As the table was floated, we measured the noise from error point again.

     We tried to determine if the noise bump we saw were from the window, so we place an extra window in front of a mirror [add fig] and compare to the noise when there was no window. The results are not different that much.

     From a quick look, by adjusting the input power, from 1mW to 10mW. The shape of the noise from error point changes substantially. This could be come from RFAM or scattering.  I'll measure the noise vs input power after I optimize everything first. RFAM, beam splitter, back reflection have to be optimized.

we have replaced the second leg which was leaking. It could be that the legs are simply to old and the rubber got brittle. As far as i know the table has never been operated floating as it had the suspended reference cavity on it since the beginning. We operate at around 85 PSI, maximum operational pressure is specified at 100PSI so that should be OK. The second leg started leaking after one day of floating operation. We disassembled one today to have a close look but we can't really tell where the leak is. We will check with some pressure to see what's broken within the next few days. Let's see what the other legs do in the near future. We still have plenty of "spares", as Aidan bought two new sets (taller ones) for the old tables in the TCS lab. So we have 8 short ones which are currently not used (and seem to be newer than the ones we currently have). And they still work as passive legs.

we got the new nitrogen bottle this morning. We could check the price for nitrogen and compressed air bottles to see if it's worth getting a regulator for compressed air bottles in the future. But i think we should stick to the nitrogen. Once we fixed the leaking leg we shouldn't need much air anymore...

Today we did noise hunting stuff. We focused on ground loop and noise from RFPD. The noise from ground loop is probably come from the power supply of the RFPDs, and noise from RFPD looks very suspiciously similar to beat. More investigation is needed to be done.

  ==Ground Loop Problem==

     We saw a lot of odd harmonics from 60 Hz peak (180Hz, 300Hz, and so on). This indicates that we have a ground loop somewhere. We find that the table has the same ground from the cables (which come from the electronics shelf). We tried unplugging all the cables that connected to items on the table ( PDs, RFPDs, EOM), but it did not fix the ground loop. We have not tried unplugging the laser's cable yet. However, we think we know where the ground loop comes from. The power supply for the RFPD are three pins +/- 15 V with ground. Its ground is the same as power supply's ground which directly connects to the RFPD.  To fix this, we have to use a power supply from the electronics shelf, and we haven't tried to fix this yet.  This will be tested later.

==Noise at High Frequency==

We switched back to old FSS to compare the noise level. [add fig, details]

==Noise from RFPD==

   We checked the signal from error point of RCAV FSS loop. It did not changed with power input of the laser (from 1mW to 10mW), so we think it might come from the detection point, RFPD. We check the noise at error point of the FSS loop (old FSS) when

1.   the beam incident on RFPD was blocked, so only electronics noise is observed, and
2. the beam reflected from the cavity (off resonance) and fall on the RFPD, so we know what is the effect from the incident beam on RFPD (RFAM, scattering etc) and
3.  a mirror is used to reflect the beam back to the RFPD in front of the cavity, so we know what is the effect from the incident beam without contribution from the cavity.

 Fig1: setup to check noise from detection point.

We measured the slope of the error signal (1.1e5 Hz/V) which is used for converting the noise at error point to absolute frequency noise.

The result is shown in comparison with beat measurements (red and magenta) below:

Note: I have only coating noise in the noise plot so that the graph does not look to busy. The whole plot can be found in the matlab fig flie attached below.

Comment about the result:

1.      The result from case (2) (beam reflected off the cavity) has a shape very similar to that of beat measurement. It is only about a factor of 10 away from the beat. If we assume that the calibration we got from error point could be off by a factor of 3, and the noise in beat measurement includes the similar effect from ACAV's path which might have larger calibration. It is very likely that light falling on the RFPD is the origin of the bump at 100 Hz. 
2.      When we used a mirror to reflected the beam of in front of the chamber (case 3), the noise went down considerably. This means that the scattering bump come from the window of the chamber or the cavity itself. From inspection, we think it may be the window, because it is not very clean. A new AR window might be needed. 
3.     The electric noise at the detection point (case 1) is already higher than coating thermal noise level. RFPD has to be modified later.

To do: fine adjustment on beam splitter and waveplates, to reduce an

i've build a small baking chamber which we can use for small parts like posts, sensor and heater assemblies, wires, screws etc. It's a short (10" long) 6" CF reducer Tee with two heaters and a k-type thermocouple on it. We still don't have a RGA, but if we clean all parts carefully (made of materials we know that they are not a problem in general) and vacuum bake them for a few days we reduce the risk of contaminating our cavities by a lot. Current wait time for parts submitted for baking is ~2month. Would be great to find a slightly larger chamber to fit the larger parts in there as well (i don't want to use the (now) spare refcav chamber).

We can buy a tank of Nitrogen/ pressurized air  from Caltech warehouse, the extension is x4891. They have delivery service as well.

The vertical motion of the stack was modeled.

• In the current configuration, the two vertical resonances are to be located at 13.6Hz and 48.8Hz with Q of 63 and 18
  (Aha! You can see the suspicious peaks in the latest beat noise spectrum.)
   
• If the upper stage is suspended by four Stainless Steel spring (Frank suggested McMaster Carr 9435K91), the resonances move to 3.3Hz and 34.7Hz.
  The upper stage will be almost completely decoupled and will have very high Q that is determined by the loss of the SS springs.
   
• Currently there is no vibration isolation at 10Hz, and a factor of 200 at 100Hz. This will be improved to 10 and 10^-4, respectively.
  

Model parameters:

Current configuration

Mass of the upper stage, m1: 14.27 kg --- Frank's estimation
Mass of the lower stage, m2: 5.44 kg --- Frank's estimation
# of elastmers at the upper stage, n_spring1: 8
# of elastmers at the upper stage, n_spring2: 8
Spring constant of the elastmer, k1&k2: (0.396)*(2*pi*43)^2 --- RTV615 block, 43Hz resonance with 0.396kg mass
Damping factor of the elastmer, gamma1 & gamma2: 0.396*(43*2*pi)/20 --- Q=20 assumed based on M. V. Plissi, et al, Rev. Sci. Instrum. 69, 3055 (1998)

Modified configuration

# of elastmers at the upper stage, n_spring1: 4
Spring constant of the elastmer, k1: 1529N/m --- McMaster Carr 9435K91 (8.75 lbs./inch)
Damping factor of the elastmer, gamma1: 0 (no loss)

The attached zip-file contains a Mathematica file with full derivation/calculation and Matlab files with TF calculations.

We measured seismic noise (on three directions) on the table when the table was floated and not floated. Seismic is substantially reduced at frequency 10 Hz and above.

Note: the data is calibrated to velocity/rtHz. ( I checked the 40m entry, I used this calibration here. The data was corrected for 50Hz pole as well.

(after calibrating, the result look similar to what Jan did last year, see here psl:435)

==guralp cmg-40T setup==

We used Guralp CMG-40T 3-axis seismometer (borrowed from 40m with Jenne's help) to measure the noise on the table. The setup on the seismometer is "Broadband velocity". Signals from "Low velocity" channels are used to acquire the data.

Voltage supply is 12V, the output ground of the box is fake ground and it should not be connected to the ground of the oscilloscope/ spectrum analyzer. I used float ground on the spectrum analyzer, so it should be ok.

The voltage output for each channel is +/- 10V.

The position was tuned by adjusting the legs of the seismometer, so that the position readout for each channel is close to zero (we got less than +/-0.3V), the manual says it should be below 3.5V, so we are ok.

==seismic improvement==

3 directions and the corresponding resonant frequencies are

1. Vertical    (2.5 Hz)
   
2. North-South: horizontal transverse motion of the cavity (normal to cavity's beam line)  (1.8 Hz)
   
3.  East-West:  beam line direction       (1.7 Hz)

The seismometer has flat response up to 50Hz, the data has been corrected for the transfer function of the seismometer.

***The noise is getting worse at low frequency below 3 Hz is typical [link to Newport], and the signal above 50Hz are mostly noise, so both signals from floated and unfloated table are similar. (2011/10/06)

Transfer function from the floor to floated table. [Newport]

===beat improvement===

Then we checked the beat signal. We just made a quick check to see if the peaks between 10-100Hz would change or not, so we did not try to optimize the loop or anything. The result (red) improves, those peaks are decreased significantly, but it is bad that noise at frequency above 100 Hz goes up a little bit. This might be from the un optimized servo. So, the next step will be checking the servo, noise at high frequency.

today we tried to make a plan of how we can improve the seismic isolation in an effective way without to much redesign. From previous Q-measurements of the stack we know that the spring constant of the RTV is very high, but we actually never measured it. I though it might be good to simply replace the current RTV springs by conventional compression springs, stock parts if possible to make things even easier. So Koji and i measured the spring constant  of the RTV and we picked two different springs from Mcmaster (stainless steel), which can support the high load (~14.3kg including cavities and shields, screws etc.). Those springs have a much softer compared to the RTV. Koji will summarize the results of the RTV characterization in a different post.

According to first calculations we can shift the resonance frequency by about a factor of 10 by simply replacing the RTV by the stock springs, without any other changes. It will be slightly less than 3Hz. The only problem we have to solve is how we damp it. We have a couple of ideas and will figure that out when we have everything sitting on the HEPA bench for testing.

Below the mechanical properties of the individual components of the stack:

top stack, 6" wide, SS304 : 10.18kg
copper tube (rad shield) : 0.94kg ea. (2 total)
cavity (incl. mirrors) : 0.96kg ea. (2 total)
PTFE posts : 39g ea. (4 total)
shield posts : 8.4g ea. (6 total)
screws, 1/4-20, 3/4" long, SS 18-8 : 7g ea (8 total)
screws, #8-32, 3/8" long, SS 18-8 : 2g ea (12 total)
total  weight of top stack including shields, screws etc : 14.27kg

center stack, 4" wide, SS304 : 5.44kg

RTV springs, D=0.42", h=0.43" -> k~29000

McMasterCarr stock compression springs
spring 1: 9435K91 - 302 SS Precision Compression Spring 1.5" Length, .42" OD, .042" Wire, k=1530
spring 2: 9435K88 - 302 SS Precision Compression Spring 1.5" Length, .42" OD, .038" Wire, k=1020

had to replace one of the valves/regulators of the table. It started releasing air after a few minutes without any reason. Adjusting it a couple of times didn't fix the problem, after a few minutes at started again. After replacing it with a spare one everything works now fine. Tara measured the seismic noise with the Guralp of the floated and unfloated table and the effect on the beat signal which will be posted soon.

  I readjusted the lenses for mode matching a bit more and the visibility for both cavities are now ~93%. We will check beat measurement with floated table tomorrow.

we moved a nitrogen bottle from the tiltmeter lab to the CTNlab to float the table. Last time i tried this we realized that the pressure of the compressed air supply in the lab is not high enough (~35PSI max). As we now know that we need more seismic isolation we wanted to give it another try to see how much gain from floating the table and to determine how much we have to change the actual cavity support.

After connecting the table to the bottle and trying to adjust the regulators on the legs we saw that three legs seem to work fine but the forth one didn't move at all and one could hear that some air was leaking somewhere from inside the top of the leg, which we couldn't further investigate (no space when attached to the table). However the regulator seem to work just fine. So we decided to replace that leg. We took the pallet jack from the 40m to lift the table. After replacing the leg and adjusting the valves the table is now floating.

The maximum pressure for the legs is specified as 100PSI. In order to float the table we need about 90PSI, at 80PSI only part of the table floats. So we have to stick to the nitrogen bottle (or compressed air). We have a bottle rack so it's not a problem. We just have to run a hose to there which we have to buy. The hose has an O.D. of 1/4" and less than .17" I.D. We got the Guralp from the 40m to measure the isolation we gain from floating the table which we will do tomorrow.

.

As the cavities' height changed, I adjusted the lenses to fine tune the mode matching for RCAV, the visibility is ~85%

We might want to use a bigger beam splitter (the current one is 10mm cube) where the beam split to ACAV and RCAV paths, the spot radius is ~ 3mm. It might cause some diffraction problem.

The next step is to check the beat signal with the previous FSS. We suspect that something might be wrong with the TTFSS. If the old FSS can give us with better signal,

we will use the old one.

The beat signal I measured today(orange) has noise level close to what we had before (blue,purple), so I'll try to check it again tomorrow. Meanwhile, this is the plot with beat signal when we modified the suspension (removed the spring), and the cavities were in two separate chambers (blue). The noise level around 20-100 Hz is very similar to what we have right now. Since the spring filters mostly vertical seismic, these peaks should be mostly vertical seismic coupling into the system.

http://gravity.ucsd.edu/research/optical_seismometer.php

Tara showed me a quick plot which showed the spectrum taken with the new (current) setup and the one taken when we removed the springs from the wire suspension. They look pretty identical between 10Hz and 100Hz or so. So it is likely that we see a lot of vertical seismic. I've measured a little bit at low frequencies to see where we are and we are better than before now, i would guess an order of magnitude or so, without any optimized stabilization, alignment and loops. Tara will take a nice set of measurements tomorrow and make a nice plot. The peak at 6.7Hz or so is actually the horizontal motion of the two top stack plates (not only the top plate). I used our pzt-shaker to shake the table and even with a small signal i could increase it until i got scatter noise bumps around 100Hz. So we have to damp this somehow. Is there an easy way to get some more vertical isolation? what about putting the top plate on a few springs instead of rubber? How much do you typically get when floating the table? I don't have realistic numbers for that...

We also don't have a seismometer. Jenne took her's back and Jan shipped the others back to where they came from i think. i think we should get one which we can share between labs in bridge which we can keep for longer. We needed one quite often in the past couple of months and i guess once we start with the cryo cavity we will even more frequently. Any idea where to get a cheap one? We don't need an STS-2 or so... Seismic sucks anyway in bridge...

The problem with moving supports is that the spacer has pretty wide groves. But we have to think about a clever support anyways. Currently it's sitting on viton in the groves of the spacer, which, according to comsol, is very close to the optimum position but who knows in reality.

By taking the passive transfer function between a vertical seismometer on the table and the individual cavity signals, we should be able to see which direction to move the cavity supports so as to minimize the seismic coupling.

Our first iteration probably will have a sign error, but by making a few iterations we ought to be able to home in on a better support. Also, we should compare the theoretical estimate with the measured coupling in units of strain/(m/s^2).

made another step from 26.6degC to 30.6degC (4 Kelvin). Checked beat frequency to be ~72.192MHz @26.6degC, so we can re-use the iLIGO VCO when operating at high enough temperatures.

The initial step of 100mK was only a test to see if everything works. Local temp fluctuations have still a lot of influence to the precision of the reading, so the following numbers are rough estimates. The larger step will hopefully result in a better value.
The Marconi feedback signal changed by ~1.2V with a tuning coeff. of ~284kHz/V, so this gave a change of 341kHz total, so about 3.41MHz/K.

The original sensitivity can be estimated as

a_SiO2 = 5.5e-7  , L_spacer = 203.2mm  -> dL_spacer = 111.76nm/K

with FSR = 738MHz

-> df = 155MHz/K

changed set point from 26.5degC to 26.6degC for a first test at 2:42PM local time on fb2. Tuned the VCO feedback to be ~zero. Tuning coeff same as mentioned in post 682

lost lock of both cavities, beat frequency was still drifting. We currently have no slow feedback to the npro temp as the new fss servo is still not connected to the daq.

created a new PEM channels, C3:PSL-PEM_RMTEMP.  Sensor are connected to the temp box for the cavity. Will probably replace one of it with an OOL sensor we have to attach next time we remove the insulation. The other channels (formerly ACAV) will be used for the in-vac radiative heaters and stack temperature monitoring. If we need a second box around everything we have to check channels again. Don't want to reboot crate now as i have to check all channels and don't want to kick the heater now. Will do this tomorrow. Channel will not be recorded until tomorrow.

reconnected the VCO monitor signal to the Marconi tuning input. Tuning input range is set to 400KHz, tuning coeff is 283880 Hz/V, so ~284kHz/V

couldn't find some channels. Some of the connected (aux) signals seem to be physically connected to the DAQ but not assigned or loaded.
Here the current list of correct assigned channels (old channel names, will change soon):

C3:PSL-RCAV_SENS1 : VME3123, C0 S12 (sensor 1)
C3:PSL-RCAV_SENS2 : VME3123, C0 S13 (sensor 2)
C3:PSL-RCAV_SENS3 : VME3123, C0 S14 (sensor 3)
C3:PSL-RCAV_SENS4 : VME3123, C0 S15 (sensor 4)

C3:PSL-ACAV_SENS1 : VME3123, C0 S8   (sensor 1)
C3:PSL-RCAV_SENS2 : VME3123, C0 S9   (sensor 2)
C3:PSL-RCAV_SENS3 : VME3123, C0 S10 (sensor 3)
C3:PSL-RCAV_SENS4 : VME3123, C0 S11 (sensor 4)

C3:PSL-RCAV_OOLTEMP: VME3123, C0 S1  (we don't have an ool sensor, this is a room temp sensor)
C3:PSL-RCAV_TEMP        : VME3123, C0 S2  (this is the sum of all 4 sensors minus the voltage reference)

C3:PSL-ACAV_TEMP        : VME3123, C0 S5  (this is the sum of all 4 sensors minus the voltage reference)

C3:PSL-ACAV_VCOMON: VME3123, C0 S7 (monitor signal for VCO tuning)

connected but not used: S3 + S4   (aux temp sensors)

S6 was connected to the frequency counter. BNC cable is still connected, channel name should be C3:PSL-FSS_FREQCOUNT (have to check)

11:47pm : the current VCO frequency @26.5degC is ~71.294MHz still slightly drifting towards higher frequencies - will wait a few more minutes before i step it up.

11:56pm : now 71.318MHz @26.5degC

12:27am : now 71.373MHz

12:42am : now 71.394MHz

1:54am : now already 71.497MHz

2:39pm : now steady at 71.858MHz

when playing with the setup i've noticed that we are actually very sensitive to temperature fluctuations (more than i expected), so the absolute length (or CTE?) must be more different than we expected. So i've added the thermal insulation, drilled four new holes in the end caps, connected the temp sensors etc. Thermal control loop is running now. Made some changes to channels (mostly range limiuts, warning levels etc) and added four new software channels which i will use to generate absolute calibrated temp signals. The difference between the sensors is about 1K. I know it does not matter for the stabilization but i hate it as you never know if you have a gradient or degraded sensor or so...

I'ver started heating it up to 26.5C (you may ask why this number: laziness -  actually i've started the heater before finishing the insulation and that was the actual temperature at the time i finished it). The LO for the AOM already increased to 70.6MHz (from 64MHz). So my guess is that once we reach 30C we are back in range for the iLIGO VCO.

Will change the temperature over the weekend in a couple of steps to measure the temperature sensitivity  (differential length change with temp). This will tell us how good the thermal stability has to be to not be limited by thermal fluctuations. We then can make a decision on further things like second, external box, passive foam or metal box with active control etc..

 Beat signal from double cavities in the same chamber was measured. At DC to 100Hz, it seems to be dominated by seismic. Above 100 Hz, the frequency's laser noise is the limiting source. The setup has yet to be optimized. This is for a quick check to see how beat signal changes with the new seismic isolation setup.

      I measured beat signal after Paul and Frank locked ACAV. At frequency above 100 Hz, beat signal changes with RCAV's gain setup. Error signal from RCAV's mixer out matches the shape of the beat signal ( I did not do the  calibration with error signal slope, just observed the displays on SR785). However, at low frequency (DC - 100 Hz), the beat signal does not change with anything, RCAV gain, ACAV gain, PLL gain, so I'm quite certain that it's the real signal we see here.

Tuning range on PLL loop was 10kHz, the calibration is 7kHz/V. A peak close to 7 Hz might come from the beam line transverse mode of the stack.

At DC - 10 Hz, the noise is lower than before. This might be the result from common mode suppression. However, the two spacers are not identical. They have significantly different holes' sizes. This may explain why the seismic cancellation is not that good at higher frequency.

We might have to suspend the whole vacuum chamber to win against seismic.

 Beat path was setup. Beat RFPD sees the signal around 125 MHz. Beat measurement will be done tomorrow after ACAV is locked.

found the box with the beam splitters Dmass bought almost 2 years ago but never unpacked or used. They are super-polished 50:50 beam splitters for 532&1064nm but optimized for 1064nm.There are 16pcs total, so i don't see why we can't use 3 of them for our beat setup. We now have only SP optics in the critical beam paths except for the windows of the vacuum can, all lenses and wave plates where required. I hope this will reduce the amount of scattered light a little bit. The new setup only uses a minimum of components.

thought a little bit about how to create the difference of 127MHz. Here are two ways how we can do this right now:

We only use one AOM (as before). We can't use the second one as we would have to use it way beyond it's operating range (which we already tried Friday) or in second order where we don't get enough light if double passed.
So using only one we have the following options:

1. we use the Isomet 19" 80MHz OM driver which can be tuned down to 63.5Mhz using an external voltage. The frequency tuning knob can't go that low.
   We know that the phase noise is very high and the output signal is not even close to a sine wave. So this is OK for alignment but i would not use that for an actual measurement without detailed noise characterization.
    
2. we get one of the Marconi's from the 40m and use our 2W RF amplifier which we already used Friday. The phase noise is is very very low and we don't need a lot of tuning range, so it's even better.
   This guarantees that we don't sit on the phase noise of the VCO even if we only have a low UGF for the second loop.

New beat layout, both paths travel with the same distance before combining at the beam splitter. I'll try if this layout work tomorrow.

I'm a bit worried that the CCD cameras might not fit on the table. 

 I made a mistake in the last entry. When I measured the frequency difference, I did not notice that two beams path were shifted at different orders from AOM.

One was +160MHz, another was -160 MHz. So I re measured the frequency difference again, with both beams shifted by 160MHz, the frequency difference between

the two cavities is 127 MHz. Without thermal adjust on the cavities,This falls into a hard-to-work-with frequency region with the current VCOs we have.

To lock both cavities at the same time with df = 127 MHz, for now,  we use 0th order beam on RCAV's AOM, so that the beam's frequency is not shifted, and use a VCO [model's number] to drive ACAV's AOM at 63.6 MHz, so the frequency different between the two paths is 127 MHz, from this both cavities should be able to lock simultaeneously.

We should be able to see beat signal by this Tuesday.

dL~200nm + n*532nm, so you can't say what the common mode suppression really is. The only thing we know is that the length are not off by 1mm or so (which you would see and easily be able to measure), but it could be anything below that. Next time we open it we should use a good digital caliper and measure the actual lengths. But we don't have one which is long enough.

we can borrow the 2GHz PD from the cryo experiment for a quick measurement. Will get the shields/heaters next week and have the sensors. So we might be able to add those before the meeting, but i would do the test at 289MHz first to see what the noise looks like. If it's terrible we might have to work on other things as well. e.g. the cavity support.

closed the valve of the refcav chamber and turned on the ion pump. Initial current was 7mA and went down to .3mA within an hour, so looks pretty good.

 I aligned the beam to ACAV and RCAV and measured the frequency difference between the two cavities to be 289 MHz.

I used SLOWDC for calibration, the beam resonates in ACAV and RCAV at

 DClevel                                  RCAV                 ACAV

     1st resonance       -0.0520 [V]                0.0088  [V]

      2nd resonance       0.1046                      0.1661

     dV                           0.1566                        0.1573

 For a quick calculation, I used nominal FSR = 737 MHz. The calibration is then FSR/dV = 4700 MHz/V.

The frequency different between the two cavities give (0.1661 - 0.1048)*4700MHz/V = 289 MHz.

(I think the frequency might be too high for the beat PD, we may need to have thermal control before we can measure the beat)

 use df/f = dL/L, the differential length between the two cavities is   dL =  df * L * lambda / c ~ 0.2 micron.

So common mode suppression for thermo-elastic or cavity sagging due to seismic can be approximated to be ~ dL/L = 0.2 micron / 0.2 m =  1e-6.

We did ring down measurement of the new seismic stack with double cavities on top. Q and f0 are 25.85, 6.35Hz for horizontal transverse mode, and 21.53, 6.96 Hz for beam line mode.

 Left) beam line, translational motion, f = 6.96 Hz, Q = 21.5                                   Right, horizontal tranverse motion, f = 6.35Hz, Q = 25.9.

     We used shadow sensing technique to measure the ring down of the top stack with respect to the table. The cavities were placed on the stack, we will add copper tubes and heaters later, so the resonant frequencies might change from these measurements a bit).  We noticed that the motion of the top stack was almost similar to the motion of the middle stack. They moved together. We have to think how the transfer function of the stage looks like before calculating the new noise budget.

   We put two cavities in the same chamber. Next, we will measure the frequency difference between the two cavities when they have the same temperature to determine which frequency we need for the second AOM.

Before we put the stack back in to the chamber, we measured Q from two modes of the stack(with cavities on it).

The modes are longtitudinal motion (translation along the beam line) and horizontal transverse motion [see this entry]. Other modes i.e. pitch, yaw die away too fast to be measured with ring down measurement.

The cavities' mirrors were checked and cleaned. ACAV's mirrors had a trace of smudge which is now cleaned.

Since the pendulum suspensions were removed, the cavity's height is reduced from 7" to 6" above the table. The periscope's height were adjusted accordingly.

The beam is aligned to RCAV, the visibility is ~85%, it decreases because the position of the cavity's height changes.

Now I'm trying to align the beam to ACAV. Once the beam aligned, we can lock the beam and measure what is the differential length of the two cavities, hence the frequency.

the shields/heaters for the two cavities will be delayed by an unknown amount at this point. I'm still waiting for the two polished copper tubes and mounting parts to be finished. Also the Kapton heaters still didn't arrive.Even if we would get everything beginning next week at will take at least a week to get it cleaned, heaters wired, glued, cleaned and baked. Same for temp sensors etc. So we will try to use the cavities without it for a first shot. We might be lucky and the difference in frequency is not that big. (370MHz max possible). If we are lucky it's small (and we can slightly tune it by choosing the absolute temperature if they are not equal in length)). We can still work on the alignment and locking and upgrade next week or so. But even if we have a beat at 300MHz we can check the performance of seismic isolation and scattered light and see if we made a big mistake...

checked if the cavities fit on the new support, they do . We've put the assembled new stack top plate in vacuum over the weekend to clean it a bit more. Also turned on the heaters which won't do much but won't hurt. Will move everything into the other, final chamber and start aligning the cavities on Tuesday.

please be careful near the HEPA bench in 058E (PSL lab). I started the baking of the plate. The plate is sitting on the bench and is hot. It's powered by a HP power supply with 70V sitting on the floor. Warning signs are posted. The heater is insulated but anyways, be careful when near.that area. Setup will be removed Friday around noon. Same for the baking of the viton in Vladimir's lab. Heater and vacuum parts are hot, also the pump itself! Everything is hot, not just warm! Don't touch it.

yesterday i got the new top stack plate from the machine shop. They couldn't finish the entire job on time so i had to tap the holes and do the surface cleaning (sand blasting, grinding and polishing to clean the surface and reduce surface area) myself.
Getting a machined surface would have taken until mid next week (minimum, no guarantee) which would have delayed things too much to get a measurements done before the LVC meeting. Looks pretty nice, has kind of a mirror finish.

I started cleaning it this afternoon, the final sonic cleaning is underway right now and i will start air baking tonight (talked to Bob and he said that this is enough). Also boiled the viton in isoprop and it's getting baked in vacuum at the moment (for 24h). The teflon parts and screws i already baked.

We already removed the insulation from acav and vented it. So we are ready to install the new stack once the plate and the teflon is ready. As i don't know how bad the new parts are (outgassing)  thought about assembling the new stack and putting it in the acav chamber to pump on it over the weekend. Then we transfer the (cleaner) stack to the final chamber next week. We can then align both beams into the cavities and set up the beat parts while pumping on it with the turbo before we switch over to the ion pump. Or should we put everything together and start pumping Friday? Any suggestions? I tend to do it in two steps as i don't want to ruin the cavities and and if you ask Bob we should have gone through the whole class A cleaning process (which would take a month or more at the moment because they are somewhat of behind)

I prepared the VCO for RCAV loop. The signal output is quite distorted. I'll try another VCO to see if it will give better signal.

       A VCO is used for driving an AOM to shift the frequency of the laser beam. For RCAV loop, we have a LIGO prototype VCO to drive the AOM.  The VCO was not ready to be used when I first got it. A couple of calbes got detached from connections, and I didn't know the power input level. I could not find the schematic yet, Peter helped me figure out the power input, which is +/- 24V and how to fix the unattached cable. 

Fig1: detached connections. Now they are fixed.

    Once the box was fixed and the cable for power supply was ready, I checked the signal output. The output was high passed, and attenuated with 24dB attenuator. The signal did not look quite good compared to the LIGO VCO box we have been using, see fig 2 and 3.

 fig 2. Left) signal from the prototype VCO for RCAV. It is quite distorted.              Right) signal from the current VCO box (the good one) for ACAV. 

   There are also another 80 MHz VCO (fixed frequency). I'll check that VCO if it has better output. We don't really want to use the fixed frequency because we have to tune the frequency later for lower the beat frequency.

Here is the measured transimpedance of the RFPD.

I measured this by taking the optical transfer function using the Jenne Laser. The setup is shown below:

I needed to readjust my measured optical transfer function to account for the difference in power between the two diodes. Before, I assumed that all the energy that I measured using the ophir powermeter at the points before the photodiode was not actually going into the photodiode. For the tested photodiode, I got 89% of the voltage that I was supposed to measure if I used the quantum efficiency from the data sheet and no loss. The predicted measurement was calculated using the equation V=P*QE*l*e/(h*c)*Z, where V is the dc voltage output, P is the incident power on the photodiode, l is the wavelength, and Z is the DC transimpedance. However, I only measured 49% of the predicted dc voltage from the reference photodiode. This is most likely because not all of the light is hitting the reference photodiode. The diode is quite small, and I was unsuccessful in focusing the light on the diode even further. So I decided to just use the dc voltage and calculate the photocurrent. I can then use the ratios of the photocurrent to adjust for the difference in power of the incident beams. I can do this because the responsivities are just about the same and I'm going to assume the losses in both diodes are equal. I then multiply by the transimpedance of the reference photodiode to obtain the transimpedance of the tested photodiode. A much more cohesive, clear, and eloquent explanation for obtaining the transimpedance will be included in my final report :>] (I'm a bit tired now :>[ ).

Here is the characterization of the shot noise and the input referred noise floor at 35.5MHz.

I made two different types of fits and obtained two different answers for the transimpedance and for the shot noise intercept. I should also note that the parameter fittings seem to change (not exactly sure by how much) when I took more and more data, i.e., the parameters fittings didn't seem to be converging to a single number.

So in the end, we have 3 different answers for the transimpedance at 35.5MHz: Z= 1231, Z=1297, and Z=1397. I also have two different answers for the shot noise intercept: i_int=0.41mA, 0.48mA. For the shot noise measurement, I am inclined to say that I would trust the fit with the green and orange line the most, just because the shot noise has to be proportional to 2ei_dc, so we only had to figure out one parameter (which was done by finding the intersect between the input referred noise level and the line that forms from the shot noise).

However, I guess we should be a little bit more conservative and say that the shot noise intercept can be as high as 0.48mA. For the Z intercept, I am inclined to say that I would trust the same fit just because it was really the only parameter in this fit and the shot noise has to follow that line. So maybe Z=1397. I perhaps assumed to many things when I calculated the transimpedance from the optical transfer function, such as equal responsivities, equal losses, and that the referenced PD transimpedance is accurate. However, I don't really know how accurately shot noise can be measured. Perhaps the optical transfer function is more reliable in obtaining the transimpedance.

Sorry for the rambling. I'm a bit tired. Also, sorry this took so long. I basically spent two weeks trying to figure out why my shot noise measurement wasn't following theory and why my two measurements contradicted each other. It was a matter of hunting down two factors of two :<[.

So over the whole summer, I've put together, little by little, the simulink model for the path that locks the laser to the reference cavity. I'm going to attach the matlab files in a zip folder. I've tried my best to comment as best as possible and to make the programs as robust as I could possibly think of.

I am also attaching the LISO models to the TTFSS paths and for the fits to the measured PZT and EOM Actuator transfer functions. I'm going to list the links of past elogs that refer that mention these fits. I'm also going to attach the fits themselves.

LISO TTFSS

EOM Actuator Measurement Procedure

PZT and EOM Actuators Fit

A more complete description of my measurements can be found in my surf report

I am still not satisfied with these fits and would like some feedback in improving them. I am most worried about my measured EOM Actuator transfer function fit.

I put most optics on ACAV path. I have not tried to lock the cavity yet. I'll install ACAV RFPD next.

 I recalculated the mode matching so that the spot radius in AOMs is 100 um. Now the visibility of RCAV is 90%.

     From the previous mode matching calculation, the spot radius in AOM is 220 um. This was too large for ISOMET AOM and caused beam distortion. The AOM was designed for much smaller spot radius (50 - 110 um). So I recalculated to make the spot radius inside the AOM to be 100 um. This spotsize is small enough for ISOMET and not too small for Crystal Tech AOM.

Rise time is 35 ns (28.5MHz) for 100 um radius in ISOMET AOM, diffraction eff ~80%. This should be sufficient for our less than 1MHz bandwidth loop.

For the new layout, I have to remove the Faraday isolator behind the EOM for another lens. I'll try to intall it back later.

I forgot that 1.5x higher noise is in length fluctuation (Sx(f)). Since we measure the frequency which relates to length as df/f = dL/L. As the spotsize becomes smaller (same mirrors, shorter cavity), the length noise is going up by a factor of 1.5 when the cavity is shorter by a factor of 4 (20cm to 5 cm), see the quote. Thus the frequency noise is going up by 1.5x4 = 6. Still not very significant if we want to boost up the thermal noise level.

Here is the characterization of the shot noise.

I ended up with two different answers for the shot noise intercept. When I fitted it to Z(2e(i_{dc}+i_{int})^{1/2}, where Z is the transimpedance, i_dc is the measured data, and i_int is the shot noise intercept, I got 0.27mA.  When I fit two lines for the input referred noise level and the shot noise, the shot noise intercept is at 0.17mA.

Also, everytime I get more data, I end up with a different answer :(

While I'm at it, I will post the transfer function of the PD.

[koji, frank, tara]

Today we changed the AOM in RCAV setup from Crystal Tech AOM to Isomet AOM. The beam shape distorted and cause the reduction in cavity visibility from 75% to 63%

For RCAV, we will use ISOMET AOM because it will operate at a fixed frequency.  ISOMET AOMs have weird f dependent impedance, while Crystal Tech, see elog PSL 59 , so we don't want to use it for keeping laser locked to a cavity.

As the beam radius is quite large ~350 um for the ISOMET, the spot shape becomes more elliptic causing the visibility to reduce.

 I will check if I can find an easy solution for mode matching to make the waist smaller at the AOM (~150um in radius) or not. If this is not possible, I'll see how shot noise level will increase due to residue reflected light on RFPD.

I checked the RCAV loop performance with the new layout (double pass AOM + semi fixed RF summing box). The TF looks better, but it seems that the gain is still not enough. The suppressed laser noise (measured at mixer out) is still higher than the coating thermal noise .

          Since the cavity can be locked, I checked the loop performance of the new setup. A few changes in this setup are

•  double passed AOM
•  RF summing box is modified so that it has a notch at 35.5 MHz for high voltage input for feedback (previous value was 21.5 Mhz) but the impedance as seen from 35.5 MHz LO input is ~5-10 ohm instead of 50.  

==Setup==

  I kept most of the setup similar to what we have used before. The power input = 1mW, RF amplitude adjust = 10, RF Phase adjustment = 5.06V (so that the error signal appears symmetric), with phase flip = 180. The phase flip for FAST actuator on the TTFSS is at (-). Common/Fast gain =544/773.

The measurement follows the procedure in this entry:PSL 592.

==Result==

fig1: Transfer function of RCAV loop. UGF is at 33kHz. (this will be compared with the calculated TF later)

The result looks much better compared to the previous measurement entry:PSL 594. It is very likely that the modified RF summing box which now has a notch at 35.5 MHz improves the loop performance. The RF summing box does not have 50 ohm impedence match for RF input yet. 

       With the current gain setup, the suppressed laser noise is still higher than coating thermal noise (this is bad because to be able to measure the coating noise, the noise of the laser must be lower than that of the coating). The suppressed laser noise can be measured at mixer out. Use the error signal slope (0.675 MHz/V) to convert to frequency noise.  By increasing more gain (common/fast), we can suppress more laser noise but oscillation occurs. I'm not sure yet what causes the oscillation. The phase margin seems to be enough, so we may have to investigate more.

fig2: Lasers' suppressed noise measured from mixer out (red) is still higher than coating noise level. The beat plot was taken before I switched the RCAV servo to the new TTFSS

   The peak to peak value of the error signal is only ~ 80mV (used to be up to 200mV). If we fix the impedance of the RF summing box so that more power from LO can couple into the EOM, we might get larger error signal.