Comment to OSEM LED screening preparation (Click here to view original report: 114)

[Abstract]

For measuring the LED intensity noise, the large DC voltage spoils the dynamic range of the FFT analyzer. The AC coupling could be usable to overcome the situation, so the difference between the AC and DC coupling inputs of 35670A's are measured. We should include the fit function for the calibration of measured data with the AC couping input.

[Measurements]

Measuring an input signal (or noise) with Ch1 and 2 of 35670A simultaneously and make it calculate the ratio Ch2/Ch1. A signal input is divided by a T-shape stuff and input into each channel.

• (a) Freq. response of Ch2/CH1 are measured with a random noise input created by the 35670A itself.
• (b) Swept sine mode to measure the same thing.

For the both cases, Ch2 is set to AC float and Ch1 to DC float. (I've tried other combination for just a confirmation.)

[Results]

The results of (a) and (b) are consistent, and  the fit transfer function is approximated by

f(x) = (i*x/x0)/(1+i*x/x0)

and

 x0 -> 0.485(5),

where I just assume f(x) -> 1 as x -> infinity.

You can use this function for the calbration when you use this 35670A with the AC coupling input (CH2).

Note: I'd like to avoid SR560s for the OSEM screening setup, as it appears to have large noise comparing to what I want to measure in the low frequency region, and each SR560 looks to have different noise level one by one... maybe our SR560s are getting old since they had been bought in 1998; one of them even sometimes has a jump in the DC offset during the measurement.

Strange AC noise

In order to check if the spiky noise decreases after placing the clean booth,  long time acquisition of AC noise was performed (1 hour).

I couldn't find spikes, but we see a strange beaviour: the noise increase with time to a very high level (1mV std). Then another acquisition of 5 minutes was performed and it can be seen the noise suddenly decrease alone at a certain point. I'm trying to find a reason to this strange behaviour. The booth fan are ON, I think they never stop alone. Maybe some fan of the air conditioning or other noisy devices, I don't know.

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

I used a vacuum cleaner to clean the floor inside the clean booth and around it. The floor is not smooth, it's not easy to remove all the dust. 

Clean booth over the optical table

Members: Manuel, D.Tatsumi,

we clean up the optical table, remove unnecessary things, clean all cables, stick labels on cables, clean the floor.

We moved the clean booth from the upper floor to the optical table downstairs. And we moved the wheel desk from downstair to upstair.

OSEM LED screening preparation

LEDs of OSEMs should be screened by measuring their intensity noise beforehand, so a setup for this business is prepared.

It is a quite simple setup holding a LED and PD closely with Thorlabs jigs (see pic).

For the driver for the LED and PD, today I just use the Okutomi-kun's circuit (made based on the actual OSEM satellite box design, but a little bit modified).

• PD: Hamamatsu S1223-01
• LED: OP232 (I'd like to change this to TSTS7100...)

When the LED is turned on, the DC ouput of the PD driver is 1.5 V; the transimpedance is 38.3k Ohm, so the photocurrent would be about 40uA. According to the specification sheet, the PD would have about 0.55 A/W at the wavelenth (OP232's light is 890nm), so the corresponding light flux collected by the PD would be around 72uW. Probably the current supply to the LED became too narrow in the Okutomi-kun's circuit (which was Akutsu's direction, sorry). The power supply of the driver is +/-14V, so even if the LED become brighter, the PD can deal with it.

Anyway, firstly I just setup the FFT analyzer (Agilent 35670A) so that it has DC coupling (the "range" was around 1.5V), but the dynamic range of the measurement system gets ruined due to the large DC voltage described above (see the graph attached).

Then today I simply set AC coupling and then the "range" can be reduced to around 12mV, and the measurements appear to be meaningful so far. As I'm not sure well what will happen in the lower frequency range when you select the AC coupling mode, so it would be better to use SR560 to cut DC signal because we can measure the transfer function and can be calbirated afterwards.

AC noise with Function Generator

I measured again the noise of the photodetector. The chopper is off and the lockin trigger comes from a function generator set at 430Hz. In order to check if the noise could come from the function generator, I measured the noise with the probe laser ON and with the probe laser OFF.

The AC noise std with the Probe ON is 18microV as usual.

The AC noise std with the Probe OFF is 0.7microV. Very similar to the case Chopper ON - Probe OFF measured before (0.8microV).

This means that the noise doesn't come from the function generator.

Other possible noise sources could be the HeNe power fluctuation or jitter position fluctuation due to some vibrations.

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

During many noise acquisitions I noticed that sometimes (one time in 3 or 4 minutes) there is a strong peak of noise, about 2  order of magnitude higher than the usual noise. The system manual speaks about them and give this explanation: "There are various sources of spiky noise including dust particles crossing the probe beam path." 

AEL: Satellite box preliminary test

A satellite box prototype was delivered to NAOJ last week, and was insitu revised/modified, having discussions with AEL group.

So the prototype circuit in NAOJ is not the final version; some capacitaces are removed due to their wrong assembly directions, D101 diodes are removed due to similar reason, and a Q101 transistor of Ch1 is removed during the investigation of the circuits.

Anyway, a basic measurement, the noise level of the output of the circuit is preliminarily measured. A PD (S1223-01) in a old design holder is connected to the circuit with a barrack conversion board, and the output's differential outputs are directly connected to a FFT analyzer (Agilent 35670A) with a front end setup of DC float.

The measurement are done during the room lights are turned off as much as possible, and the PD is covered by an easy carton box. So far, for some reason (probably large 50Hz signal), the FFT's range should be set no less than 158 mV, and the most of the measured data are buried under the measurement system noise.

Anyway, the vertical axis is converted to input equivalent current noise density; divided by 2 due to its differential output, and divided by 38.3k Ohm of the transimpedance; see http://gwdoc.icrr.u-tokyo.ac.jp/cgi-bin/private/DocDB/ShowDocument?docid=3499

Note that this is not a test for OSEM but just a basic measurement of the circuit itself. To say something about OSEM, the output should be calibrated to the unit of "m/rtHz".

I'm wondering why the 50Hz apperas such large in the circuit; comparing to the measurement result of an oplev driver so far, it appears a little bit too large...

Calibration history

Members: Manuel, D.Tatsumi.

We made the calibration of  bulk and surface reference samples, in three different days.

For the bulk calibration, we found a difference of about 2% among the days.

For the surface calibration, we found a difference of about 10% among the days.

The plots show the AC/DC ratio during the scan along the z axis.

AC noise

Adjustment of breadboard in MCE

Member: R. Takahashi, N. Ohishi, K. Doi

June 11
1) Clean the floor of clean booth.
2) Open the chamber.
3) Adjust the height of breadboard. The height was set to 490mm +/- 1mm from the base plate.
4) Measure the position of the breadboard by the straight edge and the pendulum. The position was deviated by -1.0mm, -1.75mm from the center of the chamber.

June 12
1) Measure the rotation of the breadboard by the straight edge and the pendulum. The position of support bolt was on the Y line within +/- 0.5mm.
2) Pick up some parts from the bottom of chamber. We found M10 long bolt, Al foil, I-bolt, and CCP base.
3) Close the chamber.

AC noise -Lockin Time constant; Photodetector noise; Pump power needed

I acquired 3 minutes of the AC noise in the following conditions;

 - pump off

 - probe on

 - front door closed

  - various lock in integration time constants;

As shown in the plot , the higher the time constant, the lower the noise.

But we have to choose it depending on the total acquisition time, for example we cannot spend a week to acquire a surface map.

I report the values:

timeconstant (ms)	Noise std (microV)
10	47
30	14
100	18
300	9.5
1000	5.2

We remind the specification of the system says between 5 and 25 microV at time constant: 100ms; and chopper frequency: 380Hz.

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

Then we switch off the HeNe probe and measure the noise of the photodetector with 100ms time constant.

The photodetector noise standard deviation is 0.8 microV (see the histogram below) and it includes the detector noise and the lockin noise

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

For absorption of 22% we have about 350mV AC signal at 30mW pump power;

The signal should be at least 2 or 3 times the noise (18microV * 3);

For 12ppm we expect 19microV of AC signal at the same pump power (30mW)

This means that a pump power of 100mW might be sufficient to see a good signal for 12ppm. It follows that we need 10W to see absorptions of 0.1ppm

AC noise - Chopper frequency

I tuned the lockin gain to have a better resolution for little signals.

I acquired 3 minutes of noise for different modulation frequencies, in the following conditions:

 - Probe ON;

 - Pump OFF;

 - Front door closed;

 - Lockin time constant 100ms;

 

I made an histogram for every acquisition and I computed the standard deviation in order to find the best chopper frequency.

The lowet value of the noise standard deviation is at 430 Hz modulation frequency. The second best value is for 480 Hz, as used previously. The highest noises are at multiples of 50 Hz, as expected.

Oplev's pillar test installation to the tunnel

Abstract: we did a test installation of an oplev's pillar to the tunnel.

Workers: Uchiyama and Akutsu

The test installation has been done around (1) PR3 chamber, (2) IFI chamber (for IMMT2 mirror), and (3) MCF chamber (for MCo mirror).

For all cases, the bolt holes on the floor are all dirty, and the M24x130 bolts are hard to fit to them without cleaning the bolt holes; even after the cleaning, it was so hard,

so ratchet socket wrenches are used to screw. Actually we didn't have M24 washers today, so the legs have not been fully fixed.

 (1) for PR3,

It is known that the pillar and the chamber's hinge will mechanically interfer, and the inconvenient fact was confirmed. We just slide the pillar a little from the planned location, and found a so-so place, but not know wether the optical table on the pillar can be installed or not. When the pillar stands, the top surface is almost horizontal without shims (shown by a whater leveler), and this means the floor surface here and the pillar's upper and lower surfaces are well manufactured.

(2) for IMMT2 mirror

Not much mechanical intereferences, but one concern is space between a barrelhead attached to a gate valve between MCF and IFI chambers. Still the bolts were hard. When the pillar stands, the top surface is almost horizontal without shims (shown by a whater leveler), and this means the floor surface here and the pillar's upper and lower surfaces are well manufactured. The height from the top surface to the beam line is measured to be 112.5 mm (the beam line height is just estimated from the chamber structure). Then the oplev beam height from the optical table on the pillar can be estimated 112.5 - 27 = 85.5 mm. The expected beam height was 85.85~85.9 mm, so it could be in the tolerant range.

(3) for MCo mirror

Not much mechanical intereferece, but one concern is space between a reducing flange and the pillar. Still the bolts were hard. Unfortunately, the floor surface for the pillar is not so good, and we need two pieces of shims to make the top surface horizontal. The height from the top surface to the beam line is measured to be 115 mm (the beam line height is just estimated from the chamber structure), which should be the same as the one arund IFI chamber, but now it is obvious to have 2.5 mm difference. This difference would be a little bit large if you take into consderaton about measurement errors. Then the oplev beam height from the optical table on the pillar can be estimated 115 - 27 = 88 mm. The expected beam height was 85.85~85.9 mm, so it could be a little bit larger than my expectation, but can be tuned.

PSD calculations

I got the two-dimensional map of the BS from Hirose-san.
I will try to implement these data into the "LightTools" software in the next days.

At the same time, I tried to fit the one-dimensional PSD with the K-correlation model with only limited success.
It seems that this particular model is not suitable for the data...or the data aren't enough.
From the map it should also be possible to calculate the two-dimensional PSD and to fit these data instead of the one-dimensional one.
I would like to try these calculations also during the next days to see how far I can use the PSD for the scattering Paper I am now writing.

Updated JASMINE scatterometer

The laser for the JASMINE scatterometer is now focussed on the sample.
I used a lens with f=200mm; the size of the focal point is now 0.25 - 0.3 mm.

Measurements can again be taken.

I will do some test measurements tomorrow on a Titanium sample.

Flexi circuits

Finally Tatsumi-san gets the (usable) flexi circuit prototypes for OSEMs, and he gave them to me. I firstly check whether the soldering can be done, and the results so far are good.

Assy of Oplev's box prototype

An oplev's covering box prototype is assembled. Needs more additive modifications.

AC noise

 I measured the AC signal to estimate the noise in the following conditions:

No sample installed ;

Chopper frequency: 478 Hz ;

Lock-in time costant: 100 ms ;

Tuned to maximum DC signal ;

Front door closed ;

Motors moving but time wait 500 ms ;

 

Measure are taken in 3 different situations

 - Pump open at 30mW --> figure empty20150612.jpg ;

 - Pump  closed --> figure emptyclosed20150612.jpg ;

 - My voice singing around 478 Hz like here --> figure manuelvoice478hz.jpg ;

I measured 60 microV with silent, and 2.2 mV with acustic noise at chopper frequency.

 

The manual  says, at 380 Hz and with motors not moving, the noise is between 5 and 25 microV.

I will make measurement for many chopper frequencies and look for the minimum.

Spectrophotometry of two reference samples

By using spectrophotometer (Shimadzu SolidSpec 3700) at NAOJ ATC

transmissivity and reflectivity of the reference samples were measured.

From these measurements, absorption value can be obtained.

SURFACE SAMPLE:    Newport FRQ-ND02          22.0%

BULK SAMPLE:            Schott glass  NG-12         37.5%     ---> 37.5% / 0.36 cm = 104% /cm

Both of values are almost consistent with catalog spec. (22.2% and 116% /cm).

Glueing test of MasterBond EP30-2

I did the following things.

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

[ Mixuture ]

EP30-2 (A) : EP30-2 (B) : borosilica = 10 : 1 : 0.55 = 5 g : 0.5 g : 0.275 g

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

[ Cure schedule ]

24 - 48 hours        at 75 Fdeg = 21 Cdeg

2.0 - 2.5 hours      at 200 Fdeg = 93 Cdeg

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

Test 1: Cure Testing

Provide heating by hot plate. Set the temperature to be 90 Cdeg (~200 Fdeg).

Leave the glue for 15 minutes. After the heating, cool it for 1 minute.

Because it is smooth and hard, it is a well-mixed batch.

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

Test 2: Glue flag and flag base

After the test 1, I made glueing two sets of (dummy) flag and flag base.

One is heated at 90 Cdeg for 3.5 hours. Another is placed at room temperature (~24 Cdeg).

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

iKAGRA AS port BRT design

Uploaded a draft of the design document for the iKAGRA AS port BRT (in Japanese)

http://gwdoc.icrr.u-tokyo.ac.jp/cgi-bin/private/DocDB/ShowDocument?docid=3624

I and Michimura-kun also had a skype meeting on how to take care this task...