After Jamie fixed the third party repo issue with Ottavia, he was able to upgrade it to Ubuntu 12.04 Precise Pangolin. But its network stopped working.

I tried to fix its issues by ifconfig and GUI, but what it really wanted was for me to put the network cable back into its eth0 slot. The eth1 network card appears not be working anymore.

All seems fine now. Next I will mount the shared user disk from linux1 and put in a .bashrc.

Last night after checking cabling and turning on ISS, we tried several times to handoff to REFL_DC but it didn't work at all.

Some issues:

1. The ISS was injecting a lot of very low frequency power fluctuations because of bad AC coupling.
2. The SR560 @ LSC rack was saturating a lot with the x10 gain that Jenne and Rana put in; we turned it back to G = 1.
3. The ISS was also saturating a lot. We turned it off around 4 AM, but still no success.
4. The ALS sequence for finding the Red Resonance takes too long (~2 minutes), so we're trying a faster scheme tonight.

 We want there to be ~16000 cts of signal per quadrant on the optical levers. I think that most of the QPDs have been modified to have 100k transimpedance resistors.

From the attached 90 day trend, you can see that the ETMX, BS, PRM, and SRM are really low. We should figure out if we need to change the lasers or if the coating reflectivities are just low.

Steve, can you please measure the laser powers with a power meter and then reply to this entry?

Another possibility is that we are just picking a dim beam and a brighter one is available.

 Still no success tonight

• We took CARM OLTFs at various CARM offsets and could clearly see the peak in the optical TF (in once case ~2.5kHz), which gave us an indication of our offset (~200pm)
• REFLDC effectively sees the same plant TF as the transmission signals plus a zero at ~110 Hz, at all offsets under 1nm, from my simulations; this pushes up the optical resonance and causes a loop instability when we try to handoff. 
• We need to make the CARM OLTF steeper to suppress this instability, but also to make a good crossover with the AO path, which otherwise has too similar of a slope around the UGF, as we saw with our one arm test. 
• We're thinking of trying to turn the AO path on with REFLDC while keeping the arms on SQRTINV signals. This may be tricky, but if we can get the loop bandwidth above the optical peak, it'll be a lot easier to deal with, and transfer digital control to REFLDC as well. 

For some reason or another, I decided that we should see if the optical lever servos were injecting too much noise into the test masses. The ITMs are much worse than the ETMs and I am afeared that they might be making the main noise for our arms in the 20-40 Hz region. Jenne is checking up on these feedback loops to see what's up.

To estimate the actuator gains of the mirrors, I turned on 1 count drives from LSC/CAL oscillators into the LSC drives of each test mass at the frequencies shown in the plot with the resulting peaks showing up in in POX/Y with the single arm locks in red. I will leave these going permanently, but with 0.1 count ampltiudes; we need to make it so in the scripts.

We need to go back and have a look at all of our optical lever control filters, and make sure they make sense. 

In particular, we should have a look at the ITMs, since they have a huge amount of motion around 10Hz. 

Notes:  ETMX shouldn't have that lower notch.  The bounce mode Qs should be lowered.

 I'm in the process of filling this table

 I should replace ETMX He/Ne laser

I took loop measurements of ETMX pit and yaw, and set the upper UGF to be ~6Hz for both.  This required a pitch gain of 25, and a yaw gain of 16.

The spectra look similar to what they were before Steve did the swap.

I believe that the Xend aux laser was turned off earlier today, for Steve's work swapping out the oplev.  When I went down there, the red "off" LED was illuminated, and the LCD screen was showing something.  I pushed the green "on" button, and I immediately got green.

Also, I saw that the 24Hz roll mode was very rung up on ETMX.  I looked at the FM5 "bounce roll" filter, and it had some old values, 12Hz and 18Hz for the resonant gains.  All other optics have the proper 16Hz and 24Hz frequencies.  I copied the BS oplev bounce roll filter over to ETMX pit and yaw (both were wrong), and loaded them in.  The mode is starting to ring down.

 ETMX oplev qpd gain has to be increased.

 Atm3, Oplev sum read 12,000 counts when the qpd was disconnected ?

           Dark qpd was zero and normal He/Ne incident on qpd was 1,730 counts.

 This QPD circuit (D980325-C1 ) uses the nice OP497 Quad FET opamp as the transimpedance amplifier. It has a low enough current noise, such that we can increase the resistors (R1-4) up to 100k and still be Johnson noise limited. We should also make sure that the compensation caps (C3-6) are ~2.2 nF so as to not destabilize the opamp. f_low = 1/2/pi/R/C = 730 Hz.

I will do the swap later today unless someone else gets to it first. (note: check for oscillations w/ fast scope probe after installing)

I did these modification tonight. The slideshow of some images is attached. Instead of 100k, I used 97.6k thin film, since this seemed like an oddball size that doesn't get used otherwise. I forgot to measure the dark noise of the quadrants before doing the swap, but comparing the pit/yaw/sum before/after the swap shows that the signal is basically unchanged (since pit/yaw is normalized by SUM), but that the noise is lower by a factor of a few above 100 Hz due to being above ADC noise now. Previously, it was bottoming out at ~10 prad/rHz. Since the signal is unchanged, I guess that the calibration and therefore the loop gain should not have changed either...

And the sum went up by almost 10x as expected from the resistor change.

My SURF week-1 work...

Motivation:

To measure the transimpedence of  the Broadband photodiode (D1002969-v8), using a New focus photodiode (1611) as reference. The amplitude modulated Jenne Laser (1.2mW) was used. 

The steps involved in getting the transimpedence are as follows:

Acquiring data

• Get 2 sets of data from Network Analyzer Agilent 4395: One set of data will be for the transfer function of Ref PD over RF out. The other set for Test PD over Ref PD.
• The following conditions were set:

1) Frequency sweep range: 1MHz to 200 MHz.

2) Number of Points sampled in  the range: 201

3) Type of sweep: Logarithmic

• Set the NA to give the corresponding transfer function values in dB and also Phase response in degrees.
• Save the data into floppy disk for processing on the computer (The wireless way of acquiring data was not working when the experiment was conducted )

Plotting

• The matlab code attached (TransimpedencePlot.m) will then give plots for the absolute values of transimpedence in V/A.
• Logic involved in the code:
  • Transimpedence = Voltage response / (Responsivity of the photodiode * Power incident) 
  • Responsivity for BBPD is taken as 0.1 A/W and for NF1611 as 0.68 A/W as given in their datasheets.
  • Voltage response of Test PD w.r.t RF output of NA (in dB) = Voltage response of Test PD w.r.t Ref PD (in dB) + Voltage response of Ref PD w.r.t RF output of NA (in dB) 

 Results

The Plots of transimpedence obtained are attached (results.pdf) . The results obtained for BBPD is consistent with the ones obtained before, but the same method and code gives a different transimpedence for 1611.

The transimpedence of NF 1611 was obtained to be around 4-5 V/A which is very much off-track compared to the one given in the datasheet (elog: 2906).

The transimpedence of  Broadband photodiode (D1002969-v8) was around 1200 - 1300 V/A for most of the range, but the value started falling as the frequency approached 100 MHz. This result is consistent with DCC document: T1100467-v2.

How is the modulation depth assumed in the calculation?

If you don't know the modulation depth, you can't calibrate the transimpedance of each PD individually.

A recap of the ringdown measurements made at the 40m in mid 2012, the hardware that was installed for the same and results from the measurements.

IMC ringdown:

Hardware installed 
A trans PD was installed (elog 7122).
This PD does not exist in the trans path anymore.

Measurements
The polarity in the MC servo was flipped with the MC WFS disabled and the PD trans signal was used to look at the cavity ringdown. 
Ringdown time = 13 us
Cavity finesse from the measurement = 453 (inconsistent with actual finesse). 

Attachment: Ringdown measurement and fits

PMC ringdown:

Hardware installed
The AOM was installed before the PMC. The AOM was driven by the driver installed on the PSL table. (RF power ~1.5W @ 80MHz @1.0V modulation input). An RF switch was installed to control the AOM driver input. ZASWA-2-50DR+ was installed.
Note: The AOM was used by the ISS crew after this. So the RF switch has been removed and the AOM is no more a part of the ringdown setup.

Measurements
No measurements were made as we could not obtain relevant TTL signals to control the switch remotely.

To do this better:

1) Just step the analog input to the AOM (i.e. no RF switch) and measure the PMC trans output with a fast DC coupled PD. IMC should be unlocked and disable during this part. We want the PMC ringdown to be faster than 1 us.

2) Re-install a fast PD in the MC trans path without disrupting the existing MC trans QPD setup.

3) Measure IMC ringdown and fit the data to find the cavity losses.

4) Think about how to use the Isogai, et. al (2013) technique to better measure the losses, taking into account the mirror transmissions.

 Attached is the weekly work plan / equipment requirement / lab expert's presence needed for the upcoming week.

Main Problem:

The frequency counter (FC) takes in an analog RF input(signal) and outputs the frequency of the signal(Ranging from 1 MHz- 6000 MHz) in the digital domain (into a processor). The FC samples the data with a given sample rate( user defined) which ranges from 0.1 s to 1 s(faced problems in fixing this initially).  For data acquisition, we have been using a Raspberry Pi(as a processor) which is connected to the martian network and can communicate with the computers inside the 40m.  The ultimate challenge which I faced( and been knocking my head off from past two-three weeks) is the synchronization of clocks between the Raspberry Pi and the FC i.e the clock which the FC uses to sample and dump data( every 'x' s) and the clock inside the raspberry pi( used  in the loop to wait for a particular amount of time the frequency counter takes to dump successive data).

Steps Taken:

• To address this problem, first I added an external clock circuit which monitors the Raspberry Pi and the FC to dump and read data at a particular rate(which is equal to the sampling rate of the FC)In detail: http://nodus.ligo.caltech.edu:8080/40m/10129. 
• While doing so, at first the level trigger algorithm was used which means that the external clock frequency was half as that of  the reciprocal of the sampling rate and a trigger was seen every time the level shifts from +DC to -DC(of the external square wave).
• But this did not completely mitigate the issues and there were still few issues on how quickly the ADC reads the signal and R Pi processes it.
• To minimize these issues completely, an edge trigger algorithm which detects a pos edge(rising)  of the clock was used. The clock  frequency is now equal to the reciprocal of the sampling rate. This algorithm showed better results and greatly minimized the drift of the sampling time.

Psuedo Code(code attached):

open device : FC via USB-HID;

open device : ADC via I2C;

always(for t= recording time):

            read data from ADC(external clock);

            if pos edge detected:

                    read data from FC and store it in a register;

             else read data from ADC;

end

write data stored in the register to a file( can be an Epics channel or a text file);

Results:

The attached are the plots showing the time between samples for a large number of samples taken for different sampling times of the FC. The percentage error is the percentage of standard error in the timing between two samples for the data for the entire measurement. It can be inferred that this error has been cut down to the order of ms.

To do next:

• I have started taking phase measurements( analysis and plots will follow this elog) and also the PSD plots with the improved timing characteristics.

Looks good. Now you have the internal timer to verify the external clock.
If you can realize the constant rate sampling without employing the external clock, that's quite handy.

The attached are the PSD plots with improved FC timing(with the same code as in http://nodus.ligo.caltech.edu:8080/40m/10151). More plots(Phase and PSD) to follow.

New script called scripts/ALS/findRedResonance.py finds the IR resonance in less than 20 seconds.

1. Do a ~2 FSR scan of the arm over 10 seconds using the Phase Tracker servo offset ramp. (dt = 10 s)
2. Load the frame data for the TR_OUT and the Phase Tracker phase. (dt = 2 s)
3. Use Python (modern) SciPy to find all peaks with moderate SNR (dt = 3 s)
4. Take the top 3 peaks (all presumably TEM00) and choose the one closest to zero and go there.

The above is the gist of what goes on, but there were several issues to overcome to get the code to run.

1. None of our workstations have a modern enough Python distro to run either the ipython notebooks or the new SciPy with wavelets, so I installed Anaconda Python in the controls home directory.
2. In order to get the peak finder function to work well I gave it a range of peak widths to look for. If we change the speed of the ALS scan by more than 3x, we probably have to change the width array in the function.
3. I've used the find_peaks_cwt function in SciPy. This uses a Ricker ('Mexican Hat') wavelet to find peaks by doing multi-res transforms and looking for ridges in the wavelet space (I think)
4. To make the code run fast, I downsample from 16 kHz to 1 kHz right at the top. Would be better if we had downsampling in v2 of the getData function.

From the last plot:

- Subtracting the offset of 0.0095, the modulation depth were estimated to be 0.20 for 11MHz, 0.25 for 55MHz

- Carrier TEM00 1.0, 1st order 0.01, 2nd order 0.05, 3rd order 0.002, 4th order 0.004

==> mode matching ~93%, dominat higher order is the 2nd order (5%).

Eric: now we have the number for the mode matching. How much did the cavity round-trip loss be using this number?

AOM removed from the beampath and PMC relocked. 

AOM alignment:

1. Measured the initial power after PMC as 1.30W and reduced it down to 130mW.
2. Checked the power in the AOM zero order transmission before touching it. For 0-1V modulation input, the power dropped from 125uW to 98.3uW.
3. Steered the mirror right before the AOM to increase AOM zero order transmission and then carefully moved the AOM around to obtain maximum power attenuation. I repeated this a few times and the maximum attenuation that I could obtain was 125uW to 89.2uW (~30% attenuation).
Although this is not the right way to align the AOM, we do not have much options with the current setup as there is not enough separation between the zero order and first order beams and the AOM is on a fixed rigid mount.
4. I tried to dump the first order beam from the AOM and it wasn't satisfactory as well. There is barely any separation between the zero order and first order beams.

PMC relocking:

1. SInce the alignment to the PMC was disturbed by moving the AOM and the steering mirror in front of it, the PMC alignment was lost.
2. I could not relock the PMC at low power or high power. Rana had to come to rescue and fixed the alignment so that we could see flashes of PMC on the trans camera (This was done by aligning refl beam to the PMC REFL PD while giving a triangular ramp to the PMC PZT voltage).
Also I should not have tried to lock the PMC at high power as I could have been steering the beam at high power to the edges of the PMC mirrors that way and burning stuff easily.
3. Before fine tuning the alignment, I decided to remove the AOM from the beam path as there needs some work done on it to make it useful.
4. I removed the AOM from the beam path and relocked the PMC. 
5. PMC is relocked with 0.79 counts in TRANS and I measured the power after PMC 1.30W

Attachment: picture showing AOM removed from the beampath.

Goal:

To estimate the transfer function and the noise in the FC that is a part of the FOL-PID loop.

Measurements Taken:

The setup used for the measurements is described in my previous elogs.

The input modulation signal and the FC output were recorded simultaneously for a certain period of time and the phase and gain are estimated from the data.

Analysis(Data and code attached):

The recordings must contain equal number of data points(around 6000 data points in my measurements) for analysis.

The steps I followed to generate these plots are:

• Took the FFT of both FC out data(from FC) and Modulation input(from SRS via ADC).
• Estimated the phase angles at the particular modulation frequencies from the FFT data(in Matlab using  angle(x) for phase at the frequency f(x);x: is the frequency bin)
• Then for the phase of the system at a particular modulation frequency, 

                              Phase(system) =Phase(FC Signal) - Phase(Input Signal)

• Plotted the acquired phase vs the modulation frequency on a Semi-log graph.

Results:

From the plots its can be inferred that :the delay of the FC is almost 0 until the modulation of 0.1 Hz. Then there are phase shifts of  +/- 180 degrees showing that the system has multiple poles and zeroes(will be estimated after I have phase plots at few more carrier frequencies).

To Do Next: 

Phase plots for varying carrier frequencies and different sampling times.

Installation of FC inside the 40m.

If I assume 1sample delay for 0.1s sampling rate, the delay is Exp[-I 2 pi f T], where T is the sampling period.

This means that you expect only 36 deg phase delay at 1Hz. In reality, it's 90deg. Huge!

Also there are suspicious zeros at ~1.6Hz and ~3Hz. This may suggest that the freq counter is doing some
internal averaging like a moving average.

It would be interesting to apply a theoretical curve on the plot. It's an intellectual puzzle.

Rack 1Y2, I took transfer function measurements for each of the following demodulator boards: REFL11, REFL33, REFL55, REFL165, AS55, POP22, POX11 and POY11.

What I did:

1) Removed the wire at PD Input to demodulator board.

2) Put the MOD output from network analyzer into PD input of board.

3) Ran a sweep from 100kHz to 100MHz.

4) Measured the transfer function between PD RF MON and PD Input. (The PD RF MON signal came out of the RF multiplexer, so the mux is included as well )

5) Put the original wire back at PD Input.

Results:

The plots clearly show an attenuation of 20dB (factor of 10) for all the demodulator boards. This explains why my transimpedance measurements are off by 10 times.

Note: for REFL 165, there was an extra 100MHz high pass filter installed at PD Input. I did not remove this and made my measurements along with this.

To Do:

a) Modify the PDFR system to calibrate out this attenuation.

b) Measure the transfer function between the input and output of RF mux, so that we can have just the transfer function between PD input an PD RF MON (for documentation's sake)

 Thanks Koji , for your  hint for the brain teasing puzzle. I was looking into Filters that are usually used in devices like counters, DSO and other scopes. I found that , to improve the quality of the measurement one of the best approach  is averaging. I looked deeper into averaging and found out this:

There are two general use-cases for averaging . The first, successive sample averaging, takes a single acquisition and averages between its samples. The second, successive capture averaging, combines the corresponding  samples of multiple captures to create a single capture. Successive sample averaging is also called boxcar filtering or moving average filtering. In an implementation of this type of averaging each output sample represents the average value of M consecutive input samples. This type of averaging removes noise (one of the reasons the noise level was not bad: http://nodus.ligo.caltech.edu:8080/40m/10151) by decreasing the device's bandwidth(could be one of the reasons why the FC operates in 4 different frequency ranges). It applies an LPF function with a 3dB point approximated by  0.433 * s / M, where M is the number of samples to be averaged, and s is the sample rate in samples per second. 

Now I tried verifying the 3 dB points in the gain plots I generated :

For 1 s Sampling time : the 3 dB point for such a Boxcar filter should be at 0.433* 1/M. If we assume that it averages for 2 samples, M=2 which gives the 3dB point at 0.288 Hz but occurs somewhere between 0.3 and 0.4 Hz.  (http://nodus.ligo.caltech.edu:8080/40m/140619_120548/GainVsFreq.png)

For 0.1s Sampling time: the 3dB point should be at 2.17 Hz and in reality is 2.5 Hz(http://nodus.ligo.caltech.edu:8080/40m/140701_211904/gain.png).

Also, This type of filter will have very sharp nulls at frequencies corresponding to signals whose periods are integer sub-multiples of M/s. As seen my previous plots (http://nodus.ligo.caltech.edu:8080/40m/10118 , http://nodus.ligo.caltech.edu:8080/40m/10070) there are sharp nulls at frequencies

0.4 Hz for 1S sampling time and

at 1.5 Hz,3 Hz for 0.1 S sampling time as correctly predicted.

The moving average filter is  L-sample moving average FIR, with the frequency response as:   H(ω) = (1/L) (1 − e^{− jω L})/(1 − e^{− jω})..

There is an overall delay of (M - 1)/2 samples from such a length-M causal FIR filter. 

The expected bode plots for such a filter with L= 5 is attached(attachment 2).

The expected bode plots for such a filter with L= 4 is attached and compared with the measured.

RXA: When comparing two things, please put them onto the same plot so that they can be compared.

Using these numbers for both arms (Modulation takes away .2*.25 = 5% power, mode matching takes away 7% after that), I get the following from my data from March:

Xarm loss is 561.19 +/- 14.57 ppm

Yarm loss is 130.67 +/- 18.97 ppm

Obviously, the Xarm number looks very fishy, but its behavior was qualitatively very different when I took the data. ASDC would change from ~0.298 to ~0.306 when the Yarm was locked vs. misaligned, whereas the xarm numbers were .240 to .275. 

In any case, I'll do the measurement again tomorrow, being careful with offsets and alignment; it won't take too long. 

I repeated the exact steps above and made sure everything was back where it should be after I was done.

Reason I had to retake the measurements:

My script for acquiring data from the AG4395A network analyzer was such that it first acquired the magnitude data from channel 1 and then recorded phase data from channel 2 without holding its trace. Hence the phase and magnitude data were not exactly in sync with each other. So, when I tried to fit the data to a model using vector fitting, I ended up with very bad results.

I have now changed every single script relating to the network analyzer to just get the real and imaginary data in one go and then calculate the phase using this data.

The fitting process is now in progress and results will be up shortly.

 The goal of the measurements we made ( my previous 3 elogs) was to characterize the laser frequency thermal actuator that is a part of the FOL- PID loop.

For this we made indirect TF measurements for the thermal actuator by looking at the PZT response by 1)arm cavity( ETM ,ITM) displacement  and 2) temperature offset excitation. The goal was to do something like getting G1=TF3/TF1 and G2=TF3/TF2 and ultimately dividing G2/G1 to get TF2/TF1 with correct calibration. The final TFs obtained are the X and Y arm TFs for Laser frequency response vs temperature offset in(HZ/count). The calculations  in detail are:

Obtained    G1 = PZT response/ Temperature Offset (count/count): (in detail here )

Obtained    G2 = PZT response/  X and Y arm displacement( count/ count) : (in detail here)

Calibrated G2 to count/m ( in detail here)

Divided G2/G1 to get X and Y arm displacement/ Temperature Offset( m/ count) to get G3

Did these calculations:

dL/ L = dF /F

F = c/lambda ;Lambda = 532 nm  ; L = 

X arm length = 37.79 +/- 0.05 m

Y arm length = 37.81 +/- 0.01 m

TF: Laser Freq/ Temperature Offset = G3 *F/L       (HZ/Count)

The calibration coefficients for the ends  are :

X End:  [23.04 +/-  0.23 ]* 10^3  (HZ/Count)

Y End:    [18.71 +/-  0.2 ]* 10^3 (HZ/Count)

For the TFs of the temperature actuator on laser frequency I used ITMs for both the arms. The bode plots for the calibrated( HZ/Temp Count) are attached.

 For the X-Arm Thermal Actuator, I calculated the TFs at two different frequency ranges and combined the results where the coherence is high(>0.7). At 1 Hz the coherence was not as good as the other frequencies(due to the suspension resonance at 0.977 Hz).

The poles and zeroes are estimated after fitting this data using Matlab vectfit tool.The  graphs showing fit and measured values are attached.

Y arm Thermal Actuator:

5th order TF fitted: 

Gain: 9000

Zeroes:

z1 = -0.9799;

z2 = 2.1655; 

z3 = -2.9746- i * 3.7697

z4 = -2.9746+ i * 3.7697

z5 =  95.7703 + 0.0000i 

Poles:

p1 = -0.0985- i* -0.0845

p2 = -0.0985+ i* -0.0845

p3 = -0.6673- i* -0.7084

p4 = -0.6673+ i* -0.7084

p5 = -8.7979.

X-arm Thermal Actuator:

5th order TF fitted: 

Gain = 20

Zeroes:

z1= -305.7766

z2 =   -18.2774

 z3 =  -16.6167

 z4 =   -1.2486

 z5 =   28.1080

Poles:

p1  = -0.1311 - 0.1287i

p2 =  -0.1311 + 0.1287i

 p3  =  -8.3797 + 0.0000i

 p4 =  -4.0588 - 7.5613i

  p5 = -4.0588 + 7.5613i

I will use get the poles and zeroes from these fitted  bode plots and use it to build the PID loop.

 Koji asked me to get the calibration of the PZT counts to Volts for the the X and Y ends. Yesterday, I went inside the lab and took some measurements from the digital readout of the PZT by giving in a DC offset(-5 to +5 volts) to PZT_Out and read out from these channels:

For X-end:  C1:ALS-X-SLOW_SERVO1_IN1

For Y-end:  C1:ALS-Y-SLOW_SERVO1_IN1

Since a 20dB attenuator was placed in the path of X-arm readout while taking the Transfer functions(Detail), I did the calibration measurements without removing it from the path. However, for the Y arm there was no attenuator in the readout path.

The obtained calibration values are :

X- arm PZT : [146.3 +/- 2.37 ]  counts/Volt 

Y- arm PZT :  [ 755.1 +/- 3.6]    counts/Volt

The attached are the fit and data plots for the above calibration.

1) Don't be brainless. Redo the fitting of the Y arm. Obviously the fit is not good.

2) How can you explain the value from the ADC bit and range?

e.g. +/-10V range 16bit ADC => 2^16/20 = 3276.8 count/V

 The PZT seems to saturate at around +/- 3500 counts. So for the Y arm, I excluded the saturated points and fitted the data points again.

As for the calibration number, we expect the 3276.8 count/V for +/- 10 V range of a 16 bit ADC but the number is ~800 count/V. I couldn't figure out a reason why the number is so different.

The new calibration values are :

X- arm PZT : [146.3 +/- 2.37 ]  counts/Volt   (with a 20 dB attenuator included in the path)

Y- arm PZT :  [ 797 +/- 3.6]    counts/Volt  

I will get the calibration in MHz/V of PZT actuation and check whether these numbers make any sense.

To make MC auto locker running correctly, mcdown and mcup were revised

I tried it by unlocking MC several times. It seems OK. Let's see how it works.

Nominal gains for locking (to be taken care by mcdown)

C1:IOO-MC_REFL_GAIN
was 16 and is 19 now.

C1:IOO-MC_VCO_GAIN
was 9 and is 9 now too.

C1:PSL-FSS_MGAIN
was missing and now +13

C1:PSL-FSS_FASTGAIN
was +23.5 and is now +20.0

Nominal gains for operation ( to be taken care by mcup.

C1:IOO-MC_REFL_GAIN
was 19 and is 19 now too.

C1:IOO-MC_VCO_GAIN
was 25 and now uses ezcastep (ezcastep C1:IOO-MC_VCO_GAIN=9 +1,16 -s 0.1)

C1:PSL-FSS_MGAIN
C1:PSL-FSS_FASTGAIN

ezcawrite C1:PSL-FSS_MGAIN `ezcaread -n C1:PSL-STAT_FSS_NOM_C_GAIN`
ezcawrite C1:PSL-FSS_FASTGAIN `ezcaread -n C1:PSL-STAT_FSS_NOM_F_GAIN`

C1:PSL-STAT_FSS_NOM_C_GAIN`  is +18
C1:PSL-STAT_FSS_NOM_F_GAIN`   is +20

- Put the circuit diagram of the sum amp on/in the circuit enclosure and associate it with an elog [done].
- Update the circuit diagram of the pomona box [done]
ALL DONE

It seems that the MC auto locker and the FSSSlow PID servo survived a night.

PC Drive is still angry occasionally. We want to know what this is.

The summing amp is prepared to allow up to use bipolar full range of the FSS box output

This means that the FSS fast PZT output is now nominally 0V and can range +/-10V.

- FSS Box has the output range of +/-10V

- Thorlabs HV amp MDT694 accepts 0V ~ +10V

- This circuit add an offset of +5V while the main signal is attenuated by a factor of 2. The offset voltage is produced from the voltage reference IC AD586.

- In addition, a summing node and voltage monitors before and after the summing node are provided. They are useful to test the crossover frequency of the fast/PC loops.

- The output noise level at 10kHz was ~60 nV/rtHz. The transfer function of the circuit was measured and flat up to 100kHz. The phase delay is negligible at 10kHz and less than 3deg at 100kHz

- Although the schematic was drawn in Altium, the board is a universal 1U eurocard and all wires were hand soldered.

Reasoning to choose the current parameters:

FSS Common: 18dB
FSS Fast: 20dB

Attachment 1:
Openloop transfer function of the IMC loop with the nominal gain setting. The UGF is 176kHz and the phase margin is 48 deg.
This is about 3 time more bandwidth than the previous setting. (Good)

It is visible that the TF has sharp roll off around 1MHz. I wonder if this comes from the demodboard LPF and/or the PMC cav pole.
In fact, according to Manasa, the PMC has the ringdown of 164.6ns which corresponds to the cavity pole of 967kHz. So this must
be there in the OLTF.

From the plot, the order of the low pass is about 5. Subtracting the slope by the cavity pole, the order is four. If I look at the TF of the minicircuits
LPFs (this entry), the phase delay of the filter at 1/10 of the cut off freq is ~30deg. And the order of the filters are maybe 6th elliptic?
So it's not yet clear if the LPF is causing a significant phase delay at 180kHz.

More significantly, the gain margin at ~1MHz is way too small. This is causing a big servo bump at that frequency as seen in Attachment 2.

In total, my recommendation is to move the LPF freq up by x2 or x3, and give a mild LPF above 500kHz.
This requires some modeling as well as try and error.

Attachment 2:

This figure is to explain how the common FSS gain was set. By increasing the gain, the UGF is increased and we can enjoy more supression (from red to purple).
The more gain, however, the more servo bump we observe above the UGF. The gain was chosen so that the total PC feedback does not exceed 3V.

Attachment 3/4:

This figure explains how the fast FSS gain (namely crossover frequency between fast and PC) was set. When the fast is low (red) the phase margin between two loops
are plenty and therefore the openloop TF is smooth. But the PC's frequency domain is large and has to work more (in rms). As the fast gain is increased, the actuation
by the PC is offloaded to the fast PZT (that's good). But eventually the phase margin is not enough and the dip start to show up (purple). This dip cause worse closed loop TF,
as seen in Attachment 4, or even an instability of the loop eventually. So the fast gain was set somewhere in between (green).

 The PZT actuation on the laser frequency in MHz/V ( assuming the previous calibration here of the PZT count/V) is :

X- arm: 33.7 MHz/V

Y- arm: 14.59 MHz/V

This number seems to be wrong by a factor of 10. 

So we[I and EricQ] decided to trace the cables that run into the ADC from the PZT Out. We found a black LEMO box in the path to ADC,which is  an anti-aliasing filter for each input channel. However,in theory the response of this filter should be flat up until a few kHz i.e. for  the DC gain it should be 1. But we will manually test it and look at the DC gain of the LEMO box.

LISO Fit for the IMC open loop TF. The data and liso source for the fitting were attached in the ZIP file.

I noticed now that the open loop TF I measured has too less phase delay.
I used the closed loop TF to estimate the openloop TF.

Looking at this comparison, I'm afraid that the superboost was not on during the measurement.
I need a new measurement to design MC loop modification to give the AO path for broader bandwidth.

                  2005              ALL oplev servos use Coherent DIODE LASERS # 31-0425-000, 670 nm, 1 mW

    Sep. 28, 2006              optical lever noise budget with DC readout in 40m,  LIGO- T060234-00-R, Reinecke & Rana

    May  22, 2007              BS, SRM & PRM  He Ne 1103P takes over from diode

    May  29, 2007              low RIN He Ne JDSU 1103P selected, 5 purchased sn: T8078254, T8078256, T8078257, T8078258 & T8077178 in Sep. 2007

    Nov  30, 2007               Uniphase 1103P divergence measured

    Nov. 30, 2007               ETMX old Uniphase 1103P  from 2002 dies: .............., running time not known......~3-5 years?

    May 19, 2008               ETMY old Uniphase 1103P from 1999 dies;.....................running time not known.....~    ?

    Oct.  2, 2008                ITMX & ITMY are still diodes, meaning others are converted to 1103P earlier

                     JDSU 1103P were replaced as follows:

   May 11, 2011                ETMX replaced, life time 1,258 days  or 3.4 years

   May 13, 2014               ETMX , LT 1,098 days or 3 y

   May 22, 2012               ETMY,  LT 1,464 days or  4 y

   Oct.  5, 2011                BS & PRM, LT 4 years,  laser in place at 1,037 days or 2.8 y

   Sep. 13, 2011               ITMY  old 1103P &    SRM    diode laser replaced by 1125P  ..........old He life time is not known, 1125P in place 1,059 days or 2.9 y

   June 26, 2013              ITMX 622 days or 1.7 y    note: we changed because of beam quality.........................laser in place 420 days or 1.2 y

  Sep. 27, 2013               purchased 3 JDSU 1103P lasers, sn: P893516, P893518, P893519 ......2 spares ( also 2 spares of 1125P of 5 mW & larger body )

I have measured the current boosted MC CLG below 100kHz with an SR785. Swept sine only could get me down to 10kHz, but I was able to get down to 5kHz with a noise-injection measurement. 

I am attaching the SR785 outputs, which are in dB and Degrees. Additionally I pruned the areas of bad coherence out of these, and merged them to provide data files for the CLG and OLG in Real,Imaginary format.

 I did some further measurements, to try and see what corresponds to what. In the end I performed four measurements:

1. Closed loop gain measurement on SR785: Source to MC exc, T'd to channel one. Test 2 to channel two.
2. Open loop gain measurement on SR785: Source to MCexc, Test 2 to channel one, Test 1 to channel two.
3. Closed loop gain measurement on AG4395: RF Source to MC exc, T'd to R input. Test 2 to A input.
4. Open loop gain measurement on AG4395: RF Source to MC exc, Test 2 to R input. Test 1 to A input.

I then converted OLGs to CLG and vice-versa with CLG = 1/(1-OLG)

Here are two plots showing the measured and inferred loop TFs for both closed and open. 

The best agreement seems to be between the directly measured OLGs. Maybe I did something weird with the CLG measurements, or input impedances are distorting things ... 

All data is attached, along with code used to generate the plots. 

The closed gain I meant is the AO path: Use IN2 to excite the MC loop and measure IN1 using MON2(?).
In order to obtain the open loop gain from this meausrement, the gain mismatching needs to be compensated, though.

This measurement is to correctly predict the AO path response from the open loop transfer function.

Anyway, the openloop gain seems nicely measured. I'll try to predict AO path response from this.

Eric's OLTF turned out consistent with the AO path TF that has been measured by me on Jul 31 (entry 10322).

Attachment 1:
Updated empirical fit of the open loop TF by LISO.
In this fit, I gave some of the poles/zeros associated with the boost manually set so that I can use them for the servo design.
LISO itself can make better fitting if all of the variables are moved.

Atatchment 2:
The OLTF data and LISO source for the fitting.

Attachment 3:
Comparison of the AO path TFs. The red one was measured directly on Jul 31. The TF is normalized at the low frequency.
The blue was estimated from the OLTF model given above. They are well consistent now.

Attachment 4:
Now some servo design was tried. In the new design (blue), zeros of the super boost frequency was moved from 20kHz to 30kHz
with the hope of having flatter AO response. The improvement is very little while costing costing above 100kHz. Note that the vertical
axis is intentionally in a linear scale. In fact, the AO response is much improved compared to the one before the MC UGF was increased
(shown in magenta). We have a flatter response both in magnitude and phase.
Therefore I think there is no need to tweak the boost frequency for the AO path.
I'd rather recommend to inspect the high frequency LPFs to earn more gain margin at 1MHz as explained in entry 10322.

Attachment 5:
This figure shows the comparison of the TFs for the current and new design trial, just in case someone is interested in to see.

      May  13, 2014             ETMX,  .............laser in place 90 d

      May  22, 2012             ETMY, 

     Oct.  7,  2013             ETMY,  LT  503 d  or  1.4 y............bad beam quality ?

     Aug. 8,  2014              ETMY,  .............laser in place   425 days  or  1.2 y

We measured the TF of the MC Demod board today.

We set the Marconi to +3dBm and drove the PD IN port of the demod board, starting at 29.5 MHz. Then we looked at the beat signal amplitude in the output of the demod board. So this is a transfer function but with mag only. Plots from Q below.

Rana took the demod board out and took pictures of it. Inside, the post mixer low pass is a SCLF-5 from mini-circuits. This has a lot of cutoff down low. Since the purpose of this filter is only to cutoff the 2f-1f and the 3f-2f products, we need to have a lot of attenuation at 29.5 MHz. One day, we may want to re-instate that notch for the (3*f1- f_MC) beat frequency, but for now we want stability.

So, I recommend that we (Steve) get 3 each of the SCLF-10 and SCLF-10.7 from Mini-Circuits Tuesday morning. Maybe we can put them into a spare board?

Also, we should probably remove the 140kHz:70kHz lead filter which is in the MC servo board. Its out of date. I think it would be fine for us to get a 7-15 kHz UGF for the CM servo and the MC can basically do that already. Mainly we want to fix the high frequency shape to get more stability.

After the measurements and photos, we had to reset the MCWFS offsets to get the WFS to not break the lock. Seems very sensitive to offsets. Hopefully Andres will give us a new Gouy phase telescope.

SCLF-5!? It's surprising as the cut off of the OLTF is just above 1Hz. cf this entry

This means that not the demod board but MC or FSS boards seem to have large attenuation above 1MHz.

In this situation, does SCLF-10/10.7 really help us?