The freq fluctuation of the beat note has been measured with the following condition

• The IR beam only locked to the MC. The green beam locked to the arm
• Both of the IR and green locked to the x-arm

Calibration
- The output of the freq divider is already calibrated to have the unit of MHz.

- The transfer function between the analog PFD channel and the digital PFD output was measured to be -23dB = 0.7.
  The gain of the XARM-FINE channel was changed to 0.7 such that the output is calibrated in MHz.

Results

- I have not checked the analog noise level of the analog PFD path. We may need more whitening gain (by icreasing the gain of SR560).

- The analog PFD is always better than the digital PFD above 20Hz.

- Both the digital and analog PFD showed good agreement below 20Hz.
  Note the measurement was not simultaneous.

- When the arm is locked with the ETMX being actuated , the fluctuation of the arm length must be stabilized by a huge factor
(~10^5 according to Kiwamu's entry) However, we only could see the stabilization factor of 30.

As this residual is the difference of the freq noise felt by the IR and the green,
this is a real issue to be tackled.

- The RMS fluctuations of the arm with and without the IR beam being locked are 2MHz and 0.1MHz,
which correcponds to the arm length motion of 250nm and 13nm, respectively.
Ed: I had to use 532nm in stead of 1064nm. The correct numbers are 130nm and 7nm.

- Without the IR locked, The typical peak-to-peak fluctuation of the beat freq was 10MHz.

I found that some flakiness of the beat signals comes from the RF components for the beat detection.
They are touching the racks in an indefinite way. If we move the components the output of the analog PFD
goes crazy.

Once Kiwamu is back I will ask him to clean up all of the green setting in an appropriate way.

I wonder why POY11 has the dark noise level of 90nV/rtHz that is 5 times larger than that of POX (18nV/rtHz)
even though the Q are the same (~15) and the transimpedance is better (3.9k instead of 2k).

What cause this high noise level?
What is the expected dark noise level?

Suresh is saying 375mW and 0.375mW. Let's wait for his update of the actual power.

Also he is not using EPICS, there may be the factor of two missing for now.

I propose to use C1:ALS-xxx_xxx for the names of the green related channels, instead of GCV, GCX, GCY, GFD...

Like C1:SUS or C1:LSC, we name the channels by the subsystems first, then probably we can specify the place.

We can keep the names of the processes as they are now.

As per Kiwamu's request I made a light touch to the input steering and the mode matching lens.

Here V_ref and V_trans are C1:IOO-MC_RFPD_DCMON and C1:IOO-MC_TRANS_SUM, respectively.

Result: Visibility = 1 - V_ref(resonant) / V_ref(anti_reso) = 1 - 0.74 / 5.05 = 85%

What has been done:

• Alignment of the steering mirrors before and after the last mode matching lens
     V_ref: 2.7 ==> 2.2, V_trans: 34000 ==> 39000
• Moving of the last mode matching lens away from the MC (+ alignment of the steering mirrors)
     V_ref: 2.2 ==> 0.74, V_trans: 39000 ==> 55000

- Put priority on the list

- Put names on the items

- Where is the CDS TO DO ==> Joe

-

- Remote disconnection of the greeen PDH 

- What is the situation of the PD DC for the LSC PDs?

- SUS Satelite box Resister replacement ==> Jamie

- IMC mode matching ==> Jamie/Larisa 

- Mechanical shutters everywhere

- SRM OPLEV Connection

- MC OAF

- Better LSC whitening boards

- DAFI 

[Rana / Koji]

The MC servo loop has been investigated as the MC servo was not an ideal state.

With the improved situation by us, the attached setting is used for the MC and the FSS.
The current UGF is 24kHz with phase margin is ~15deg, which is unbearably small.
We need to change the phase compensation in the FSS box some time in the next week.

- We found the PD has plenty of 29.5MHz signal in in-lock state. This was fixed by reducing the LO power and the modulation depth.

- The LO power for the MC demodulator was ~6dBm. As this was too high for the demodulator, we have reduced it down to 2dBm
by changing attenuator to 12dB (at 6 oclock of the dial) on the AM stabilization box.

- The RF power on the MC PD was still too high. The PD mush have been saturated. So the modulation slider for 29.5MHz was moved
from 0.0 to 5.0. This reduced the 29.5MHz component. (But eventually Koji restored the modulation depth after the servo shape has been modified.)

- The openloop gain of the loop has been measured using EXC A/TEST1/TEST2. The UGF was ~5kHz with the phase mergin of ~10deg. 

- This quite low phase margin is caused by the fact that the loop has f^-2 shape at around 4k-100kHz. The reference cavity has
the cavity pole of 40kHz or so while the IMC has the pole of ~4kHz. Basically we need phase lead at  around 10-100kHz.

- We decided to turn off (disable) 40:4000 boost of the MC servo to earn some phase. Then MC did not lock. This is because the LF gain was not enough.
So put Kevin's pomona box in the FAST PZT path (1.6:40). By this operation we obtain ~75deg (max) at 560Hz, ~35deg at 5kHz, ~20deg at 10kHz.

- In this setup the UGF is 24kHz. Still the phase margin is ~15kHz. This phase lag might be cause by 1)  the MC servo circut 2) PMC cavity pole

NEXT STEP

- Put/modify phase lead in the FSS box.
- Measure the PMC cavity pole
- Measure and put notch in the PZT path
- Increase the UGF / measure the openloop TF

[Koji / Rana]

- Since the MC servo had UGF up to ~20kHz and huge servo bump at 50kHz, we needed more phase between 20kHz to 100kHz.

- Today a phase compensation filter in a Pomona box has been inserted between the MC servo box and the FSS box.
  This is a passive filter with zero@14kHz and pole@140kHz. We obtain ~60deg at around 50kHz.

- After the insertion, the lock of the MC was achieved immediately. The overall gain as well as the PZT fast gain was tweaked
  such that the PC feedback is reduced down to 1~2.

- The OLTF has been measured.
  The insertion of the filter change increased the UGF to 130kHz even with "40:4kHz" and double super boost turned on.
  The phase margin is 54deg. Quite healthy.

- Rana modified the existed Auto Locker script.
  It is now continuously running on op340m!
  We made a couple of testsif it correctly relock the MC and it did. VERY COOL.

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

NEXT STEPS
- Measure the PMC cavity pole
- Measure the circuit TF and try to shave off the phase lag.
- Measure the PZT resonance of the NPRO and put notch in the PZT path
- Increase the UGF / measure the openloop TF

We can modify the freq divider circuit to make it a comparator.

PMC TRANS/REFL on MEDM showed red values for long time.
TRANS (a.k.a C1:PSL-PSL_TRANSPD) was the issue of the EPICS db.

REFL (a.k.a. C1:PSL-PMC_RFPDDC) was not physically connected.
There was an unknown BNC connected to the PMC DC output instead of dedicated SMA cable.
So they were swapped.

Now I run the following commands to change the EPICS thresholds:

ezcawrite C1:PSL-PMC_PMCTRANSPD.LOLO 0.8
ezcawrite C1:PSL-PMC_PMCTRANSPD.LOW 0.85
ezcawrite C1:PSL-PMC_PMCTRANSPD.HIGH 0.95
ezcawrite C1:PSL-PMC_PMCTRANSPD.HIHI 1

ezcawrite C1:PSL-PMC_RFPDDC.HIHI 0.05
ezcawrite C1:PSL-PMC_RFPDDC.HIGH 0.03
ezcawrite C1:PSL-PMC_RFPDDC.LOW 0.0
ezcawrite C1:PSL-PMC_RFPDDC.LOLO 0.0

As these commands only give us the tempolary fix, /cvs/cds/caltech/target/c1psl/psl.db was accordingly modified for the permanent one.

grecord(ai,"C1:PSL-PMC_RFPDDC")
{
        field(DESC,"RFPDDC- RFPD DC output")
        field(DISV,"1")
        field(SCAN,".1 second")
        field(DTYP,"VMIVME-3113")
        field(INP,"#C0 S32 @")
        field(EGUF,"10")
        field(EGUL,"-10")
        field(EGU,"Volts")
        field(PREC,"3")
        field(LOPR,"-10")
        field(HOPR,"10")
        field(AOFF,"0")
        field(LINR,"LINEAR")
        field(LOW,"0.0")
        field(LSV,"MINOR")
        field(LOLO,"0.0")
        field(LLSV,"MAJOR")
        field(HIGH,"0.03")
        field(HSV,"MINOR")
        field(HIHI,"0.05")
        field(HSV,"MAJOR")
}

grecord(ai,"C1:PSL-PMC_PMCTRANSPD")
{
        field(DESC,"PMCTRANSPD- pre-modecleaner transmitted light")
        field(DISV,"1")
        field(SCAN,".1 second")
        field(DTYP,"VMIVME-3123")
        field(INP,"#C0 S10 @")
        field(EGUF,"10")
        field(EGUL,"-10")
        field(EGU,"volts")
        field(PREC,"3")
        field(LINR,"LINEAR")
        field(HOPR,"10")
        field(LOPR,"-10")
        field(AOFF,"0")
        field(LOW,"0.8")
        field(LSV,"MINOR")
        field(LOLO,"0.85")
        field(LLSV,"MAJOR")
        field(HIGH,"0.95")
        field(HSV,"MINOR")
        field(HIHI,"1.00")
        field(HSV,"MAJOR")
}

This is the grand plan we talked about in the beginning of the meeting.

• (Kiwamu) X-end Green cleaning up / Prep for DRMI
• (Bryan) Y-end Green
• (Suresh) Help Bryan / RF (w. Kevin)
• (Jenne) MC WFS / Y-arm IR alignment / MC adaptive feedforward (incl. CDS)
• (Koji) LSC
• (Joe) CDS cleaning up
• (Jamie) Help Joe / Noise Budget
• (Larisa) PMC scan / PSL photo&diagram
• (Barbarela) ASS

This is the log of the work on Wednesday 23rd.

1. Power Supply of the freq divider box

Kiwamu claimed that the comparator output of the freq div box only had small output like ~100mV.
The box worked on the electronics bench, we track down the power supply and found the fuse of the +15V line
brew out. It took sometime to notice this fact as the brown-out-LED of the fuse was not on and the power
supply terminal had +15V without the load. But this was because of the facts 1) the fuse is for 24V, and 2)
the large resistor is on the fuse for lighting the LED when the fuse is brown out.

I found another 24V fuse and put it there. Kiwamu is working on getting the correct fuses.

2. MC locking problem

After the hustle of the freq divider, the MC didn't lock. I tracked down the problem on the rack and found
there was no LO for the MC. This was fixed by pushing the power line cable of the AM Stabilizer for the MC LO, which was a bit loose.

For bode plot:

USE LOG-LOG plot for the amplitude

USE LOG-LINEAR plot for the phase

Search "Bode Plot" on web

I thought that "m40m" was the traditional alias for the sitemap...

rossa:~>alias
...

m40m ${medm_base} ${medm_newtail} &
...
sitemap medm -x /cvs/cds/rtcds/caltech/c1/medm/sitemap.adl

rossa:~>set|grep medm
medm_base       medm
medm_newtail    -x /opt/rtcds/caltech/c1/medm/sitemap.adl
medm_tail       -x /cvs/cds/caltech/medm/sitemap.adl

 I went through the entries.

1. Give us a photo of the day. i.e. Faraday, tilted lens, etc...

2. After all, where did you put the faraday in the plot of the entry 4466?

3. Zoomed-in plot for the SHG crystal show no astigmatism. However, the zoomed out plot shows some astigmatism.
How consistent are they? ==> Interested in seeing the fit including the zoomed out measurements.

Since last Friday I have been testing the broadband RF photodetector in order to figure out the capability of S3399 with the similar circuit as Matt's BBPD
We also like to figure out if it has sufficient performance for the 40m green locking.

The circuit diagram is shown in the first attachment. The RF amplifier is attached at the diode while the reverse bias voltage is applied at the other side of the diode. The amplifier's input impedance is used as the transimpedance resister. Note that the bandwidth of this configuration is limited by the RC filter that consists of the junction capacitance of the diode, the series resistance of the diode, and the transimpedance resister. This cut off freq is in general lower than that cut off obtained with the usual transimpedance amplifier which has the readout resister at the feedback path of the opamp.

The transfer function of the PD is measured using Jenne's laser. At the reverse bias voltage of 30V, the -3dB bandwidth of 178MHz was obtained. This is quite high bandwidth for the most of the applications at the 40m.

Because of the low transimpedance the low-noise level of the RF amplifier is very crucial. Recently we can obtain an ultra low noise RF amplifier like Teledyne Cougar AC688 which has the NF of 0.9dB with the bandwidth between 10MHz - 600MHz. Next step will be to obtain this kind of amplifier to test the noise performance.

Last Thursday, I tested Newport Servo Controller LB1005 with the X_arm green PDH servo.

The setup and the settings I could lock the arm is depicted in the attached figure.
To lock the cavity, follow the steps below

1) Toggle the switch to the "lower" position. This disengages the servo and reset the integrator.

2) Toggle the switch to the "middle" position. The zero freq is set to the "PI corner" freq. At the low freq the gain is limited
at the value of "LF Gain Limit". This gives us a single pole at the low freq.

3) Once the lock is acquired, toggle the switch to the "upper" position. This moves the pole freq to DC, resulting in the complete integration of the signal at the low frequency.

I measured the openloop transfer function (attachment 2). The amp is quite fast and exhibits almost no phase delay upto 100kHz.
The UGF was 10kHz with the phase mergin of ~45deg. I had to tune the input offset carefully to stay at the center of the resonance.

Long time ago, I looked at the manual of the video switcher.
http://media.extron.com/download/files/userman/Plus_Ultra_MAV_C.pdf
Here is the summary. This will be the basic of the more sophisticated switching program which may have GUI.

In principle, you can manually control the matrix via telnet. At the console machines, you can connect to the matrix using telnet

telnet 192.168.113.92

This opens TCP/IP port 23 of the specified machine. You will receive some messages.
Then type some command like:
--------------------

• 1*2!       (connect input#1 to output#2)
• 1,           (save the current setting into preset1)
• 1.           (restore the setting from preset1)

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

Basicaly that's all. There are many other features but I don't think we need them.

We can create a simple program with any of the language as any of the language has the capability of the TCP/IP connection.
e.g. C, Perl, Python. Tcl/Tk
Any of them are fine.

Now what we have to think about is how to implement the interface in the epics screen (or whatever).
It needs some investigation how the people is thinking as the ideal interface.
But, first of all, you should make the above three operations available as a simple UNIX command like:

videoswitch -i 192.168.113.92 1 2
videoswitch -i 192.168.113.92 -store 1
videoswitch -i 192.168.113.92 -recall 1
(There is no such command yet. These are showing what it should be!)

This can be done by a single day work and our life will be much better.

Responsivity of SGD-444A

I started to modify another green PD set.

It so far has the transimpedance of 240 Ohm on CLC409 for the RF output.

It shows the BB output upto ~100MHz.
The measurement shows the transimpedenca of ~90Ohm which is ~25% smaller than the expected gain of 120Ohm.
It is calibrated based on the transimpedances of Newfocus 1611 (10kOhm and 700Ohm for AF and RF).

The next step is to change the transimpedance resister to 2k and replace the PD to S3399 Si PD, which has the diameter of 3mm.
Then, the noise level will be measured. (and replace the RF opamp if necessary)

51 Ohm for CLC409

The datasheet of CLC409 uses 25Ohm there. This is to cancel the input bias current of the two inputs of the opamp.

The source impedance (series) of SGD444 is 50Ohm. So I used 50Ohm for the + input shunting.

However, I could probably use anything between 0-50Ohm as the datasheet itself tells that the bias currents are
not related between the two inputs. In addition, I am not sure how much the real series resistance of the PD is.

1kOhm for OP27

This resister is to ensure the (+) input to have a high impedance at high frequencies.

As far as OP27 is behaving as an ideal opamp, the (+) input has a high impedance.
Also if the inductor behaves as the ideal inductor, no photocurrent comes to the AF path.

However, if both of the op27 and the inductor show similar impedances to the RF transimpedance of 240Ohm,
the AF path absorbs some photocurrent and affects the RF transimpedance of the RF output.

We know that the inductor has a self resonance where the shunt capacitance take over the impedance of the coil.
Above that frequency, the inductor is no longer the inductor. The self resonant freq of this inductor is ~300MHz. It is OK, but not
too far from the freq of interest if we like to see clear cut off at around f>100MHz.
Also OP27 is an AF amplifier and I had no confidence about the input impedance of the OP27 at 100~300MHz.

If I put 1k in the (+) input of the OP27, I can ensure the entire AF path has the impedance of ~1k (at least 500Ohm even when L and OP27 are shorted).
I think the chip resister easily works as a resister up to 1GHz.

Correction:

The (-) input has been decoupled by the capacitor. So the series resistance of the PD is not the matter.
In this sense, we should use 0Ohm for the (+) input shunting.

[Jamie, Jenne, Koji]

We installed the new c1lsc and started the process.

We still need to configure bunch of the EPICS variables, matrices, and some of the filters.
This should be done in order to transmit the signals to the suspensions.
Jenne is going to work on this task tomorrow (Friday) morning,
and Koji will take over the task afternoon/evening.

Target: To lock the Michelson with the new RF/LSC

Status

RF generation box: READY - already ready to go to the IOO rack. (Suresh)

RF distribution box: In Progress - the internal components are to be connected. (13th evening - Suresh)

Placing PD and CCD: Done - PD and CCD on the AP table (13th Afternoon - Aidan, Larisa with supervision of Kiwamu)

Cabling1: Done - PD signal AP table to the demodulator (13th Afternoon - Jamie with supervision of Suresh)

Cabling2: Done - RF generation box (IOO Rack) to the demodulator

Demodulator: In Progress - Test and install (13th night - Kiwamu with supervision of Suresh)

LSC model: Done - Run the new LSC model. (It is named as "C1LST" so far) (13th evening - Jamie)

LSC medm: Done on 14th - Modify the current LSC medm screens Update the EPICS database Adjust the matrices (- Jenne with supervision of Koji)

[Jenne Koji]

The PD signals are transmitted to the suspension now.

The trigger thresholds were set to -1. This means the triggers are always on.

While Kiwamu was working on the RF cabling at the LSC rack, I removed 80% of SMA cables which were not connected anywhere.
The rack is cleaner now, but not perfect yet. We need patch panels/strain relieving for heliaxes, cleaning up of the RF/LO cables, etc.

I have made a small python script to handle the video matrix.

It is too far from the perfection, but I release it as it is already useful in some extent.

The script is in the /cvs/cds/rtcds/caltech/c1/scripts/general directory.

usage:

videoswitch.py in_ch_name out_ch_name


in_ch_name is one of the followings

MC2F, IFOPO, OMCR, FI, AS_Spare, ITMYF, ITMXF, ETMYF, ETMXF,
PMCR, RCR, RCT, PSL_Spare, PMCT, ETMXT, MC2T, POP, IMCR, REFL,
MC1F, SRMF, AS, ETMYT, PRM, OMCT, Quad1, Quad2, Quad3

out_ch_name is one of the followings

Mon1, Mon2, Mon3, Mon4, Mon5, Mon6, Mon7,
ETMY, MC1, PSL1, PSL2, ETMX, MC2, CRT9,CRT10,Projector,
Quad1_1, Quad1_2, Quad1_3, Quad1_4,
Quad2_1, Quad2_2, Quad2_3, Quad2_4,
Quad3_1, Quad3_2, Quad3_3, Quad3_4


Just a curiosity:

I just wonder how you have distingushed the difference between _111 and _111.

They are equivalent alone themselves. Have you looked at the contexts of the lines?
Or you just did not have the larger matrix than 16x16, did you?

I tried the idea that the PRC can resonate f1 and f2 at the same time if the arm gives the reflection phase to f1 and f2 with the ratio of 1 vs 5.

The details are described on wiki. The point is this removes all of the PRC/SRC/asymmetry mumbo jumbo.

The calculated cavity lengths for f_mod of 11.065399MHz are:

• Arm Length: 37.7974 [m]

• PRC Length: 6.7538 [m]

• SRC Length: 5.39915 [m]

• Asymmetry (lx-ly): 0.0342 [m]

Here is the actual values derived from the photos.

• Arm Length: 37.54 [m] (0.26m too short)

• PRC Length: 6.760 [m] (6mm too long)

• SRC Length: 5.415 [m] (16mm too long)

• Asymmetry (lx-ly): 0.0266 [m] (8mm too long)

- AC coupling for the comparator circuit of the green locking

In order to relieve the power consumption of the RF buffer, ac coupling circuits have been added.

The ac coupling before the buffer amp helps to relieve the power consumption in the chip.
But because of the distortion of the signal (and the limitation of the bandwidth), the output still has some DC (~0.6V).
Therefore, the output is also AC coupled.

Note that the BW pin of BUF634P should be directly connected to -15V in order to keep the bandwidth of the buffer.

The drawings are also uploaded on the green electronics wiki

The Martian wireless bridge has the ethernet cable inserted in the wrong connector.

It should be inserted to one of the four port. Not in the "INTERNET" connector.

Once the connector has been changed, the martian net as well as the internet became accessible from the laptops.

[Rana, Koji]

REFL55 was modified. The noise level confirmed. The PD is now ready to be installed.

Kevin's measurement report told us that something was wrong with REFL55 PD. The transimpedance looked OK, but the noise level was terrible (equivalent to the shotnoise of 14mA DC current).

Rana and I looked at the circuit, and cleaned up the circuit, by removing unnecessary 11MHz notch, 1k shunt resister, and so on.

I made a quick characterization of the PD.

First page:

The transimpedance ws measured as a function of the frequency. The resonance was tuned at 55MHz. The notch was tuned at 110MHz in order to reject the second harmonics. The transimpedance was ~540V/A at 55MHz. (For the calibration, I believed the DC transimpedance of 50V/A and 10000V/A for the DC paths of this PD and #1611, respectively, as well as the RF impedance (700V/A0 of #1611.

Second page:

Output noise levels were measured with various amount of photocurrent using white light from a light bulb. The measurement was perforemed well above the noise level of the measurement instruments.

Third page:

The measured output noise levels were converted into the equivalent current noise on the PD. The dark noise level agrees with the shot noise level of 1.5mA (i.e. 22pA/rtHz). In deed, the noise level went up x~1.5 when the photocurrent is ~1.4mA.

I think the installation of the PD DC signals are quite important. What to do
1) Connect the DC signals to the right top whitening board (be aware that there may be the modification of the whitening circuit).
2) Reconfigure the LSC model such that the DC signal is passed to the right channels (modify the left top part of the model)

Done. C1:PSL-PMC_PMCTRANSPD was improved from ~0.75 to 0.87.

Comparison between Hamamatsu S3399 and Perkin Elmer FFD-100

These are the candidates for the BB PD for the green beat detection as well as aLIGO BB PD for 532nm/1064nm.

FFD-100 seems the good candidate.

Basic difference between S3399 and FFD-100

- S3399 Si PIN diode: 3mm dia., max bias = 30V, Cd=20pF

- FFD-100 Si PIN diode: 2.5mm dia., max bias = 100V, Cd=7pF

The circuit at the page 1 was used for the amplifier.

- FFD-100 showed 5dB (= x1.8) larger responsivity for 1064nm compared with S3399. (Plot not shown. Confirmed on the analyzer.)

- -3dB BW: S3399 180MHz, FFD-100 250MHz for 100V_bias. For 30V bias, they are similar.

I've got confused

1) Are these the DC responses of the coils? If that is true, we need to specify the resonant frequency of each suspension to get the AC response.

2) Are these the AC responses well above the resonant freqs? In that case, The responses should be x.xxx / f^2 [m/counts]

The PMC exhibited the reduction of the transmission, so it was aligned.

The misalignment was not the angle of the beam but the translation of the beam in the vertical direction
as I had no improvement by moving the pitch of one mirror and had to move those two differentially.

This will give us the information what is moving by the temperature fluctuation or whatever.

Minicircuits ERA-5SM was used for the RF amp of the BBPD. This amp is promising as a replacement of Teledyne Cougar AP389
as ERA-5SM gave us the best performance so far among the BBPDs I have ever tested for the aLIGO BBPD/Green.

The -3dB bandwidth of ~200MHz and the noise floor at the shotnoise level of 0.7mA DC current were obtained.

1. Introduction

The aLIGO BBPD candidate (LIGO Document D1002969-v7) employs Teledyne Cougar AP389 as an RF amplifier.
This PD design utilizes the 50Ohm termination of the RF amp as a transimpedance resistance at RF freq.

However, it turned out that the bandwidth of the transimpedance gets rather low when we use AP389, as seen in the attachment2.
The amplifier itself is broadband upto 250MHz (the transfer function was confirmed with 50Ohm source).
The reason is not understood but AP389 seems dislike current source. Rich suggested use of S-parameter measurement
to construct better model of the curcuit.

On the other hand, the RF amplifiers from Minicircuits (coaxial type like ZFL-1000LN+), in general, exhibit better compatibility with PDs.
If you open the amplifier case, you find ERA or MAR type monolithic amplifiers are used.

So the question is if we can replace AP389 by any of ERA or MAR.

2. Requirement for the RF amp

- The large gain of the RF amp is preffered as far as the output does not get saturated.

- The amplifier should be low noise so that we can detect shot noise (~1mA).

- The freq range of the useful signal is from 9MHz to 160MHz.

The advanced LIGO BBPD is supposed to be able to receive 50mW of IR or 15mW of 532nm. This approximately corresponds to
5mA of DC photocurrent if we assume FFD-100 for the photodiode. At the best (or worst) case, this 5mA has 100% intensity modulation.
If this current is converted to the votage through the 50Ohm input termination of the RF amp, we receive -2dBm of RF signal at maximum.

This gives us a dilemma. if the amp is low noise but the maximum output power is small, we can not put large amount of light
on the PD. If the amp has a high max output power (and a high gain), but the amp is not low noise, the PD has narrow power range
where we can observe the shotnoise above the electronics noise.

What we need is powerful, high gain, and low noise RF amplifier!

Teledyne Cougar AP389 was almost an ideal candidate before it shows unideal behavior with the PD.
Among Minicircuits ERA and MAR series, ERA-5 (or ERA-5SM) is the most compatible amplifier.

Considering the difference of the gain, they are quite similar for our purpose. Both can handle upto -2dBm,
which is just the right amount for the possible maximum power we get from the 5mA of photocurrent.

3. Test circuit with ERA-5SM

A test circuit has been built (p.1 attachment #1) on a single sided prototype board. 

First, the transfer function was measured with FFD-100. With the bias 100V (max) the -3dB bandwidth of ~200MHz was observed.
This decreases down to 75MHz if the bias is 25V, which is the voltage supplied by the aLIGO BBPD circuit. The transimpedance
at the plateau was ~400Ohm.

Next, S3399 was tested with the circuit. With the bias 25V and 30V (max) the -3dB bandwidth of ~200MHz was obtained although
the responsivity of S3399 (i.e. A/W) at 1064nm is about factor of 2 smaller than that of FFD-100.

The noise levels were measured. There are many sprious peaks (possibly by unideal hand made board and insufficient power supply bypassing?).
Othewise, the floor level shows 0.7mA shotnoise level.

The RF amplifier of the prototype BBPD has been replaced from ERA-5SM to MAR-6SM.
The bandwidth is kept (~200MHz for S3399 with 30V_bias), and the noise level got better while the maximum handling power was reduced.

MAR-6SM is a monolithic amplifier from Minicircuits. It is similar to ERA-5SM but has lower noise
and the lower output power.

The noise floor corresponds to the shotnoise of the 0.4mA DC current.
Now the mess below 50MHz and between 90-110MHz should be cleaned up.
They are consistently present no matter how I change the PD/RF amp (ERA<->MAR)/bias voltage.
I should test the circuit with a different board and enhanced power/bias supply bypassing.

Discussion on the RF power (with M. Evans)

- Assume 5mA is the maximum RF (~50mW for 1064nm, ~15mW for 532nm). This is already plenty in terms of the amount of the light.

- 100% intenisty modulation for 5mA across 50Ohm induces -2dBm RF power input for the amplifier.

- Assume if we use MAR-6 for the preamplifier. The max input power is about -18dBm.
  This corresponds to 16% intensity modulation. It may be OK, if we have too strong intensity modulation, we can limit the power
  down to 0.8mA in the worst case. The shot noise will still be above the noise level.

- In the most of the applications, the RF power is rather small. (i.e. 40m green beat note would expected to be -31dBm on the RF amp input at the higherst, -50dBm in practice)
So probably we need more gain. If we can add 10-12dB more gain, that would be useful.

- What is the requirement for the power amplifier?

• Gain: 10~12dB
• Output (1dBcomp): +3dBm +Gain (13dBm~15dBm)
• Noise level / Noise Figure: 3nV/rtHz or NF=14dB
  The output of MAR-6 has the votage level of ~7nV/rtHz. If we bring the power amplifier with input noise of ~3nV/rtHz,
  we can surppress the degradation of the input equivalent noise to the level of 10%. This corresponds to N.F. of 14dB.

Search result for Freq Range 10-200MHz / Max Gain 14dB / Max NF 15dB / Min Power Out 13dBm
GVA-81 is available at the 40m. ERA-4SM, ERA-6SM, HELA-10D are available at Downs.

Conversion between nV/rtHz and NF (in the 50Ohm system)

SN1: Connect signal source (50Ohm output) to a 50Ohm load.
Power ratio between the noise and the signal

SN1 = (4 k T (R/2)) / (S/2)^2

SN2: Connect signal source (50Ohm output) to an RF amp.
Only the voltage noise was considered.

SN2 = (4 k T (R/2) + Vn^2) / (S/2)^2

10 Log10(SN2/SN1) = 10. Log10(1 + 2.42 (Vn / 1nVrtHz)^2)

e.g.
Vn: 0 nVrtHz ==> 0dB
Vn: 0.5 nVrtHz ==> 2dB
Vn: 1 nVrtHz ==> 5dB
Vn: 2 nVrtHz ==> 10dB
Vn: 3 nVrtHz ==> 13.5dB

BBPD update

- The BBPD circuit has been constructed on the aLIGO BBPD board

- It still keeps 200MHz BW with FDD-100 Si PD for the 100V bias.

- The noise spectrum has been cleaned up a lot more. It shows the noise level of the 0.4mA shotnoise between 9-85MHz.
The noise at 160MHz is the noise level of the 1mA shotnoise.

Some of the noise peaks at around 97MHz came from the bias voltage.

What to do next

- Confirmation of the performance with the original aLIGO BB PD configuration.

- Notch filter for 9MHz (for aLIGO).

- Implementation of a power amplifier. (issues: power supply and heat removal)

Key points of the power supply installation

• We followed the grounding configuration for KEPCO except for the signal ground connection
• AC power supply has been obtained from the local power strip. This also provides chassis earthing (for safety)
• The chassis is connected to the shieldin of the DC supply cable. The other end should be isolated.

• The low voltage side of Sorensen's DC outputs are connected in order to share the same reference  level.
• The ground level is provided from the cross connect. The cable is connected between the cross connect ground to the sorencen.
  Unlike the KEPCO case, this cable does not have the current return, but just is to define the voltage level of those Sorensens.

• New AC&DC cables have been nicely strain-relieved.

Incidentally, a blown fuse on 5V line at 1Y2 rack was found during the intallation of Sorensens.
The fuse (5A 125V) has been replaced and fixed.

When I plugged the fuse in, I heard some sound like relays were switched. Are there any relays in the LSC rack?

It was a 9th fuse from the top as seen in the picture.

I have made my own AutoCAD WS account and put the latest 40m layout.
AutoCAD WS is a free cloud service which enable us to browse/edit the DWG files.

You can view/play/print it from the following link even without making the account.

https://www.autocadws.com/main/publish?link=RVFmNEtOd3EzNVFtRXcx

If you make your account you can actually save it although the editing capability is somewhat limited.

Notes:
- It needs the Flash plug-in and Internet connection.
- I think Safari is the most stable platform on Mac. (i.e. I had some malfunctions with Firefox and Chrome)

[Koji, Steve]

DC power supplies for the RF generation box are now in place. They are the top two of the 6 Sorensens in the OMC short rack next to 1X2.

We made the connections as we did for the RF distribution box, the power supplies labele, and the cables strain-relieved.

The power supply is not yet connected to the actual RF generation box. This should be done by Suresh or someone with the supervision of him.

Note:
We have two +18V supply on the short OMC rack, in total. One is for the RF source, the other is for the OMC PZTs, whitening, etc.
This is to avoid unnecessary ground loop although the grounding situation of the OMC side is not known to me.

- Found the inductor which shunts the positive input of MAX4107 was not touching the ground.
This left the positive input level undetermined at DC. This was why MAX had been saturated.
The PCB has a cut, so it was surprising once the circuit worked.

- Resoldered the inductor to the ground. This made the circuit responding to the intensity-modulated beam.

- But the resonances and the notches were totally off, and the 200MHz oscillation has resurrected.

- Attached 40Ohm+22pF network between the neg-input of MAX and the gnd. This solved the oscillation.

- Made the tuning and the characterizations. The PD is on Kiwamu's desk and ready to go.

More to come later

- Resonant at 55MHz. The transimpedance is 258Ohm. That is about half of REFL55 (don't know why).

- 11MHz&110MHz notch

- The 200MHz oscillation of MAX4106 was damped by the same recipe as REFL11.

The DC Transimpedance of POP55 was increased from 50 Ohm to 10010 Ohm. There is the offset of 46mV. This should be cancelled in the CDS.

Here is the conclusive result for the circuit configuration for aLIGO BBPD and 40m Green PD.

- Use Mini-circuits MAR-6SM for the RF preamplifier. The 50Ohm input impedance is used for the RF transimpedance.
  The maximum output is ~4dBm.

- Use Mini-circuits GALI-6 for the RF middle power amp. The gain is 12dB and the amplifier is linear up to +17dBm. i.e. This is still linear at the maximum output level of MAR-6SM.

- The total RF transimpedance is ~2k. The DC transimpedance is also 2k.

- The bandwidth is 80MHz with FFD100 and internal 25V bias. When S3399 is used, the bandwdith goes up to 180MHz
although the responsivity of FFD100 at 1064nm is better than S3399 by a factor of 1.5. At the 40m we will use S3399 for the green BB PD.

- By adding an LC network next to the PD, one of the unnecessary signal can be notched out.
As an example, 9MHz notch was placed for the FFD100 case.

- Noise level: ~10pA/rtHz as a floor noise level at around 30MHz. This corresponds to the equivalent dark current of 0.4mA.

Matt has finished the PCB layout. We will order small first batches, and stuff it for the test. Some of these will be the 40m green PD.

The full characterization of REFL11 is found in the PDF.

Resonance at 11.062MHz
Q of 15.5, transimpedance 4.1kOhm
shotnoise intercept current = 0.12mA (i.e. current noise of 6pA/rtHz)

Notch at 22.181MHz
Q of 28.0, transimpedance 23 Ohm

Notch at 55.589MHz
Q of 38.3, transimpedance 56 Ohm