff [Hz/W] x  Pin [W] x RIN [1/rtHz] xFinesse/pi,  or

coupling [Hz/W]    =   1.4 mHz/rtHz  / Pin [W]  / RIN [1/rtHz] / (Finesse/pi)

                            =   0.44 Hz/W

This is 5 order of magnitude lower than what I measured !!! and the calculated noise is only one order of magnitude lower than the coating. It means that the calculation is not correct and we cannot ignore the effect. we will run into it before reaching coating thermal noise if the effects on both cavities aren't canceled.

 

  595   Wed Apr 27 22:09:10 2011 taraDailyProgressNoiseBudgetRIN coupling to frequency noise

 I list the plan for measuring the effect of RIN to Frequency noise.

 

     1) Measure the TF between FASTMON / RCTRANSPD. FASTMON [V]will be converted to frequency [Hz]. RCTRANSPD will be converted to power of the laser [W]. This will be useful once we know the coupling coefficient and want to project the noise in the noise budget.

      1.1) TF between FASTOUT and FAST MON will be measured. We observe the signal from FASTMON, but the signal from FASTOUT will drive the laser, so they will not be exactly the same because of other electronic components along the paths, for example, low pass filter at 10Hz for FASTOUT.  We already know that the calibration at FAST modulation at the laser input is 3.09 [MHz/V].

     2) Amplitude modulate the laser power by modulating the polarization of the laser via EAOM. The beam is sent to a PBS which allows only one polarization out. Thus the amplitude of the transmitted beam through the PBS will be modulated.  I choose to modulate at 10 Hz, because the frequency is not too high so the effect is small, and the frequency is not too low that it might be masked by seismic noise.The modulating voltage is 2Vpkpk, I want the Vmod to be large enough, so that I can see the signal, but not too large because we want the system to have linear response.

     3) Measure the PSD of FASTOUT, RCTRANSPD, make sure that the peak at 10 Hz is visible.

     The input power is 1mW, I will increase more power later. I want to see if  the signal is measurable at this power level. If so, the higher power should increase the signal by the same factor.

    

    To sum up, I'll repeat the measurement with

    1) changing the laser power

    2) changing the modulation depth

   3) changing the modulating frequency

    and see if the effect behaves like what we expect. It should be smaller as the laser power or modulation depth decrease, or at higher frequency.

ELOG V3.1.3-