40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  Cryo Lab eLog, Page 58 of 61  Not logged in ELOG logo
ID Date Author Type Category Subject
  202   Sat Jun 18 13:17:14 2011 FrankThings to BuySensorshigh temperature Germanium sensors

this company makes them: http://www.adsem.com/

The price for regular thermistors SE-1M is ~$40 each. The price for the resistance value in the range of 0.5-1MOhm at 77K is about $10 more, lead time 3 weeks.
With calibration they are about a factor of two more expensive, but i think we don't need absolute calibrated ones as we don't know the exact temperature we need and we have out PT sensors which follow a standard curve (
DIN43760 and IEC751). Alpha is about -2.3%/K for the NTCs which is much higher than any other sensor at 125K.
Minimum order is $100.

Below the datasheet for the SE-1M. They don't have any datasheets on their website.

SE-1M_NTC_thermistors_data_sheet_196_6-2011.pdf

  201   Fri Jun 17 23:21:58 2011 FrankCryostatDrawingsvaccum system

here a picture of what i think the vacuum system could look like. There is a flex hose missing from the valve on top of the 6-way cross to the tee attached to the other valve (the valve for the outer insulating vacuum).
There are two gauges, one for the insulating vacuum (down to 10-5 Torr), mounted on a tee, which we don't need (the tee) but i wanted to see if there is enough space to add one just in case we want to add another feedthrough over there.
The other is a better model which can measure down to 10-8 Torr. There are also two 24pin connectors on the cross and an open port for the needle valve for the exchange gas. The two open ports are connected to the LN2 reservoir.

Full_system.jpg


Will add the schematic for it later.

 

  200   Fri Jun 17 17:09:47 2011 FrankMiscGeneralProject management software - Ganttproject

Quote:

i've checked out the other software Rana suggested - works fine. Here the project file and an example  pdf-report i've exported. Import/Export to MSP also works fine.

 Cryo.gan

Cryo.pdf

 

Didn't add the stuff from the old one. Files can be also found on the svn in the Cryo folder

 

I updated this and put in on the wiki:

Wiki Link

  199   Fri Jun 17 14:56:52 2011 JennyDailyProgress Running COMSOL remotely

Here goes my first elog entry. Today I figured out how to launch COMSOL remotely on the 40-meter computer (Megatron):

I’m using the terminal window (bash) on my Mac OS X.

1)   log onto nodus, but use ssh –X instead of just ssh (the -X allows you to run applications remotely and have the windows appear on your local screen. I found that tip here: http://maketecheasier.com/use-ssh-x-forwarding-to-run-remote-apps/2009/06/22)

2)   type ssh –X megatron and enter the password when prompted

3)   enter comsol &

 

A COMSOL Multiphysics window should open in X11. The first time I did this it opened the window but did not load the graphics. I quit everything and started again, logging on to nodus. The second time everything worked.

 

Questions:

1)   Can I save files to my local laptop while working remotely? How do I do this? Or do I have to save them remotely and then transfer them to my computer?

2)   When I launch comsol, in my terminal window it says: Xlib: extension “RANDR” missing on display “localhost:10.0”  I’m not sure what this means. Could it be something important to pay attention to?

 

To do list:

1) Learn how to navigate COMSOL 4.0.

2) Study heat-transfer (radiation and conduction) equations.

3) Build simple 1-D model, simulate heat conduction over time, try to extract the transfer function for the system.

  198   Fri Jun 17 02:14:23 2011 FrankMiscGeneralProject management software - Ganttproject

i've checked out the other software Rana suggested - works fine. Here the project file and an example  pdf-report i've exported. Import/Export to MSP also works fine.

 Cryo.gan

Cryo.pdf

 

Didn't add the stuff from the old one. Files can be also found on the svn in the Cryo folder

 

  197   Thu Jun 16 23:47:50 2011 DmassLaserLab WorkMore (non)progress on diode laser locking

I had some more non success on laser diode locking.

I changed to P polarization so that I could tolerate a larger error signal RMS and still be linear, in hopes that this would be "good enough" to get a spectrum. Nothing I tried worked. Frank came down and played for a bit. Nothing he tried worked.

PMC in P-Pol (from sweep):

  • FSR = 714 MHz
  • Finesse = 103 +/- 28
  • Cavity BW = 6.9 +/- 1.8 MHz

Using LB1005 to lock the PMC PZT to the Emcore diode in P pol

  • When I sweep the PZT looking at the error signal, it appears rather thick (noisy)
  • The noise is mostly at 1.5 MHz when I sweep near resonance (Vrms ~ 30mV)
  • When I am off resonance the 1.5 MHz signal is ~ 2 mVrms
  • I'm assuming this is frequency noise, though am slightly confused by how clear of a frequency it is (I would expect the noise around there to be not-so-featureful)
    • Maybe some RF pickup in the LD/Driver?
  • The LPF for the error signal is 1.9 MHz, which is very close to the frequency I'm seeing, could this be resonating?
  • Locking the PMC PZT to the Emcore LD, I tweaked the gain and PI crossover frequencies, but was unable to get the error signal visibly suppressed below it's peak to peak value.
  • Next up...

We tried the existing "Bias-T" (not actually a bias T?) serial RC circuit board which Frank procured to actuate directly on laser current (initially with a 27 Ohm resistor and a 30nF cap in it, swapped later to a 1uF).

We locked the PMC to the laser diode loosely, then started trying to feedback onto the laser current directly with a pickoff of the error signal using an SR560 (and some RF attentuators to limit the current, intially XX was 60 dB)

BiasT.png:


 N.B. I don't know what the diode impedance or the "choke" inductances are, and the initial resistance in the RC was 27 Ohms (so ~66 kHz pole without thinking about the diode impedance)

  • When we turned the SR560 on with a single pole at high frequency, we were able to see changes in the transmission and reflection (while the PMC was locked to the laser)
  • We saw the noise go WAY up when we flipped the sign, so convinced ourselves that the sign was right
  • We saw almost no change in the transmission (or error signal) until a high enough gain setting that we saw some visible oscillation

Frank had some 1uF SM caps so we swapped the 30nF cap for a 1uF cap, moving the RC pole to ~2 kHz. I don't know how OK this is since we didn't change the "choke" inductor, but we did not appear to cook the diode (or the diode driver).

We repeated the above, trying to lock using laser current with various different SR560 gains and pole locations (so we have a plateau loop to handle high frequency)

  • For each pole frequency location, we were able to turn up the gain until the transmission and reflection started to visibly get worse (oscillating)
  • We never visibly saw the transmission/reflection/error signal get significantly better.

Unless I can get these laser diodes locked well enough that I can take a transfer function and trust its linearity, I don't see how to figure out the transfer function of current to frequency on the diode.

After this semi failure, Frank wanted to see what happened when we tried to lock using the ITC510 modulation input.

Having anything plugged into the modulation input showed some strange hopping behavior. It looked like something was jumping rail to rail, even with a 50 ohm terminator on it.

  • The ITC510 says this is a bipolar input (+/-10V) with a small signal 3dB bandwidth of 200 kHz (I'm assuming this means that's where the first pole is)
  • The input impedance is 10kOhms

I tried using the Newport 6000 Driver for the Emcore diode (it is what we used before) but I kept getting a Voltage compliance error (which shutdown the driver) when I turned the current on, whether or not I increased the current setpoint from zero.

I changed ITC510 drivers and saw the same behavior as above.

Tomorrow Rana will come beat me with sticks until I learn how to make good servos.

 

  196   Tue Jun 14 23:37:51 2011 FrankCryostatVacuumfeedthroughs

after counting all connections we might want i think we want more than 19pins per connector, especially as the standard 19pin connectors already require KF40 flanges and are ~$600 per feedthrough.

So last week i checked what cryo companies use for their cryostats and i found that most of them including well known ones as Oxford use Fischer connectors for RT vacuum connectors.
Also other large scale experiments which operate at cryogenic temperatures like accelerators (Fermilab, Cern, Desy) use those connectors so i think they are a good alternative.
The advantage of them is that they com in a smaller package so that a 27pin connector fits into a KF25. I double checked that, see CAD images below (24pin connector shown, top and bottom in a KF25 flange).

The 27pin version is pretty common, but currently not in stock. However the 24pin (and lower pin counts) version is used by the cryo companies and in stock at several places.
I also requested a quote from the local Fischer distributor, where the connectors are only half the price. The total for a 24pin feedthrough is $200 including air-side connector.
So i would go for those if no one has a better idea. We can still decide if we want an in-vacuum connector as well attached from the inside, but i won't go for it as we don't need an additional connector there.
Our cables are only about 20" long and if we solder them we have one place where we can have loose connections less. We will have some kind of breakout panel at the 77K stage too.

fischer_feedthrough_24pin.JPG

fischer_feedthrough_24pin_bottom.JPG

 

  195   Tue Jun 14 14:14:58 2011 DmassDailyProgressPlotsSummary of Measurements so Far

This is a summary of the lab work I did May 30 -> June 10 (existing entries linked to):

  1. Calculated Mode Matching between fiber output coupler and PMC (elog:162)
  2. Razorblade scans of PMC input beam mode (elog:164)
  3. PMC endcap was misalinged (elog:167)
  4. Reglued PMC Endcap (elog:171)
  5. Made an error signal (elog:173)
  6. Used an SR560 to lock the Covega laser to the PMC (elog:174), success mild.
  7. Used the New Focus LB1005 to lock the PMC PZT to the laser (elog:175)
  8. Made plots of error signals (free running and in loop) (elog:189) for the Covega and Emcore diodes
  9. Made RIN plots for the Covega and Emcore diodes (elog:189)
  10. Discovered error signal is going peak to peak, so cannot be sure about results in item 8 (elog:191)
  11. Tried to lock the Emcore diode to the PMC using the LB1005 (elog:191). Barely successful.

These are the measurements I made which are useful even though the error signal RMS was way too large for me to be sure that I understand what it means.

  • Calibrated PMC Open Loop PZT Response
    • This may be total trash as the RMS too high to really expect it to be linear. I attached the plot below
    • For one of the two measurements I did, the known DC response agreed with the low frequency response of the transfer function.
    • I think this means that I am "close" to being in the linear range of the PZT
    • If I can make a better loop to get the error signal RMS down, or I can extend the linear range a bit, I might be able to get something useful
    • I am thinking of using the the cavity in P (lower finesse) to increase the linear range of the error signal so that I can actually make a spectrum I can trust.
  • Emcore/Covega RIN
  • Cavity Finesse (S/P)
  • eLIGO (D980454-D) RFPD Transfer Functions (DC path and RF path)
Attachment 1: PMCTFs.pdf
PMCTFs.pdf
  194   Mon Jun 13 23:52:45 2011 FrankSummaryNoise Budgetresidual gas limitation

after talking to Rana this afternoon about which gauge to buy i checked again if the pressure of 10-5 Torr would be enough to not be limited by residual gas scattering.
It turned out that the simple number game i did last year was OK, but here a plot taking the full model (see P940008) into account. Matlab-file is in the svn (DYM: path/name?).

CryoCav_H2_125K.png

I only used hydrogen as a reference here. I'm too lazy to calculate the molecular polarizability for all the different gases for 1550nm (you need the refractive index for the gas at low temperatures to do so and then it depends on pressure and temperature).
Other gases will usually contribute more but i don't know how far i have to go as most of the usual ones taken into account (methane, CO2, Krypton, Xenon, ...) will be liquid already due to our low temperature so i think it doesn't make much sense.

  193   Thu Jun 9 23:34:17 2011 DmassLaserPlotsDiode Frequency Noises

I have updated the plot of the Emcore Noise Budget with some additional measurements.

I also noticed that my FREQUENCY NOISE CALIBRATION WAS IN FACT TOO HIGH BY A FACTOR OF 10!

 

This makes me less concerned about the linewidth vs RMS argument, which seemed to be "crazy" and now is around the right OOM...

  • If the noise is in fact "white" out to the laser's relaxation oscillation, then we get:
    • ~0.2kHz/rtHz out to ~1GHz (with some peak which will contribute to RMS)
    • so an RMS on a 1us timescale gives: ~sqrt( (0.15kHz/rtHz)^2 x 1 GHz ) = 4MHz (they quote ~500 kHz to 1 MHz)
    • Before it seemed that the rms was 40MHz, which seems *really* crazy when compared to the quoted ~1MHz linewidth

And also: The low pass is something very steep, I couldn't guess it's poles to undo it's filtering in the spectrum, I am guessing they are spread around a bit. Just ignore the frequency noise above 2MHz from my measurements.

Attachment 1: EmcoreNoiseBud.pdf
EmcoreNoiseBud.pdf
  192   Thu Jun 9 22:43:42 2011 FrankLaserPlotsDiode Frequency Noises

I guess you are still using the same driver. You should swap the drivers to see if the one you are currently using is broken. Try the other Thorlabs or the Newport. I have no idea why shorting the input makes everything terrible.

Quote:

Quote:

Here are plots of the respective laser diode frequency noise, with the correct calibration this time (measurement details later):

  • The PMC pole is seen around 2.8 MHz in the Emcore plot.
  • The Covega noise is assumed to be white above 30 kHz (up to its relaxation oscillation).

I am trying to figure out some quantitative answer to the following question:

"Can we lock these noisy laser diodes to our Cryo cavity by locking a PMC to the laser, then putting in a (fast) small additive offset into the error point of the laser (and suffering some small AM)"

I am currently a bit confused on the control topology, a few things seem like they might make sense:

  1. Lock the PMC to the laser
  2. Get a noisy transmission through the PMC
  3. Sweep laser current slowly, following with the PMC
  4. Pray the noise (crazy AM etc) is low enough where we can lock the laser to the reference cavity while we sweep it
  5. Have the PMC (slowly - 2kHz) follow the new low noise narrow linewidth laser which is locked to the reference cavity

OR (I think it's what I've detailed below, if its the first one, I don't know how to model "lockability")

  1. Lock the laser to the PMC
  2. Sweep PMC length
  3. When we go through the reference cavity resonance, add a FAST offset straight into the error point of the laser
  4. The PMC linewidth will be high, so we will probably drift within the resonance for some time
  5. Hand off the slow actuation to the PMC PZT as we start to drift out of resonance
  6. Final state is: Laser locked to REFCAV, PMC locked to laser

 Here is a plot for figuring out if the diodes are OK...

NOTES:

  • The error signal itself goes peak to peak in its "most locked" state. It is so noisy that I cannot say for certain that I don't go outside the locally monotonic range of the error signal.
  • I measured the error signal dark noise floor and will add it to this measurement.
  • The high frequency dip is the 1.9 MHz low pass filter after demodulation - I can't see the PMC pole in this measurement.

ATTEMPT TO LOCK THE DIODE LASER TO THE PMC:

  • Was woefully unsuccessful.
  • The coefficient is ~0.1V/GHz at the input of the ITC510 (or 630 MHz/mA at the diode current input).
  • When sweeping the PMC PZT to see resonance, I was not able to make out any real error signal or reflection dip, unless I had nothing plugged into the modulation input of the ITC510.
  • Using the LB1005, I was able to get a slightly worse "lock" than when modulating on the PMC PZT.
    • Transients from flipping the control output switch from OFF to ON would make the current of the ITC510 hit its protective rail (310 mA)
    • I changed P-I pole location, gain settings, integrator switch location, but was never able to get the integrator to not rail the ITC 510
    • I WAS able to get unstable transmission (and I checked my feedback sign) at certain settings, but the reflection dip was exceedingly low, and the error signal never visible shrunk (railed peak to peak the whole time)
  • Maybe I would have more success with the Covega, since it's current to frequency is 90 MHz/mA, a factor of 8 lower than the Emcore. I doubt it will make THAT much of a difference.

 

  191   Thu Jun 9 14:27:15 2011 DmassLaserPlotsDiode Frequency Noises

Quote:

Here are plots of the respective laser diode frequency noise, with the correct calibration this time (measurement details later):

  • The PMC pole is seen around 2.8 MHz in the Emcore plot.
  • The Covega noise is assumed to be white above 30 kHz (up to its relaxation oscillation).

I am trying to figure out some quantitative answer to the following question:

"Can we lock these noisy laser diodes to our Cryo cavity by locking a PMC to the laser, then putting in a (fast) small additive offset into the error point of the laser (and suffering some small AM)"

I am currently a bit confused on the control topology, a few things seem like they might make sense:

  1. Lock the PMC to the laser
  2. Get a noisy transmission through the PMC
  3. Sweep laser current slowly, following with the PMC
  4. Pray the noise (crazy AM etc) is low enough where we can lock the laser to the reference cavity while we sweep it
  5. Have the PMC (slowly - 2kHz) follow the new low noise narrow linewidth laser which is locked to the reference cavity

OR (I think it's what I've detailed below, if its the first one, I don't know how to model "lockability")

  1. Lock the laser to the PMC
  2. Sweep PMC length
  3. When we go through the reference cavity resonance, add a FAST offset straight into the error point of the laser
  4. The PMC linewidth will be high, so we will probably drift within the resonance for some time
  5. Hand off the slow actuation to the PMC PZT as we start to drift out of resonance
  6. Final state is: Laser locked to REFCAV, PMC locked to laser

 Here is a plot for figuring out if the diodes are OK...

NOTES:

  • The error signal itself goes peak to peak in its "most locked" state. It is so noisy that I cannot say for certain that I don't go outside the locally monotonic range of the error signal.
  • I measured the error signal dark noise floor and will add it to this measurement.
  • The high frequency dip is the 1.9 MHz low pass filter after demodulation - I can't see the PMC pole in this measurement.

ATTEMPT TO LOCK THE DIODE LASER TO THE PMC:

  • Was woefully unsuccessful.
  • The coefficient is ~0.1V/GHz at the input of the ITC510 (or 630 MHz/mA at the diode current input).
  • When sweeping the PMC PZT to see resonance, I was not able to make out any real error signal or reflection dip, unless I had nothing plugged into the modulation input of the ITC510.
  • Using the LB1005, I was able to get a slightly worse "lock" than when modulating on the PMC PZT.
    • Transients from flipping the control output switch from OFF to ON would make the current of the ITC510 hit its protective rail (310 mA)
    • I changed P-I pole location, gain settings, integrator switch location, but was never able to get the integrator to not rail the ITC 510
    • I WAS able to get unstable transmission (and I checked my feedback sign) at certain settings, but the reflection dip was exceedingly low, and the error signal never visible shrunk (railed peak to peak the whole time)
  • Maybe I would have more success with the Covega, since it's current to frequency is 90 MHz/mA, a factor of 8 lower than the Emcore. I doubt it will make THAT much of a difference.
Attachment 1: EmcoreNoiseBud.pdf
EmcoreNoiseBud.pdf
  190   Thu Jun 9 14:18:18 2011 DmassDailyProgressSchematicsTimeline PDF

Timeline!

Attachment 1: TimeLine.pdf
TimeLine.pdf TimeLine.pdf TimeLine.pdf
  189   Thu Jun 9 02:44:15 2011 DmassLaserPlotsDiode Frequency Noises

Here are plots of the respective laser diode frequency noise, with the correct calibration this time (measurement details later):

  • The PMC pole is seen around 2.8 MHz in the Emcore plot.
  • The Covega noise is assumed to be white above 30 kHz (up to its relaxation oscillation).

I am trying to figure out some quantitative answer to the following question:

"Can we lock these noisy laser diodes to our Cryo cavity by locking a PMC to the laser, then putting in a (fast) small additive offset into the error point of the laser (and suffering some small AM)"

I am currently a bit confused on the control topology, a few things seem like they might make sense:

  1. Lock the PMC to the laser
  2. Get a noisy transmission through the PMC
  3. Sweep laser current slowly, following with the PMC
  4. Pray the noise (crazy AM etc) is low enough where we can lock the laser to the reference cavity while we sweep it
  5. Have the PMC (slowly - 2kHz) follow the new low noise narrow linewidth laser which is locked to the reference cavity

OR (I think it's what I've detailed below, if its the first one, I don't know how to model "lockability")

  1. Lock the laser to the PMC
  2. Sweep PMC length
  3. When we go through the reference cavity resonance, add a FAST offset straight into the error point of the laser
  4. The PMC linewidth will be high, so we will probably drift within the resonance for some time
  5. Hand off the slow actuation to the PMC PZT as we start to drift out of resonance
  6. Final state is: Laser locked to REFCAV, PMC locked to laser
Attachment 1: EmcFreqNoise.pdf
EmcFreqNoise.pdf
Attachment 2: CovFreqNoiseHz.pdf
CovFreqNoiseHz.pdf
  188   Mon Jun 6 00:16:41 2011 DmassDailyProgressLab WorkPLOTS!

Here are some plots of the measurements I've been making:

  1.  Reflection, Transmission, Error Signal, and Drive signal while sweeping the 00 PMC resonance
  2.  Error Signal vs Drive signal
    • This gives Verr/Vdrive calibration
    • Also have 10MHz/Vdrive from sweeping the PZT by hand
    • Have error signal calibration (Hz/V)
  3.  Error Signal Spectrum
  4. Calibrated Error Signal Spectrum
  5. PMC PZT Open Loop Transfer Function
    • I will add the diagram of the setup used to make this measurement
    • Above 500 or so Hz (my UGF) this is just the open loop transfer function of the PMC
    • I assumed that the 12kHz peak was its first resonance of the PMC, which means that below that the response is flat in phase and magnitude
    • The error signal calibration drifts between locks (I am inferring this since the gain at which the loop goes unstable drift changes between locks)
    • Since I know the PZT calibration (10MHz/Vdrive), I calibrated the OLTF to agree below the 12kHz resonance.
  6.  The Open Loop Transfer Function of the LB1005 driver
    • This was taken with the integrator disengaged
    • Below about 1 kHz phase is flat at 90 degrees
  7. Open and Closed loop error signal
    • Open loop error signal (control signal) calculated from measured open loop transfer function
  8. Covega RIN
    • I think I hit the PD noise floor around 1 kHz
  9. Emcore RIN
    • I think I hit the PD noise floor around 1 kHz
    • Both diodes showed nonstationarity (glitches) in the low frequency parts of the spectra while I was taking them. On the ~10 second time scale I would get large increases in the noise below around 50 Hz

Emcore frequency noise measurements will have to wait... I kept the fiber coupler in the same place and removed the Covega fiber from it, and then put the Emcore fiber in. The beam was misaligned going into the EOM, so I think I probably have to realign the input beam into the EOM. Yay!

Attachment 1: PMCSweep.pdf
PMCSweep.pdf
Attachment 2: ErrCal.pdf
ErrCal.pdf
Attachment 3: ErrorSpecV.pdf
ErrorSpecV.pdf
Attachment 4: ErrorSpecHz.pdf
ErrorSpecHz.pdf
Attachment 5: PZTOLTF.pdf
PZTOLTF.pdf
Attachment 6: LB1005OLTF.pdf
LB1005OLTF.pdf
Attachment 7: ErrandCtrl.pdf
ErrandCtrl.pdf
Attachment 8: CovegaRIN.pdf
CovegaRIN.pdf
Attachment 9: EmcoreRIN.pdf
EmcoreRIN.pdf
  187   Sun Jun 5 20:12:48 2011 FrankCryostatCryocryo glue/epoxy

started making a list of stuff people use for cryogenic applications to see which one we might want.
Most people don't care about outgassing as they don't have optics inside, but we do. We don't wanna ruin our cavity.
So here a list of stuff i've found: @Warren: please add whatever you guys used so far.

Stycast Epoxy 2850-FT, Catalyst 9

  • highly versatile, nonconductive epoxy resin system for cryogenic use
  • Primary use for Stycast is for vacuum feedthroughs or permanent thermal anchors
  • Excellent thermal cycle reliability
  • for permanent attachment
  • excellent low temperature properties
  • poor electrical conductor
  • low cure shrinkage

price: $200 / 20 packets, 2 g each  link
            $100 / 1kg link
            $100 / 3lb link  +  $15 Catalyst 9 link
           

 

VGE-7031 Varnish (formerly GE 7031 Varnish, also known as IMI 7031)

  • Low temperature glue for thermally anchoring wires at cryogenic temperatures
  • Does not outgas after baking and can be used in vacuum down to 10-9 Torr
  • Can be air-dried or baked
  • widely used to mount temperature sensors
  • Nonpermanent attachment
  • easy to apply and remove
  • can be easily removed with methanol

price:  $95 / pint link
           15 Euro / 20ml link
            $72 / 100ml link

http://www.ib.cnea.gov.ar/~experim1/Barniz.htm

 

 

Stycast 1266 clear epoxy

  • 2-part clear epoxy - Used for encapsulating electrical connectors and other assemblies used at cryogenic temperatures
  • High thermal conductivity, low susceptibility
  • Very low viscosity when mixed
  • Thermal contraction of 1,1% at 4K
  • Room temperature curing

price: $160 / 1kg (part A+B) link
           $70 / 25oz part A link + $50 / 7oz part B link

 

Other epoxies from Master Bond :

http://techbriefs.firstlightera.com/EN/Microsites/1/Master+Bond+Inc/CryogenicEpoxyAdhesivesSealantsandCoatings-forUseatCryogenicTemperatures

http://www.masterbond.com/sg/masterbond_cryosg.pdf

http://www.eso.org/sci/facilities/develop/detectors/optdet/manuals/CleanRoom/msreport/

 

Probably better for Silicon:

lartpc-docdb.fnal.gov/0002/000249/001/fermilab-tm-2366-a.pdf

Outgassing database from NASA: http://outgassing.nasa.gov/cgi/uncgi/search/search_html.sh

  186   Sun Jun 5 17:14:57 2011 FrankLab InfrastructureGeneralgas cylinder rack

we have to ask PMA to install another gas cylinder rack in the lab for the exchange gas cylinder (Helium(?)).
I think a good spot would be next to the hydrogen bottle. It's close enough to our experiment and there should be enough space
Anywhere next to the table is complicated because of the electrical cabling on the wall and we do not have a lot of space around the table.

Price for Helium, Research Grade, 99.9999%, large 300CF bottle is ~$80
(standard industrial/medical grade Helium is ~$65, but purity is less and contains moisture and nitrogen which will freeze or becomes liquid which we don't want to happen)

We don't need a lot of gas so we could also go for very small bottles which we can mount almost anywhere.
But i guess it's more difficult to get research grade in smaller bottles.

Any suggestions?

  185   Sun Jun 5 14:02:20 2011 FrankThings to BuyCryoin-vacuum cryo connectors

i thought a bit about how we connect the individual sensors and heaters attached to our cavities after i talked to David on Friday and came up with a list of arguments for how to do it:

I think we want be able to connect sensors and heaters individualy inside the experimental chamber, instead of having one big connector there or running cables all the way up to the feedthrough connector. I'm thinking of having a small breakout panel attached to the cold plate with 3 or 4 small connectors per cavity each. Feedthrough connectors can still be high pin count connectors as we prepare them only once. Here i would still go for two 19-pin connectors or one 32-pin connector. I'm still undecided what's better. One big one is slightly cheaper, the smaller connectors are easier to get. We need a breakout panel in the rack anyways to connect to the other equipment.

Reasons for individual low pin count connectors on cold plate:

  • This can be quite handy as we can e.g have more sensors in different locations or different type of sensors and we can easily swap them without soldering.
  • We can prepare every sensor / heater with the connector and get it cleaned/baked before we install it and we don't have to deal with that in our experiment.
  • We can remove a single cavity including frame from the cold plate to work on it somewhere else
  • We can assemble everything somewhere else before putting it together
  • Easy cabling in experimental chamber as we can run cables from individual sensors/ heaters independent.
  • ..(?)

Reasons against it:

  • more connectors means (slightly) higher cost
  • more complex when having 3 temp sensors and heaters connected to one connector. Running cables more difficult. Cabling has to be planned more ahead.
  • more space required on cold plate for multiple connectors (?)


I've checked what other people use these days. Almost all of them use the PEEK insert from standard LEMO connectors. You can buy just the PEEK insulator part with the pins without the metal casing.
Cryo Shops sell those, e.g. see
here. First i couldn't clearly identify them but they looked very similar to the LEMO inserts but then i found a couple of companies which e.g. sell SQUIDs  and are way more specific in their datasheets (SQUID, SQUID2, Cryo Preamp).

So i started checking where to get the parts in the US, the part number they use is EGG.1B.310.ZLL and EGG.1B.314.ZLL.
For those who don't know the numbering scheme here a short expanation: EGG=type of connector, male/female, cable, PCB etc and keying, 1B = size and series, here size 1 and B series. 310/314 = pin configuration and count, here 10 and 14 pins. The most important part is the ending, which specifies the materials used. Most of that can be found in the
LEMO catalog.

But the catalog doesn't say anything about ZLL for the part number. Finally found it on some LEMO UK page: description

Part Number Component Description
 EGA.0B.200.LC Housing
 EGG.0B.304.ZLL Insert with contacts
 ERA.0B.220.BN Earthing crown
 ERA.0B.230.LN Press-fit ring
 GBA.0S.250.FN Washer
 GEA.0S.240.LN Hexagonal nut

 

 

 

 

 

 

 

 

 

So ZLL is just the PEEK insert with the contacts (solder cup).  ZLN means the PEEK insert used with crimp contacts.

Now i started looking for companies in the US selling only the inserts. However we could always buy the full connector and only take the insert.

HEILIND

SAGER

POWELL

CMR

Standard connectors and inserts can also be bought from

MOUSER

NEWARK

Most of the above have 2 and 4-pin connectors of the 0B series in stock, also inserts only. We can also recycle initial LIGO LEMO connectors. There should be plenty.
I prefer the crimp version as they are easier to make and you don't have to use solder without flux or clean them a lot.

 

There is also a company selling cheapo LEMO compatible connectors, but not using PEEK as an insulator: link 

  184   Sat Jun 4 15:19:31 2011 FrankThings to BuyVacuumvacuum gauge

we need to buy some vacuum gauges to measure the pressure in the experimental chamber and the isolating chamber for the cryostat.
As we want to use exchange gas for cooling the cavity we first have to pump both chambers, refill the experimental chamber with some gas and then pump it again later after cooling.
So i think what we want is two vacuum gauges, one for each volume. The flanges on the chamber are KF25, so i would go for KF25 flanges on the gauges as well to simplify connections/adapters.

The next question is how much range do we need. I think we don't need a full range gauge going from atm to 10-9 Torr or so. They are expensive, ~$2k each, depending on brand and interfaces they are coming with, and usually they need an additional display unit to read the value. Less range should be fine for us.
So i browsed the web for a while and checked what is sold these days and i found a nice device from MKS, which is a combination of two transducers, covering atm to 10-5 Torr.
It comes with local display, RS232 port, analog output, 3 programmable relay outputs etc, everything for less than $600.  Shall we go for two of it?

901P.jpg

  183   Sat Jun 4 12:35:20 2011 FrankNotesVacuumVaccum feedthroughs

List of vendors having circular multi-pin vacuum feedthroughs. D-SUB connectors won't work as we can't get enough pins in small flanges.
As feedthroughs are very expensive we probably want to go for one or two total.

http://www2.ceramtec.com/catalog07/view_subsubsection.cfm?SectionID=36&SubsectionID=237&SubSubsectionID=215

http://www2.ceramtec.com/catalog07/view_subsubsection.cfm?SectionID=35&SubsectionID=284&SubSubsectionID=262

http://accuglassproducts.com/product.php?productid=16698&cat=336&page=1

www.lesker.com/NewWeb/Literature/PDF/07-073_SubminiatureC-TypeFeedthroughs.pdf

www.testbourne.com/im/instruments/ceramaseal/products/Circular-Type-MIL-C-26482.pdf

www.testbourne.com/im/instruments/ceramaseal/products/Plugs.pdf

http://www.solidsealing.com/parts/accessories.cfm?pid=1840

www.mdcvacuum.co.uk/resources/downloads/pdfs/sec6.1.7.pdf

http://accuglassproducts.com/home.php?cat=320

http://www.mpfpi.com/Multipin%20Table%20of%20Contents.html

http://scbshop.de/epages/Budzylek.sf/en_GB/?ObjectPath=/Shops/Budzylek/Categories/Assortment/Cryogenics/Connectors

 

Vacuum channel/pin count:

2 sensors per cavity, 4-wire sensing each   = 8
1 heater per cavity, 2 wires   = 2

Total for two cavities: 20

one common heater or two additional heaters to bring everything close to 124K before fine tuning. This increases the sensitivity and reduces the noise for the final radiative heater for cavity tuning : 2 or 4

additional sensors? one for the cold plate would be good, but does not have to be in the experimental chamber. 2 or 4 pins, depending on how accurate we wanna measure that.

 

So i get a total of something between 20 and 28 depending on what we wanna do, so getting 2x 19 pins or one 32pin connector would be good.
I tend to go for 2x 19 pin as we get even more pins total and we can make everything symmetric and have one connector per cavity, so we can swap things for debugging etc.
19 pin mil-spec connectors are also more common and can be bought from different vendors, but 32 pin connectors are rare.

  182   Sat Jun 4 12:32:41 2011 FrankNotesPurchasesvaccum pump station

- personal notes-

Vacuum pump station

HiCube 80 Eco, DN 63 CF-F, with DCU 002    (PM S03 556http://www.pfeiffer-vacuum.com/productPdfs/PMS03556.en.pdf

online catalog: http://www.pfeiffer-vacuum.com/products/pumping-stations/turbo-pumping-stations/hicube-eco/onlinecatalog.action?detailPdoId=5165

All available stations: http://www.pfeiffer-vacuum.com/products/pumping-stations/turbo-pumping-stations/hicube-eco/onlinecatalog.action

Promo: http://www.pfeiffer-vacuum.com/products/pumping-stations/turbo-pumping-stations/sales-promotion/container!download?referer=1248&download=/filepool/File/Literatur_Pdf/Promotions/PT0131PE_HiCube_Eco_Promotion.pdf

 

additional options we might wanna have:

http://www.pfeiffer-vacuum.com/products/turbopumps/hipace-tm-series/hipace-tm-10-800/hipace-80/onlinecatalog.action?detailPdoId=4511

  • TVV 001 fore-vacuum safety valve (PMZ01206)
  • Splinter screen for Turbopumps, DN 63 CF-F flange (PM016312)
  • Vent Valve, G-1/8 connection w/ M8 plugin cable for HiPace 80  (PMZ01290)

 

Requested a quote on Friday

  181   Fri Jun 3 23:20:08 2011 DmassDailyProgressPMCPZT Voltage Coefficient

I measured the PMC PZT coefficient using a HV driver.

I was able to go 1 FSR (714 MHz) in 70 Volts

[Hz/V] = 10 MHz/V

  180   Fri Jun 3 02:28:41 2011 DmassDailyProgressLab WorkKoji is the man.

The following is the to-do list of measurements I want for the diodes:

  • Diode RIN
  • Diode frequency noise
  • PMC PZT Transfer Function (m/V or Hz/V)
  • Laser diode current transfer function
    • Thorlabs ITC510 transfer function to frequency (Hz/V)
    • Bias-T transfer function (I/V)

Made the following today:

  • PZT Transfer function
  • Error signal with loop closed
  • Control signal with loop closed (this will give me frequency noise within the lock bandwidth)
  • Transfer function through LB1005 with integrator engaged
    • Transfer function of cable for calibration (mostly because I don't know what the splitter does at 10 Hz)

I tried to take a transfer function from the modulation input of the ITC510 to the error signal with the PMC locked, but for some reason, the error signal of the PMC started to oscillate whenever I tried to set up the measurement. It happened any time I plugged anything with a 50 Ohm resistance into the modulation input, including a 50 Ohm terminator. This confuses me.

  179   Thu Jun 2 19:31:29 2011 FrankUpdatePurchasesitems to finalize

Below a list of items i was working on in the past which i think we can finalize but want to double check a last time.

RA: Go, go, go

  • Photodetection
    • photodiodes will be standard InGaAs photodiodes (2mm PerkinElmer, C30642GH). QE is excellent at 1550nm and we also know that we can build reasonable detectors with them. So no need to go for something different.
    • Copy new electronics design from Gyro, make a couple of small changes, mainly footprints of some parts. We need a minimum of 6 boards, 2 times PMC + RefCav and one for the beat + 1 spare.
      The PSL will need some as well (2x RefCav, beat, spare) so we can order the boards together. 6 boards are $645. Going from 6 to 15 boards is $50 more.
  • Servo Box
    • Zach will finalize the board this week. We should build one or two of the old ones with some modifications for a start and then replace them later by the new design once we know the new one is working
    • Parts for current (old) design exist.
  • Vacuum pump
    • would go for a "Pfeiffer Vacuum HiCube 80 Eco, DN 63 CF-F, MVP 015 with DCU 002", a slightly modified version of the one Alastair bought about a year ago. Will be about $5k.
    • http://www.pfeiffer-vacuum.com
  • Power supplies
    • would go for Sorensen XEL-P series. They are linear regulated and super low noise and come in a 1/4 rack style with rack adapter. So we can mount a dual 30V/3A supply and two additional 250V supplies in the rack if we want. Have front and rear terminals, can be locked so that we can bump the knobs and nothing happens :-) Do also have remote sensing, control etc.
    • waiting for a last quote. Items are in stock except for the rack mount kit.
  • Temp sensors / Cryo controller
    • will go for PT sensors in the beginning (see yesterdays entry). The cryo controller (SRS CTC 100) has 4 independent inputs, 6 analog outputs, 2 of them with power drivers to directly drive heaters. Any input can servo to any output, different TFs can be saved for different temp ranges, inputs can be switched depending on temperature, etc. So we can use one input to cool down to 124K and once reached a specified temp window we can switch to a different sensor and/or output. So heating/cooling to 124K seems to be easy as everything can be programmed within the controller. It can also run scripts for temp cycling to measure thermal expansion coefficient etc. It also has digital outputs, so we can e.g. start automated pumping the exchange gas once reached 120K or something like that. Datalogging function is nice and also has standard interfaces to network, computer etc.
  • 1550nm CCD
    • ordered one of the 1460-1625nm Near-Infrared Cameras 3 weeks ago, but they say they don't have one even if the system says one in stock. Have to poke them.
    • should think about getting a beam analyzer like the Wincam we already have. Might be nice to have for the future.
  • Optomechanics/Optics
    • more optomechanics / optics has been ordered

 

  178   Thu Jun 2 15:53:04 2011 FrankMiscGeneralCryo controller manual

got a copy of the manual for the SRS cryo controller. The device will be delivered in about 2-3 weeks. The have some delay in the production.
A copy can also be found on the wiki.

CTC100_manual.pdf


  177   Thu Jun 2 15:37:29 2011 DmassDailyProgressLab WorkKoji is the man.

Koji came down and poked around at the LB1005 with me for a bit to try to help me understand why I saw weird clipping behavior where I would expect to see a proportional signal. I managed to trick him for a little bit into being confused by what we saw until he pointed out that the answer would have been clear if I had, in fact, thoroughly RTFM ( shameface):

  • Input range of LB1005: +/-10V
  • Input range of DIFFERENTIAL AMPLIFIER AFTER WE TRIM THE ERROR SIGNAL: +/- 300 mV
  • We saw some crazy behavior with input signals to the LB1005 which were over 300 mV
  • Funny that.

He also helped me understand what was going on with the loop and suggested some better ideas on why it might be "not so hot":

  • The loop oscillates like heck at ~12 kHz with a fairly high Q peak when I turn the UGF up past ~100 Hz
  • Koji recalls seeing ~10 kHz resonances from his earlier days where he still did things like gluing mirrors onto cavities
  • If I DO have some PZT/Endcap/Glue resonances w.r.t. the spacer starting around 12k, and some of them are high Q (~100), then at a UGF of ~100Hz, I could have some terrible instability in my servo at 12k from where the peak is crossing unity gain with some random (bad) phase.
  • I will see if I can get a tiny bit more out of the UGF by switching to a 30kHz PI crossover.
    • This should make the loop have slightly less phase at low frequency, but also lower my 12kHz resonance by a factor of ~3,
    • maybe I can lock with a ~400 Hz UGF instead of 100 Hz in this case?
  176   Thu Jun 2 00:45:24 2011 FrankUpdateSensorstemp sensor selection

i had a closer look into cryo temp sensors and what we might wanna use for the first tests.
The comparison posted some time ago (link) showed that platinum RTDs seem to be a good choice in terms of stability / repeatability of the cryo controllers.

I found the following publications which go into more details, but they agree with the above statement that platinum RTDs are a good choice for 120K:

Resolution and Accuracy of Cryogenic Temperature Measurements

A review of cryogenic thermometry and common temperature sensors

I extracted some plots from the above publications. The red line indicates ~120K, our first setpoint. The curves are for different types of sensors. Detailed description below the graphs.
They show that PT RTDs are a good choice in terms of sensitivity and temp resolution for ~120K. Other sensors are much better at lower temperatures.
The platinum RTD plot is for a 100Ohm sensor. So by using a 1k or 10k PT RTD we can get even better. The cryo controller we bought can handle higher RTD values.
Later we might wanna try some high resistive NTCs/PTCs which have much higher gain and so much less noise. The drawback is that the range is very limited and there are issues with the repeatability.

tempsensor-comp-01.jpg

 

  175   Wed Jun 1 23:19:45 2011 DmassDailyProgressLab WorkError signal smaller?

I swept the PZT directly to get the V/V calibration from the PMC to the error signal.

  • In another measurement yesterday, I measured the error signal to be 200mV/2ms.
  • While sweeping with a 10 Hz signal, it is now 60 mV over 2 ms. This is a factor of ~3 smaller than what I found yesterday.
  • The only thing I did differently is sweep the PZT directly with a normal function generator, and use the laser current to change the frequency so that I can get in range of the resonance.
  • I checked my drive signal (still 23 dBm into a splitter), the magnitude of the resonance dip (to check alignment), and the total power in (via 1st waveplate)
  • My 1st guess is that I have found a region where the diode is not very "single mode".
    • This appears to not be the case - I changed back to yesterday's measurement setup and swept through a decent current range (5 mA) around yesterdays nominal values (185 mA), and the error signal did not magically grow large
    • I FOUND A 1.3 ND FILTER IN THE BEAM PATH FROM AN EARLIER MEASUREMENT I MADE TODAY WHERE I NEEDED TO LOWER THE OPTICAL GAIN. This level of awesome should not go undocumented.

Moving on...I swept the PZT directly to get the V/V calibration from the PMC to the error signal.

  • I put the drive signal of the PMC and the error signal onto a scope to get the V/V calibration (a.k.a. DC transfer function values)...
    • [ Error Signal ] = 100mV/4ms
    • [ PMC Voltage ] = 780mV/4ms
    • Verr / Vpmc = 0.128
  • I measured the PZT capacitance: C_pzt = 246.6 nF with a BK Precision LCR Meter (Model 878A)
  • I borrowed the New Focus LB1005 Servo Controller from the 40m
    • The output impedance of the unit is nominally 50 Ohms
    • The pole of the Driver/PZT is 1 / (2 x pi x C_pzt x Z_output) = 12.9 kHz
    • The LB1005 servo has a PI crossover on the front - this should be set to as close to the 12.9 kHz pole as possible
      • 10 kHz seems to be the obvious choice here, since the next highest the LB1005 has is 30 kHz
    • The unit has a switchable integrator which is specified in "low frequency gain, in dB"
      • Koji says this is the gain between high frequency and low frequency, rather than the total servo box gain
  • I think my open loop transfer function is as follows:
    • Verr / Vpmc x HFGAIN x [ pole at 12.9k ] x [ zero at 10k ] x [ pole at low frequency ] - Integrator switched off
    • Verr / Vpmc x HFGAIN x [ pole at 12.9k ] x [ zero at 10k ] x [ pole at 0Hz ]- Integrator switched on
  • The gain knob of the unit controls the high frequency gain,
    • so for a UGF of ~1k we would want a gain setting of 1 and a PI crossover of 10 kHz
    • for a UGF of ~100k we want a gain of ~100 and a crossover of 10kHz

Marginally Better Lock

  • I tuned the LB1005 (via knobs on the back) so it could only output -40 mV to +10V (this is as close to 0-10V as it got)
  • I was able to get an OK lock of the PMC to the laser (it survived me leaving the room and tapping gently on the table with the back of a ball driver - too hard and it broke)
    • Settings: Gain = 345 // LF Gain Limit = 50 dB // Integrator On // PI crossover = 10 kHz
  • I am HIGHLY SUSPICIOUS of the LB1005's linearity - I put a 1Vpp sine wave into the input at a few different high frequencies, and was unable to recover it without pretty significant disortion. I believe this should not be the case if it works as described by the manual, and if I put the P-I frequency somewhere below the frequency I am putting into the thing, I should be able to recover it rather simply. Test this more later
  • I suspect I need more DC gain (among other things) - locking an in air PMC to a noise laser without anything more than 1/f at low frequencies seems sad.
  • The UGF seems to be around 100 Hz - The loop begins to oscillate like mad at higher frequencies (12 kHz) when I push it past this. This seems not great.

Summary

The short term goal (days) is currently diode characterization. I want to take the following measurements:

  • RIN
  • Frequency Noise
  • (maybe) Driver to Frequency Transfer Function
  • (maybe) Driver to Intensity Transfer Function

I was initially going to use the laser current modulation input on the ITC510 driver to get some reasonable bandwidth (the specs say it "works" up to 50 kHz, whatever that means to them) lock of the laser to the cavity, then look at the laser control signal (and figure out how to calibrate it) to get the laser frequency noise. Frank pointed out this might not be straightforward since the transfer function of current to frequency of the diode is not entirely simple (its sort of 1/f on a log linear scale with an unknown phase response).

The alternative to the above is locking the PMC to the laser(s), then measuring the calibrated control signal going to the PZT. Where the laser is noisier than the PMC, and within the bandwidth of the loop, this is a direct measurement of laser frequency noise. This turned out to be harder than expected, and I am not sure what exactly the culprit is...some ideas:

  • LB1005 showed some nonlinear behavior, this makes the dynamics too tricky to bother thinking about
    • I could put a 20k resistor in series with the PZT to make a ~35 Hz pole and use that to lock with something like an SR560 just as an amplifier
  • The PZT is glued onto the endcap which is glued onto back of the PMC with a few dallops of glue - this could be fairly non-rigid and have some serious resonances below the PZT body modes
  • The PZT is angled w.r.t. the cavity axis (because the back mirror is wedged), so we get axial shifting as we change the length

I will take spectra of the control (and error) signals below 100Hz or so (current bandwidth with the Newport servo)

  174   Wed Jun 1 00:50:08 2011 DmassDailyProgressLab WorkLaser Locked to PMC!!

Quote:

Making an Error Signal (try #2):

  • Currently found error signal with 30 MHz sidebands!!
    • Sidebands: generated with a SR DS345: 23 dBm at 30 MHz into a broadband EOM
    • PDs:
      • PDA50B (Slow Switchable gain Ge PD) in transmission
      • Ligo resonant PD (picture included) - D980454-D
        • Frank got his hands on a few of these for the PSL lab, and we decided to test them for our error signal. Out of three PDs, each was broken in some way.
          • The first we tried oscillated like madness
          • The second's transfer function was....odd (no obvious resonant peak).
          • The third had a burn spot on the PD itself
        • We decided to see if we could use the second one in the short term, and just suffer some loss of gain
        • The PD has a DC channel and AC channel
    • Swept the PMC PZT:
      • Drove the PZT with a Thorlabs MDT693A high voltage driver
        • 0-10V input -> 0-150V output
      • Used a SR DS335 to sweep a 5V + 8*triangle(w*t) into the MDT input
  • Measured the error signal to be 18.6 V/GHz
    • Error Signal ~ 200mV/2ms
    • Reflection Dip ~ 2xCavity Bandwidth/520us
    • Cavity Bandwidth ~ 1.4 MHz
    • [200mV / 2ms] x [520us/ 2*BW] x [BW / 1.4MHz] = 1.86 x 10^-8 V/Hz = 18.6 V/GHz
    • Picture of sweep included:
      • Purple: PZT driver input
      • Yellow: Error signal
      • Green: Transmitted power
      • Blue: Reflected power
  • Choose a reasonable loop for actuating on current!
    • 0.11V at ITC510 input gives ~1GHz shift => 9GHz/V at laser diode controller
    • 18V/GHz x SR560GAIN x 9 GHz/V = G*167
    • Start with a pole at 1 Hz and a gain of 10, see what happens!

 

 Laser Locked to PMC!

  • Using an SR560, and hand tuning the PZT offset, I was able to lock the laser to the PMC for ~30 seconds at a time.
    • First I used a 10 Hz pole (checked the offset - small w.r.t. the error signal so don't care much) and slowly increased the SR560 gain to see what happened.
      • I was able to lock with a gain of 10, but the loop seemed to go unstable at higher gains
      • If my DC gain math is right, and there are no poles except the one I put there, then this should correspond to a 1Hz pole, with a UGF of ~1.7 kHz or so
    • Then I tried turning off the pole to see if there was a pole in the system I could use.
      • I could STILL lock with a gain of ~10 DC coupled through the SR560 with no filtering
      • I am suspicious, and Frank said previously that he was able to lock the (linear) cavity for short stretches just from absorption
    • The above was something silly. I had the control signal plugged into the wrong input on the ITC510. I do not know what the heck the dynamics were.
    • I plugged the right things into the right places, and was more successful.
      • With a 1 Hz pole and a gain of 20, I was able to get a definite lock.
      • I looked at the error signal and the transmission on a scope while playing with the loop and did a few sanity checks for myself
        • When I changed the sign of the feedback (using the INV button on the 560), I was not able to get transmission through the PMC when when I swept the PMC's PZT
        • I started (slowly, manually) sweeping the laser current, and when the PMC started to resonate / transmit, slow changes to the "current' knob would no longer change the current, until (what I think is) lock broke and then it would jump by ~10 mA. This is consistent with what I would expect from "closed loop" behavior if I started trying to add in an offset after the control point

I think the lock isn't tight enough to take transfer functions directly without some truly terrible happening to myself...so:

NEXT UP: Transfer function characterization of individual parts!

  173   Tue May 31 22:47:57 2011 DmassDailyProgressLab WorkError signal found!

Making an Error Signal (try #2):

  • Currently found error signal with 30 MHz sidebands!!
    • Sidebands: generated with a SR DS345: 23 dBm at 30 MHz into a broadband EOM
    • PDs:
      • PDA50B (Slow Switchable gain Ge PD) in transmission
      • Ligo resonant PD (picture included) - D980454-D
        • Frank got his hands on a few of these for the PSL lab, and we decided to test them for our error signal. Out of three PDs, each was broken in some way.
          • The first we tried oscillated like madness
          • The second's transfer function was....odd (no obvious resonant peak).
          • The third had a burn spot on the PD itself
        • We decided to see if we could use the second one in the short term, and just suffer some loss of gain
        • The PD has a DC channel and AC channel
    • Swept the PMC PZT:
      • Drove the PZT with a Thorlabs MDT693A high voltage driver
        • 0-10V input -> 0-150V output
      • Used a SR DS335 to sweep a 5V + 8*triangle(w*t) into the MDT input
  • Measured the error signal to be 18.6 V/GHz
    • Error Signal ~ 200mV/2ms
    • Reflection Dip ~ 2xCavity Bandwidth/520us
    • Cavity Bandwidth ~ 1.4 MHz
    • [200mV / 2ms] x [520us/ 2*BW] x [BW / 1.4MHz] = 1.86 x 10^-8 V/Hz = 18.6 V/GHz
    • Picture of sweep included:
      • Purple: PZT driver input
      • Yellow: Error signal
      • Green: Transmitted power
      • Blue: Reflected power
  • Choose a reasonable loop for actuating on current!
    • 0.11V at ITC510 input gives ~1GHz shift => 9GHz/V at laser diode controller
    • 18V/GHz x SR560GAIN x 9 GHz/V = G*167
    • Start with a pole at 1 Hz and a gain of 10, see what happens!

 

Attachment 1: IMG_0479.JPG
IMG_0479.JPG
Attachment 2: IMG_0480.JPG
IMG_0480.JPG
Attachment 3: LockingSetup.png
LockingSetup.png
  172   Mon May 30 16:50:24 2011 DmassDailyProgressLab WorkOops!

I misread the Covega datasheet (elog:67) and accidentally set the Thermistor resistance to 10K for ~5 minutes. The diode was not very stable at this temperature, which is around 297K = 24 C. I think this was OK since the datasheet gives 10 C as its absolute minimum rating, and 25C is still in its "operating range".

Preferred value is 7.16k ~ 30 C

 

  171   Sat May 28 17:16:42 2011 DmassDailyProgressLab WorkPMC Aligned

ALIGNMENT

I got an email with rough guidelines of "what is sort of ok" and "what is totally stupid" in terms of taking the back mirror off the PMC, so I did the following:

Changing the PMC Endcap:

  • Tapped the endcap of the PMC with a rubber mallet until the epoxy holding it on snapped
    • The glue could still hold the endcap in place via friction after I snapped it off the endcap, so I could wiggle the endcap around in place
  • I aligned a 532 nm beam to be coaxial with the 1550 nm input beam with 2 green steering mirrors ("borrowed" from doubling noise setup) using an iris in the near field, and a card in the far field
  • I pushed the endcap back on to be ~ flat
  • Aligned 532nm input beam height to 4" (+/- 0.03") and pitch to 0 deg (+/- 0.1 deg)
  • Aligned input beam yaw to ~0.5 deg of the (desired) cavity axis, and displacement to who knows what (I tried to make sure the spot hit somewhere close to the center of the back mirror)
  • I let the (lower power green) spots hit the wall in transmission and reflection (~factor of 3 different distances, so can get pitch and yaw misalignment from this)
  • I tapped gently on the (now semi free) endcap to tilt the back mirror, trying to make the 1st round trip spots overlap with the direct transmission and reflection, it was fairly easy to get the first 2-3 round trip spots to overlap at ~4" (in reflection on a black card) and ~4' (the wall) simultaneously
  • I dug around in the PSL lab for a bit (and recruited Tara) looking for the glue Frank used initially, in case there was some special epoxy, I convinced myself that there was no mysterious hidden special glue and grabbed some 30 minute epoxy
  • I used the epoxy to glue the endcap of the PMC in its new happier location

 Aligning to the PMC

  • Assuming that the alignment didn't change much in the gluing, or in the setting of the glue, I put the direct reflection on a (400 kHz) PD and looked at its output with a black N white 40 MHz scope
  • I used a 0.11Vpp 0.3 Hz (sine or triangle, unsure which) wave from a function generator to sweep the laser current about 180 mA
    • This corresponds to about 11 mA of current modulation, which is slightly more than 1 FSR of the PMC
  • I saw dips in the PD power time series while sweeping laser current, but saw nothing on a card in transmission so I was suspicious
    • I checked the Covega Datasheets (elog:67) to see that the laser was supposedly single mode over this whole range (by examining the side band suppression ratio)
    • I tweaked the laser current DC value about 10 mA to 170 mA to see if this made a difference
    • I still saw the power dips on the PD (great!)
  • I marked the current position of the input beam steering mirrors with a red sharpie
  • I slowly changed the pitch and yaw of the individual mirrors, and was able to get a dip of ~ 50% in reflected power
    • The dip was a single clean dip, but rather a small inverted forest: I think the laser noise was so high that even with such a low finesse cavity the laser frequency was shifting more than the cavity linewidth in the time it took to sweep across the resonance
    • Assuming I can lock the laser to the cavity, I am going to wait to try to do the common/differential alignment until I have a lock
  • I saw flashes of what looked like a 00 mode in transmission on a viewer card

ERROR SIGNAL

Testing the EOM:

  • Aligned the (adjustable) EOM mount to the beam with a card
  • Use SR DS345 (function generator - borrowed from ATF lab, formerly in green setup) to generate 15 MHz sidebands at +13 dBm (2.83 Vpp) - this is comparable to what I used for the broadband EOM in the PSL laser, and the spec sheet seems to allow for high voltage input (Half wave voltage = 350V @ 1550nm), so this level should be fine
    • EOM is broadband (Thorlabs EO-PM-NR-C3)
    • I used to impedance matching network on the input
  • I used the HP4395 RF spectrum analyzer to measure the beat between the carrier and its sidebands with a pickoff downstream of the EOM
    • DC Level: 5.24 V (Couldn't see it on the crummy 40 MHz scope AC coupled at 100 ns time res/div)
    • I was slightly confused about what the 4395 measurement units were...at 1Hz RBW:
      • 15 MHz peak:
        • Noise Mode:(-87.3 dBm/Hz) = (17.5 uV/Hz)
        • Spectrum Mode: (-81.5 dBm)
      • White noise level:
        • Noise Mode: (-137 dBm/Hz) = ~(26nV/Hz - didn't record)
        • Spectrum Mode: (-132 dBm)
      • I don't really understand where to file "amplitude/Hz" with regards to everything else I know about PSDs, is this just RF people talk about everything in power, so implicit in the measurement is that Voltage goes as power, so they use V/Hz as a PSD?
  • The sidebands are clearly visible without a resonant circuit! Huzzah!

Making an Error Signal:

  • To figure out where to actually put the sidebands, we should think about:
    • PDs we have on hand for photodetections:
      • Thorlabs PDA10CS (17 MHz at lowest gain setting)
      • Newport 818-BB-30 (1 GHz but annoying to align to - would be ick for optimizing alignment)
    • Cavity Bandwidth (elog:150)
      • FSR = c/2L = c/(2*210mm) = 714 MHz
      • In s-pol (currently easier because of EOM placement and waveplate number)
        • Finesse = 522
        • BW = FSR/Finesse = 1.37 MHz
      • In p-pol (requires one more waveplate)
        • Finesse = 71
        • BW = FSR/Finesse = 10 MHz
  • It seems that 15 MHz is fine in the current configuration, with the PDA10CS, assuming we don't get too much phase delay that close to its maximum specified bandwidth
    • This is probably a bad assumption in reality, but we can always start here and swap out the PD later when we want to actually push the bandwidth of the loop up
  •  The PDA10CS was "totally broken"
    • Some amplifier inside is likely borked, as it gave us silly results when we swept the PMC through resonance with

 

  170   Thu May 26 14:13:28 2011 DmassDailyProgressLab WorkDaily To Do

I'm not sure if there is anything special I need to do regarding the cleanliness of the PMC spacer itself. I know that it is not hermetic by inspection, but recall *certain people* saying that the inside should be kept clean. If I can't really mess anything too important up I'll just tweak it in some way (find an appropriate thickness shim) and reglue it.

  169   Thu May 26 06:58:52 2011 FrankDailyProgressLab WorkDaily To Do

i've checked the specs for the mirror, my fault:
Laser Mirror: FS, pl-concave, Ø=12.7-0.1mm, te=6.35±0.1mm, wedge=30min, S1: L/10, AR(0°, 1480-1630nm)<0.25%, S2(^): L/4, r=1000mm, HR(0°,1480-1630nm)>99.98% (Low Loss), T(1510nm)~0.005%

I didn't pay attention to that as so far i always glued the concave side to the pzt, not the rear side.
So we have two options:

  1. if we exchange it by the 0.5m mirror it should be fine as it has no wedge specified: (Laser Mirror: FS, pl-concave, Ø=12.7-0.1mm, te=6.35±0.1mm, S1:L/10, S2:L/4, r2=500mm, HR(0°,1450-1600nm)>99.975% (Low Loss),T at 1530nm~0,012%)
  2. we tilt the end cap a little bit

Do ypu want to try the second one? simply remove the end cap, there are only a few little spots of epoxy. You can use a 2" mirror mount, hold the aluminum cap with it (try to push it against the back of the spacer to have as little space between the cap and the sapcer as possible) and align it properly and if everything looks good add a little bit of glue and fix it to the spacer. It's not super rigid anymore but shouldn't matter as we will replace everything in 4 weeks or so anyway.

Quote:

Quote:

Figure out alignment problem to PMC

Get online document describing of all subparts of project

Come up with schedule for action based on the (est) arrival of the bigger ticket items

 

 After several futile attempts to align the PMC (as built by Frank, see elog:155), I had Koji come in to confirm that there was something wrong with the PMC alignment and not my alignment to the PMC and tried to align a green laser pointer to the PMC. This failed and looked like a back mirror pitch misalignment (yaw looked fine). A quick estimate of misalignment, probably good to 10%

Procedure for measurement of misalignment:

  • Align input beam to 4" height in near and far field
  • Align yaw into PMC so that we hit the back mirror without clipping
  • Measure the vertical displacement between direct reflection (off inside surface of 1st optic) and the transmission of the 1st round trip beam (15" from end mirror)
    • displacement was 10/32"
    • angle = ((10/32")/15") rad = 21 mrad
    • Back mirror angle is 21 mrad/2 = 10.5 mrad = 0.6 degrees
    • axial displacement is thus 1m * 10.5 mrad = 1.05 cm
    • The substrate diameter is 1.27 cm, so the axial displacement puts the beam axis OFF THE BACK MIRROR
    • THE PMC AS BUILT (GLUED) CAN'T BE ALIGNED
    • THE BACK MIRROR OF THE PMC IS MISALIGNED BY 10 mrad = 0.6 deg

Looks like we need to pull the endcap and redo the mirror placement, possibly with some beam to help align the back mirror

N.B. Frank has told me that the inside of the PMC is still clean, so I will wait for him to get back before pulling the thing apart since I don't know what I'm doing on that front.

We also may want to consider using the faster (smaller r_curve) mirror, since this is more forgiving in terms of cavity axis displacement as a function of back mirror angular misalignment. Frank had wanted to save it for use with a (possible) AOM, but I think we should just toss it in.

If we put the 0.5m mirror (which we have on hand) into the PMC:

  • For 10.5 mrad misallignment we get .525 cm displacement, which leaves us (barely) on the substrate
  • I am assuming we can do slightly better than the pitch misalignment we already have, so I believe that we can get this the cavity axis on the coating with a little care

 

  168   Wed May 25 17:41:13 2011 DmassDailyProgressLab WorkRIO Lasers

I made an entry in the CryoWiki detailing the RIO's responses to our inquiries. It is well worth checking out.

  167   Wed May 25 15:13:16 2011 DmassDailyProgressLab WorkDaily To Do

Quote:

Figure out alignment problem to PMC

Get online document describing of all subparts of project

Come up with schedule for action based on the (est) arrival of the bigger ticket items

 

 After several futile attempts to align the PMC (as built by Frank, see elog:155), I had Koji come in to confirm that there was something wrong with the PMC alignment and not my alignment to the PMC and tried to align a green laser pointer to the PMC. This failed and looked like a back mirror pitch misalignment (yaw looked fine). A quick estimate of misalignment, probably good to 10%.

Procedure for measurement of misalignment:

  • Align input beam to 4" height in near and far field
  • Align yaw into PMC so that we hit the back mirror without clipping
  • Measure the vertical displacement between direct reflection (off inside surface of 1st optic) and the transmission of the 1st round trip beam (15" from end mirror)
    • displacement was 10/32"
    • angle = ((10/32")/15") rad = 21 mrad
    • Back mirror angle is 21 mrad/2 = 10.5 mrad = 0.6 degrees
    • axial displacement is thus 1m * 10.5 mrad = 1.05 cm
    • The substrate diameter is 1.27 cm, so the axial displacement puts the beam axis OFF THE BACK MIRROR
    • THE PMC AS BUILT (GLUED) CAN'T BE ALIGNED
    • THE BACK MIRROR OF THE PMC IS MISALIGNED BY 10 mrad = 0.6 deg

Looks like we need to pull the endcap and redo the mirror placement, possibly with some beam to help align the back mirror

N.B. Frank has told me that the inside of the PMC is still clean, so I will wait for him to get back before pulling the thing apart since I don't know what I'm doing on that front.

We also may want to consider using the faster (smaller r_curve) mirror, since this is more forgiving in terms of cavity axis displacement as a function of back mirror angular misalignment. Frank had wanted to save it for use with a (possible) AOM, but I think we should just toss it in.

If we put the 0.5m mirror (which we have on hand) into the PMC:

  • For 10.5 mrad misallignment we get .525 cm displacement, which leaves us (barely) on the substrate
  • I am assuming we can do slightly better than the pitch misalignment we already have, so I believe that we can get this the cavity axis on the coating with a little care

[edit: added a picture of the green laser pointer going into the setup]

Attachment 1: IMG_0470.JPG
IMG_0470.JPG
  166   Wed May 25 13:13:15 2011 DmassDailyProgressLab WorkDaily To Do

Figure out alignment problem to PMC

Get online document describing of all subparts of project

Come up with schedule for action based on the (est) arrival of the bigger ticket items

 

  165   Wed May 25 03:36:55 2011 DmassDailyProgressLab WorkEOM razor blade scans

Couldn't quite get aligned to the PMC. Progress:

  • I modified the mode matching lens position based on elog:164
  • I took razor blade scans of the input beam to the PMC to check the mode overlap
  • I couldn't align totally blind, so I grabbed a green laser pointer and used it as an "auxiliary beam"
  • I tried to use the overlapping visible beam to align the input optics to the axis of the PMC, but failed to find the "sweet spot" after a couple hours.
  • Since this PMC has never been locked to before, I am not positive if there us any way to tell if the alignment of the mirrors themselves are OK
  • I will consult the great benevolent Koji tomorrow, and see if he thinks I am being stupid, or if there is some more fundamental problem with the PMC itself
Attachment 1: PMCInputBeam.pdf
PMCInputBeam.pdf
  164   Tue May 24 02:27:00 2011 DmassDailyProgressLab WorkEOM razor blade scans

I made some measurements of the waist size as described in elog:162 and elog:163 using a razorblade and a photodetector.

I used a Thorlabs PDA50B (1/2 cm diameter photodiode), and put the razor blade as close to its face as I could ( < 1cm from PD case).

I did a couple things with the fit (because I was curious how the answers would differ):

  1. Used a linear approximation since I took points starting at 2-3 times the rayleigh range to get the beam waist (slope ~ lambda / (pi*w_0))
  2. Used this as my number for the waist, and found its location with fminsearch ("black" numbers on plot)
  3. Did a fminsearch over waist size and location for the four points I took ("yellow" numbers on plot)
  4. Yellow fit looks better (residuals are clearly smaller by crudely inspecting the plot)

MATLAB code used:

poss = [13 10 7 5]*0.0254;
waists = 2*[0.0505 0.043 0.0345 0.029]*1e-2;
outs = polyfit(poss,waists,1);

% linear approximation when we are out of the raleigh range is:
% w(z) ~ lambda/(pi*w0) * z

w0=1550e-9/(pi*outs(1))
zr=pi*w0^2/1550e-9;
wz = @(z,zoff) w0*sqrt(1+((z-zoff)./zr).^2);
outs1 = fminsearch(@(var) sum(abs(wz(poss,var)-waists).^2),-.3);

zr2 = @(w00) pi*w00.^2/1550e-9;
wz2 = @(w00,zoff2,z2) w00*sqrt(1+((z2-zoff2)./zr2(w00)).^2)
outs2 = fminsearch(@(var2) sum(abs(wz2(var2(1),var2(2),poss)-waists).^2),[300e-6 -.15])

posarr = linspace(-.2,.35,30);

figure(1)
clf
plot(posarr/.0254,1e6*outs(1)*posarr+outs(2),'b',...
    posarr/.0254,1e6*wz(posarr,outs1),'k',...
    posarr/.0254,1e6*wz2(outs2(1),outs2(2),posarr),'y',...
    poss/.0254,1e6*waists,'rx')
axis tight
grid

legend('Linear Fit to Data','fminsearch on xwaist',...
    'fminsearch on w0 and xwaist$\;\;\;\;$','Data from Razor Scans',...
    'Location','NorthWest')
xlabel('Position (in)')
ylabel('Waist Size (um)')
title('Razorblade Waist Measurements')
ylim([0 1050])

writ(1)={strcat('Black: w0=',num2str(floor(w0*1e6)),'$\,$um, xwaist=',num2str(floor(100*outs1/.0254)/100),'$\,$in')};
writ(2)={strcat('Yellow: w0=',num2str(floor(outs2(1)*1e6)),'$\,$um, xwaist=',num2str(floor(100*outs2(2)/.0254)/100),'$\,$in$\;\;\;\;\;\;\;\;$')};
text(4,100, writ,'HorizontalAlignment','left',...
    'BackgroundColor',[.7 .9 .7])
orient landscape

Attachment 1: RazorbladeModeScans.pdf
RazorbladeModeScans.pdf
  163   Mon May 23 13:17:44 2011 DmassDailyProgressLab WorkDaily To Do

NotLab

  • Review and compare existing quotes for the windows
    • Email American Photonics for revised quote
  • Get new quotes (send emails to coating companies)?
  • Purchase PMC from local company (~4 week delivery)
    • Update drawings to A3 from A4
    • Fill out bottom right part of drawings tolerances, units, finishing, etc
    • Add drawings to svn (some files were added as bin file type - not sure if this is a problem)
    • Email drawings (which reflect the current quote) to ASCO
    • Place requisition for PMC via techmart
  • Find list (from one of the main companies) listing focal length at 1550nm as a function of EFL for BK7 lenses
    • I have used this one from the CVI catalogue before (linked here), though it only has 1319 nm for BK7. The difference is small enough where I am going assume reasonable dispersion (and reasonable extrapolation from the chart) , and stop thinking about it.
  • Figure out what to do with the response about lasers from RIO
    • Email asking for more detailed info about PLANEX lasers

Lab

  • Razor blade measurements of waist size / at EOM (as described here - more info in elog:164)
  • " " measurements of waist size / loc at PMC
  • Align to PMC sans viewer (as described here)

 

  162   Sat May 21 18:45:03 2011 DmassNotesLab WorkAligning to the PMC

More work aligning and modematching to PMC.

  1. 300mm lens at 0.14m in front of coupler => 185um @ 0.431m
  2. 400mm lens @ 31 cm from EOM waist
  3. 500 lens @ 43 cm from EOM waist

I tried the wincamD to see if I could confirm spot size anywhere, but it wasn't sensitive at 1550nm, so I go to plan B: razor blades!

I borrowed a translation stage from the gyro and mounted a razor blade onto it (picture below). For occluding a PD (from math):

  1. Measure offset of PD
  2. Measure "full power" level of PD with beam illuminating
  3. Occlude beam with razor blade so that half of beams power is blocked - call this 50%
  4. Move razor blade until power increases to 84% (1/2 + erf(1/sqrt(2))/2) [picture of derivation attached]
  5. Distance moved is half the waist size (if I go one waist in distance from the center I have to make a 2.5% measurement of power level change, this is harder with cheap scopes and
  6. Sanity check - looking at the plot from this elog from when I did the full blade scan and erf fitting, this seems consistent with the scale

 

 

Attachment 1: RazorBlade.png
RazorBlade.png
Attachment 2: ERFderiv.png
ERFderiv.png
  161   Fri May 20 15:28:55 2011 DmassNotesLab WorkAligning to the PMC

I have set out all the optics with a mode matching solution found using the javascript modematching tool from one of Franks buddies in Germany - I have previously done sanity checks so I trust it

 

Aligning to the PMC:

I am trying to align to the PMC using a PDA10CS in reflection with a pickoff mirror for attentuation. I am used to having a viewer to get the spot close enough, so will be doing this more blind than I am used to.

Technique:

  • Assume the front PMC mirror is mounted flat onto the PMC face and thus is roughly perpendicular to the table
  • Get differential height DOF by looking at the beam in front of the last steering mirror at 1/2", 3", and 11" (after reflection) - set this to parallel to table plane
  • Get common height DOF by dead reckoning the center of the PMC front mirrors, and hope that the back mirror was mounted on the same plane (it should be)
  • Guess "wildly" at the horizontal DOFs by eyeballing from the top
  • Look at reflection off the PMC on a PD
    •  Align with 10 mW
    • PDA10CS gives 10mW x (0.9 A/W) x (7.5 x 10^2 V/A w/ 50 ohm load) = 6.75V (with no optical attenuation)
  • Sweep diode current via slow DC input on the front of the Thorlabs driver
    • PMC FSR = c/2L = c/(420mm) = 713 MHz
    • Covega df/dI = 90 MHz/mA
    • 10 mA modulation gives us slightly more than 1 FSR in frequency shift
    •  0.11Vpp signal (at 0.1 Hz) gives us a ~10 mA sweep on current

I tried to use the WINCAMD from the ATF to see if it had any sensitivity to 1550, but it seems pretty clearly to have ~0 response there. I think we should getting something from dataray to solve this problem. In the meantime mode profiling is going to have to be done with razor blades.

  160   Thu May 19 12:39:17 2011 DmassNotesGeneralMeeting

Quote:

Me Rana and Frank's stand in had a meeting:

Dick gave us some drawings to review, I have distributed these via email, here are the questions

 

 

TO DO

  • Write Procedure for cycling dewar!
  • Ask Dick what the wall thicknesses are
  • Ask him how much clearance we have at windows and middle joint
  • Ask if the cold plate is where the holes are tapped or if there is an extra part as the drawings seem to indicate
  • Change the diameter of the bottom lip
  • Ask for the hole pattern on top with helicoil inserts to be called out
  • Ask if he thinks Aluminum is OK (since he has designed it as such)
  • Cymak - poke JR to do this / figure out what he's doing
  •  Make big ole chart with urrthang in it
  • Chat with Steve about life, the universe, and Dewars
  • Confirm that the Sapphire windows have similar thermal expansion to Aluminum
  • Make first version of suspension mechanical design
  • Get the windows coated - quicker over quality
  • Talk to G$ about laser quotes

Other TO DO

  • Mode match and align to PMC
  • Sweep laser current, look at trans / refl of beam through PMC
  • Lock laser to PMC
  • Find HV driver
  • Align / mode match PMC transmission to low quality REFCAV
  • Use PMC PZT to lock to REFCAV
  • ???
  • Profit

Dewar Notes:

Walls are: 1/8"

LN2 chamber is 1/" or 0.09"

G10 - 0.02"

  159   Thu May 19 00:26:29 2011 FrankThings to BuyCavityQuote from Coastline

we finally got a quote from coastline. Delivery time is ~12-14 weeks and the price is rather high see full quote posted on the wiki

  158   Wed May 18 18:18:06 2011 DmassNotesGeneralMeeting

Me Rana and Frank's stand in had a meeting:

Dick gave us some drawings to review, I have distributed these via email, here are the questions

 

 

TO DO

  • Write Procedure for cycling dewar!
  • Ask Dick what the wall thicknesses are
  • Ask him how much clearance we have at windows and middle joint
  • Ask if the cold plate is where the holes are tapped or if there is an extra part as the drawings seem to indicate
  • Change the diameter of the bottom lip
  • Ask for the hole pattern on top with helicoil inserts to be called out
  • Ask if he thinks Aluminum is OK (since he has designed it as such)
  • Cymak - poke JR to do this / figure out what he's doing
  •  Make big ole chart with urrthang in it
  • Chat with Steve about life, the universe, and Dewars
  • Confirm that the Sapphire windows have similar thermal expansion to Aluminum
  • Make first version of suspension mechanical design
  • Get the windows coated - quicker over quality
  • Talk to G$ about laser quotes

Other TO DO

  • Mode match and align to PMC
  • Sweep laser current, look at trans / refl of beam through PMC
  • Lock laser to PMC
  • Find HV driver
  • Align / mode match PMC transmission to low quality REFCAV
  • Use PMC PZT to lock to REFCAV
  • ???
  • Profit
  157   Tue May 17 10:42:52 2011 FrankThings to BuyCavityCoastline stock material properties

Optical Grade Silicon
Resistivity Ohm-cm: >20
Purity: >99.999%
Type: P
Orientation:  Monocrystalline <100>

This information is from the certificate of conformance for the 1"Ø blanks
of silicon that they currently have in house.

  156   Thu May 12 22:47:37 2011 DmassLab InfrastructurePlotsSteering Mirror Reflectivity Measurements

We bought twi]o different types of steering mirrors for the lab. I forgot that there was more than one type when unpacking and mounting the "new" ones and thus did not label them / note which is which. (OOPS!)

I can tell the difference between the two types based on the tint, but here are some measurements to make sure one of the two types of mirror doesn't suck so we can buy more (of both). One is more green if you look through the back one is more orange, and when I look at one of the room lights, there is a slightly different tint to each.

 

 

I have labeled the two with the beautifully descriptive "SM1" and "SM2" labels. Reflectivities are measured with the ultraprecise awesome Thorlabs power meter!

Attachment 1: MirrorRefls.pdf
MirrorRefls.pdf
  155   Wed May 11 23:06:16 2011 FrankDailyProgressPMCPMC prototype assembled

As getting a new spacer for the 1550nm PMC will take a couple of more weeks we decided build a temporary PMC to be able to perform our locking test.
We re-use an old glass PMC spacer and old end-cap with old pzt. The mirrors are glued to the spacer using only a minimum of epoxy at a few point, just enough to keep it in place.
That makes it easier to remove and re-use the mirrors later once we get the new parts, which also lowers the risk of damaging them while removing.

Here are some pictures:

 P1020489.JPG

P1020465.JPG

P1020469.JPG

 P1020432.JPG

  154   Wed May 11 01:30:03 2011 FrankDailyProgressDrawingstable layout draft

draft sketch of the table layout to see how we can fit everything on the table and to count parts which are still missing.
Free space  (upper half) can be either a second, independent laser source or we use the AOM in the beginning.
Long term plan is to have two independent lasers so we try to come up with a table layout which does not require rearranging everything if we change plans.
The layout below has some parts (like EOMs) as placeholders, but the plan is to still use the diode for modulation or the pzt of the PMC for locking the PMC.
Would be nice to have a layout where we can try different options without having to change the layout. Beat part is a bit squeezed but should fit. Other options are possible but not shown. Tried to make optical pathes as short and equal as possible. Cryostat can be replaced by our single ULE cavity for testing, or we put it next to it where we still have space and use it as a third cavity with a flipper mirror somewhere.

Footprints of parts are close to actual size.
Grid is 1". Table is 6ft x 4ft. Hole in center 10" diameter. Cryostat about 18" diameter.

cryolayout_v1.png

 

cryotable.vsd

  153   Tue May 10 23:52:23 2011 DmassLaserSchematicsMode Matching in new setup

I need to do the mode matching for the new layout.

 

Fiber Coupler ---> Lens1 ----> EOM ---> MMLens2 ---> MMLens 3 -----> PMC ----> MMLens 4 ------> MMLens 5 ------> Fused Silica Cavity

What are the relevant waists?

Fiber Couple Output - from the F240APC-1550 datasheet

  • They give angle of divergence: 0.075 deg ( = 2*theta: defined later)
    • This can give us the waist size since w(z) = w_0 * sqrt(1+(z/zr)^2) ==> w(z) = w_0 * z/zr    for z>>zr
    • zr = pi * w_0^2 / lambda ===> w(z) = lambda * z / (pi * w_0)
    • dw(z) / dz = lambda / (pi * w_0) = tan(theta)50
    • w_0 = lambda / (pi * tan(theta)) = 1.55 um / (pi * tan(0.075 deg / 2)) = 753 um
  • and 1/e^2 diameter at 1 focal length from lens: 1.5 mm
    • zr = pi * w_0^2 / lambda = pi * (743 um)^2 / 1.55 um = 1.12 m
    • Since they say the spot size is 2 x the waist size 8mm in front of the collimator, I'll start by assuming the waist is 1/3 inches in front of the collimator, and seeing if the mode matching is affected by this assumption (I check with a card and couldn't really tell where the waist was since the divergence was so small. I will use 8mm for now).

EOM - Thorlabs EO-PM-NR-C3

  • L/2 = zr at EOM = 1 inch
  • w0 = sqrt(zr * lambda / pi ) = 110 um

PMC - 370 um (double check)

Fused Silica Cavity - 348 um

 Available Lenses: (there is no V-coating available from our normal vendors at 1550, so we intend to take the broadband AR [1-1.55 um] coated lenses from the ATF and replace them with 1064 V-coating lenses off the Cryo Lab budget). The BBAR coating is AR.18 from Newport.

Plano Convex:

  • 50 mm  (2)
  • 75 mm
  • 88 mm
  • 100
  • 150
  • 300
  • 400
  • 500 (2)
  • 750
  • 1000 (2)

Bi Convex:

  • 38.1
  • 50.2 (2)
  • 500

 

Mode Matching:

300 mm lens 7" in front of fiber output coupler => 195 um waist @ EOM

Try 2 lenses after EOM

PMC 39 - 51 inches from EOM

  • Not sure which mirror is in PMC...
  • PMC length is 210mm (according to 40m elog)
  • R = 0.5m => 450 um waist
  • R = 1m    => 350 um

If the waist is 450 um, we can try

  • f = 150mm @ 0.16m
  • f = 500mm @ 0.744m
  • gives 435um waist @ 1.10m

If the waist is 350 um, we can use:

  • f = 150mm @ 0.168m
  • f = 400mm @ 0.72m
  • gives 360um waist @ 0.993m

 

ELOG V3.1.3-