Equipment Operation Instructions

Laser setup

This section is a summary of my notes on commissioning the TCT system. This includes notes on focusing, issues with laser trigger, MIP calibration, the test sensor itself. As the system is still in the process or being comissioned, there are some suggestions of steps to be taken.

There is information about the system on the company website: http://www.particulars.si/

Background

Transient Current Technique, or TCT, is a method to measure the behavior of one type of charge carrier in a silicon detector. By injecting a specific wavelength of laser, charge carriers will be formed up to a specific depth. For example, a 660nm laser will only penetrate a few microns in silicon, the charge carriers that flow towards the injection side (assuming it is the top or bottom) of the silicon will have a much shorter path length and therefore collection time. The signal of the solely the opposite charge carrier will be observable after that time. The injection can also be made incident to the edge of a detector to make measurements at a specific depth. A 1060nm laser penetrates 1000um microns of silicon, and if properly pulsed, can simulate a minimum ionizing particle (MIP) and characterize detectors.

In order to do TCT scans, BILPA has acquired a Large Scanning TCT system from Particulars. The system is equipped with two wavelenghts of laser diode (1060nm and 680nm), 3-dimensional motion stages with micron precision, a spaghetti diode and readout system consisting of a bias-T and external amplifier, and DRS4 oscilloscope. Particulars also provides software to run the scanning, oscilloscope, and laser simultaneously, and to record data. Properly set up and used, one can use the system to make two-dimensional scans, including edge-TCT.

Currently the system can be focused and used to complete 2 dimensional scans, however there are a number of issues that lead to make the overall behavior relatively sporadic. Most importantly, as of now I have not been able to measure a signal with the 660nm laser, though I have seen the beam.

Safety

Our laser is considered 'Class 1' (like a Bluray player) since it is enclosed. Safety glasses are not necessary. If the interlock system is compromised, the laser is class 3B and requires the use of safety glasses.

Components

The TCT system includes the following components:

- a large aluminum box with a swinging door, laser interlock, cable connections and plumbing hookups to serve as a faraday cage and darkbox.

- if the interlock is broken the laser will not be triggered; interlock is broken when the door is opened

- laser safety glasses. these should not be necessary due to the interlock. However, if the interlock is for some reason removed from the door and closed, you should wear these.

- beam focuser

- optical density filter

- optical fibers for both 660nm and 1060nm

- bias-T

- external amplifier (can use CIviDec? or Particulars)

- Keithly 2410 high voltage power supply

- TTi MX180TP? Multi-Range DC power supply

- Particulars Low pass filter for high voltage

- Particulars PS-01 L laser power supply

- Particulars laser driver

- DRS4 Oscillpscope

- Pulser

- Laser Safety glasses

- 3 stage controllers

- 3 motion stages with micron precision

- Windows PC

- Particulars PSTCT program

- Particulars Laser Control program

- XILab stage control software

- acrylic attachment to extend stage for larger detectors and/or attached readout systems, such as DECAL.

Spaghetti diode

The sensor provided in this setup is a Type 3 spaghetti diode fabricated by HPK. It is an n-on-p, 300um thick 4x4mm?? diode with 20um strips of metal at an 80um pitch on top of the silicon to match the geometry of ATLAS strip detectors. The depletion voltage is near 90V, and no gain has been observed for these detectors. [https://indico.cern.ch/event/226647/contributions/472812/attachments/373064/519024/SpaghtiDiodes.pdf]

data acquisition

trigger off laser pulse, not signal itself. Ideally the pulse needs to be attenuated, since it is too large for the DRS4 input.

General procedure for data acquisition:

- start PSTCT program

- click right arrow in upper left corner

- set stage control to xi-lab in drop down menu

- set desired wavelength in drop down menu

- set the oscilloscope to DRS4 in drop down menu

- click "init" button

- If successful you should see:

"

"

Open particulars laser control

set desired frequency

set desired pulse width percentage. This controls the amount of injected charge, a higher percentage means less charge

- test stage controllers. note orientation on box and that there is currently a ratio of 2:1 between the distance moved according to the software and that in reality. That is, if you enter 1um into the porgram the stage will really move 2um. Move each stage an amount you can notice visually, say 1000 units. Click on the 'Stage Control' tab of PSTCT program. You must have the oscilloscope off to click this tab enter 1000 into the 'delta x' box and click the right arrow next to it. Then click the left arrow to move back to the original position. The stages should have moved to the right and then to the left if you are looking into the box. If this is not the case, switch the coordinates using the drop-down menu ''. Repeat this procedure to check that +y moves the beam focuser down and and +z moves the sample stage forward

-To trigger the signal, connect the LEMO connector labeled 'TRIGGER OUT' on the RHS faraday cage to Channel 4 of the DRS4 (ideally you would use an attenuator here, since the signal is so large). Connect the 'Amp out' BNC connector on the LHS of the faraday cage to Channel 1 of the DRS4. On the 'Oscilloscope' tab, set the trigger channel to 4 and click the 'Ch4' button under the ' ' heading to display the signal on channel 4. Trigger on the rising edge of the pulse, 100mV should be an above the noise.

- sometimes the program will lose the ability to control the stages. when this happens, reboot the PSTCT program,

-to do a two dimensional scan,

- set the parameters on the 'Motion Parameters' tab

- in the 'DAQ' tab, click 'START'

- follow the prompt to choose a name for the data, then click 'Save'

- wait for the scan to finish, there should be an estimated time of completion. note that the program often freezes, I have never been able to run scan that lasts longer than approximately an hour, and even then only through repeated efforts. I think the issue may be with communicating with the stage controllers. We may want to try a new/better USB hub, or finding a way to connect the stage controllers more directly to the PC.

- if it is a focusing scan, it be analyzed with GetFocus? .C. if it is a surface scan, '' can be used. In principle for this program can be used for Edge TCT scans with slight modifications.

-In principle a 3-d scan can be preformed in this way by setting Nx, Ny, and Nz to greater than 1.

data analysis

I have the 2d-scan program here:

'plot\_scan.cpp'

and the focusing program here:

'GetFocus\_v2.C'

To execute each:

- In terminal, start ROOT (I have been using 6.12)

root -l

- Load the TCTAnalyse library (download)

gSystem->Load("TCTAnalyse.sl");

- execute the program

.x program\_name.C

In the section with the comment

//------ USER INTERVENTION REQUIRED -----------------------------------

//select the basic parameters of the analysis

one can adjust parameters to analyze a specific set of data.

For plot\_scan.cpp, the 'optic\_axis' and 'scanning\_axis' parameters are set to 0 and 2 and should not have to be changed as long as the configuration of the coordinates remains the same (that is, a left handed coordinate system with y vertical and x towards the right if you are looking into the box). The code produces a 2-d heat map and 3-d map of the collected charge. Note that due to the left handed coordinate system, the plot is inverted in z.

For the GetFocus? .C, one should generally only need to adjust the error function fit range, which can be done by changing the values of parameters 'HiLim' and 'LowLim'. One limit of the fit range is automatically the minimum of the scan at one y-value. If z is the scanning axis, set the parameter 'scanning\_axis' to 1. The initial error function fit parameters are hard-coded in the section beneath the comment '//define fit function'. Adjusting these may improve poor fits.

Issues and troubleshooting

stage communication drops for long runs

If you attempt run the scanning program for a long time, the PSTCT program tends to freeze. I believe the longest scan I have run is for about an hour. I theorize that there is an issue communicating with the stages, since the program will freeze sometimes when you are moving the stages but not doing a scan.

The scaling of the stage movements is off by a factor of 2. That is, if you input 1 $\mu$m steps into the PSTCT program, the stage actually moves 2$\mu$m each step.

An important deviation from the standard setup is that I have found the CiviDec? 20dB amplifier to have better signal to noise and fewer issues with reflection than the Particulars 02A amplifier, and so have opted to use that for measurements with the spaghetti diode.

I did investigate what might be wrong with our Particulars amplifier. I first replaced the bias-T, amplifier, and spaghetti diode of our setup with ones from RAL, and continued to observe reflections. I contacted Particulars, and sent the amplifier to be tested. They did not observe reflections in their TCT setup with our amplifier, so something may be wrong with our setup. They also sent back 2 amplifiers which I tested in our setup and still observed reflections. For the time being, the best solution is to use the Cividec amplifier and add about 6m of cable between the detector and bias-T as to delay any reflections, which I have done.

figure of reflections in multiple amplifiers.

non-linearlity in pulse height.

Triggering

I was able to trigger the 660nm and 1060nm lasers with both the Particulars laser control software and using an external pulser. For the external pulser, I found a range of settings that resulted in a pulse being fired. This range was:

I also did some exploration of the parameter space to see if the pulse height had any dependencies. I found none for the values I tested, tabulated in Table?? With the exception

[ table of pulse height vs. paramters]

I found that the following pulser settings were effective for triggering the 660nm laser:

A key issue here is that the spot intensity depends on the positioning of the laser driver, suggesting that there may be a problem with it or the fiber optic cable that connects to the focuser

Focusing

In order to find the focal length of the beam, I followed the focusing procedure using the spaghetti diode prescribed by Particulars [http://www.particulars.si/downloads/Particulars-Procedures-FocusFind.pdf]. That is, for various distances between the laser and sensor, I scanned the beam across one of the strips and measured an average pulse on the DRS4. The focus is presumably near 8.25cm for the Large TCT system, so I centered the scan there and approximated the distance with a ruler. For the focusing, I set the aperture of the beam focuser to be 8mm in diameter. Using ROOT and the TCT Analyse lirbary provided by Particulars, I fit an error function to each side of the resulting trace of the pulse height versus scanning distance at a given focal distance plotted in Figure 1. From the fit, I extracted a sigma under the assumption that the beam is Gaussian. I found the focus to be at approximately Y=12800 and beam spot to have a full width half max of approximately 10 microns. The code to do the focusing is provided here [link to program]. It is a modication of this file provided by particulars.

[attachment for focusing figure]

MIP calibration

In order to calibrate the injected charge to be approximately equal to a minimum ionizing particle (MIP), the output from the spaghetti diode was measured with a Sr90 beta source and the laser control parameters were tuned to approximately match that output. Using a [fast scope], I measured the pulse height and width of the average trace from beta injection at 100V bias to be 2.8mV and 13.4ns, respectively. The data acquisition was self triggered. I tuned the response from the laser to be 4.4mV in height with a 13ns width at 400V detector bias by setting the DAC value to 50\% in the Particulars laser control program and attaching the optical density filter to the end of the beam focuser. While the bias is different in the two measurements, the gain is 1 for the detector at both voltages, so the comparison is valid.

Some caveats for the calibration are:

- The trigger for the Sr90 data is far from ideal, as I triggered on the DUT itself, and there is poor signal to noise. To get a more reliable sampling of the pulses, a secondary trigger would ideally be implemented. At the very least, one could fit a landau to the distribution of pulse heights for the self-triggered data.

- The analysis is crude, as I am just using the trace averaging and cursors on the oscilloscope. Ideally it would be down offline with fits to distributions.

Caveats nonwithstanding, we are likely injecting charge equivalent to the energy deposited by between 1 and 5 MIPs.

Next Steps

Use a basic diode to re-do focusing, mip calibration, check for linearity across the sensor.

Check spaghetti diode, ask Simon about the wire bond, the sensor has moved around quite a bit.

Check components, work with Gregor, RAL.

Minor improvements:

attenuation of pulser output for DRS4 trigger

an IR sensitive camera to see the beam when the door is closed. An inexpensive webcam will probably suffice.

Complete an edge TCT scan.

See signal from 660nm laser, and complete a scan.

Probe Station

Probe Station

Alibava

Alibava is a readout system for micro strip detectors, named after the institutions that developed it: Liverpool. Barcelona, and Valencia. At Birmingham, as of August 28, 2018, we used the system along with a Sr90 radioactive source to characterize the charge collection efficieny (CCE) of strip detectors.

This section is to collection information about the use of the Alibava system an BILPA.

Components

The system consists of a number of components:

  • Fridge
  • Alibava motherboard
  • Alibava daughterboard
  • Sensor
  • 2 Scintillators coupled to photomultiplier tubes (PMTs) (Scintillator "B" typically gives ~5x more output than Scintillator 1)
  • Fan
  • Temerature and humidity sensors readout by Arduino board
  • Aluminum box to house Alibava daughter board, fans, scintillators and PMTs, and temperature and humidity sensors.
  • High voltage power supply: Keithley 2410) - used for sensor
  • Low voltage power supply: TTI EX354RD - used for scintillators
  • A simple breadboard to control the PMT voltages
  • A 2nd low voltage power supply when the breadboard seems to not work (ideally always use this)
  • Oscilloscope - used to monitor scintillators
  • Sr90 beta source
  • Nitrogen gas source in rhs of fridge (coming out of grey block)
  • Smaller rainbow cable, which monitors temperature and humidity, powers the fan and powers the PMTs in the scintillators
  • Signal cables of oscilloscope
  • Power cables for alibava
  • Cable blocks to replace broken ribbon cable parts (don't worry about purple and grey wires hanging out, they do nothing)
  • Large ribbon cable should always be attached
  • Python script lives at ~/temphum/python_temphum*.py, and outputs data to nohup.out and/or output_temp_hum.txt
  • Aluminum tab to block hole to radiation source

A schematic of the setup is here: alibava_schematic.pptx

Alibava Daughter Boards

There are 4 Alibava daughterboards:

  • Two 500V models that we have been able to run pedestals on, labeled ''
  • A 1000V 27nF custom designed board, this is the one currently in use, labeled with '27nF/1000V'
  • A 1000V board we have not been able to communicate with
  • Note that each board may require a different configuration of ribbon cable(s)

Annealing Sensors

  1. ONLY handle sensors with carbon, plastic or vacuum tweezers. NOT metal.
  2. Use oven at back of clean room entrance hallway
  3. Turn on.
  4. 60, 100, 120 and 130 degrees C are marked in pen (60 is just marked with a dot)
  5. Orange HEATER light will run off when the oven is at correct temperature
  6. Add a temperature sensor near the sample in the alibava setup to cross check the oven temperature
  7. Store sensor in small tin foil trays in cardboard box to left of oven.
  8. Sensor is small grey box with T1 and T2 written on it. Can use body temperature to test which is which.
  9. Standard with annealing is 60C for 80 mins.

Data Acquisition Procedure

The procedure to collect data is as follows:

Preliminaries

  1. Simon (or another technician) must wire bond the sensor to the daughter board. Send him an email and it takes him about an hour. Can ask him in person if it's urgent. You must give him board, sensor and a plastic container (bag) for the old sensor.
  2. Label the sensor bags with: sensor version, date, fluence, annealed (if it is).

Preparation of box prior to cooling

  1. If you are doing a cold run, make sure fridge is set to "super", rather than "stop" or "normal". Knob is on right side.
  2. Outside of the fridge add daughterboard to alibava box, making sure pins go in holes to hold it in place. Make sure nitrogen tube, humidity sensor are attached.
  3. Make sure temperature sensor is in heat sync on the back of the daughterboard (tape it down).
  4. Attach high voltage cable to the daughterboard.
  5. Add alibava box to fridge, making sure hole in lid is below source, and aluminum tab is removed.
  6. Connect nitrogen (on RHS on fridge from grey block) to alibava box, and connect nitrogen source (at wall) to cable going into fridge. The nitrogen should always be connected to the box when it's out of the fridge
  7. Connect smaller ribbon cable, making sure XX marks are on same side. This cable monitors temperature and humidity, powers the fan and powers the PMTs in the scintillators
  8. Make sure power supply is plugged into voltage splitter (or power supply). And all labelled cables are in appropriate places in voltage splitter, or power supply. If you are using the power supply be sure you tie the grounds its terminals to the ground for the PMT circuit.
  9. Plug high voltage power supply for sensor bias in. Cables are labelled as Alibava (or capitalised) HV. Make sure red (HV) is connected to red and black (ground) to black.
  10. Attach cable blocks using a screwdriver, which replace broken ribbon cable parts to power the PMTs. Make sure to follow the lables for "+5V", "Vcont", and "Vref."
  11. Attach the LEMO cables for the PMT signals. Cables are both color coded and labeled.
  12. Close the fridge, including the straps on the side. The only cable in the door of the fridge should be the large ribbon cable (unless we have fixed this). All others should go through the hole in the fridge.
  13. Make sure Arduino board (monitors and reads out temperature and humidity) is connected to the PC via USB, and make sure python script is running (~/temphum/python_temphum*.py). If you see sporadic readings for the temperature and humiity, it is probably due the small ribbon cable being disconnected or connected incorrectly.
  14. Make sure power is connected at wall and turned on.
  15. It is advisable to check for PMT signal on the oscilloscope and make sure you can trigger on settings that are comparable to Alibava, that is a 200mV trigger.
  16. Check the connection from the PC to Alibava by doing a pedestal run. Be sure the DAQ software is not running in emulation mode.
  17. If the above two steps are proving difficult, see trouble shooting.
  18. Connect PMT output the to Alibava motherboard. White cable to white strip on motherboard box, and green cable to red strip.
  19. Wait for the temperature to reach approximately -25C. This can take 2-3 hours, but will go faster if the daughterboard and HV supply are turned off as the current flow causes heating.

Data Acquisition with Cold sensor

  1. Bias the sensor by: set compliance to 1 microamp using "EDIT" button, up and down arrows, and then enter.
  2. Put a small bias on the sensor in each direction to check the device polarity. Use which sign of voltage gives the lower current, and apply whatever bias voltage you're testing.
  3. If you have not done so, turn on Alibava.
  4. On the PC, open the program "Alibava" to begin data aquisition.
  5. Do a run in "Pedestal" mode to check that the board and PC can communicate.
  6. Switch to "RS" mode for triggering off a signal from a radioactive source. We typically acquire 100 000 events per bias voltage, which has taken about 8 minurtes a run.
  7. In the settings menu, check that trigger is set to "OR" and that the trigger levels are -200mV.
  8. If you want to collect data, click "Log Data" and write the name of the file you want to save. The current convention, on which the analysis is dependent, is name of the form "RSRun-##C-XXX_###V" where the first ## is the temperature, XXX is the trigger configuration ("OR", "OR-GREEN", or "OR-WHITE"), and the last ### is the bias voltage.
  9. When the run is finished, adjust the bias voltage and start the next run.

Warming Alibava after DAQ

  1. When you are done with a sensor for the time being, you must decide to either leave it in the fridge or remove it. If the sensor is irradiated, you will have to keep it cold after heating and removing.
  2. To remove the board and sensor, open the fridge and quickly undo all the connections possible, this should leave only the large ribbon cable. Remove the nitrogen last. Remove the Al box from the fridge and close the fridge door.
  3. Immediately reconnect the nitrogen from a point further upstream in the line to the box. The flow will be faster here, check that it isn't too much for the sake of wire bonds.
  4. Reconnect the small ribbon cable. There is small length identical to the portion in the hole in the fridge, use that. This allows you to have temperature and humidity monitoring for the box outiside of the fridge.
  5. Wait for the temperature to reach the dew point for the ambient air. For ISO7, the humidity is ~50% and temperature ~22C, giving a dew point of 11.1C (http://andrew.rsmas.miami.edu/bmcnoldy/Humidity.html); 12 C may be safer to guarantee no condensation.
  6. Once the board is warm enough, you may now remove the lid of the box to access the contents to replace the sensor or make repairs/adjustments.

Removing the Daugterboard

To remove the daughterboard from the box in order to, for instance, give the board to a technician (Simon) for wire bonding:

  1. Remove temperature sensor from heat sink
  2. Remove humidity sensor
  3. Disconnect large ribbon cable
  4. Disconnect power supply
  5. Pull nitrogen tube out
  6. If giving to Simon for wire bonding, remove any tape from bottom as it must be level
  7. Reattach everything before the next run, as described above.

Trouble shooting

  • If the GUI for data acquitiion isn't working, close the GUI, power cycle Alibava, and open the GUI. This may take a few iterations.
  • If power cycling/rebootinh the GUI and Alibava doesn't work, the problem may be the wire bonds, particularly on either of the Beetle chips, or the large ribbon cables. You can check the wire bonds on the chips visually and the cable with a ohmeter.
  • If scintillators don't work, check power source, Vreference and Vcontrol (Vreference and Vcontrol should be ~1.1V). Check that current from power source is 5 or 6 mA. If it is 2-3mA, then probably only one PMT is working. You can use as oscilloscope to check the signals as well.
  • If you cannot get PMT bias right with the bread board, try using a power supply instead.
  • If you see strange readings in temperature or humidity, check that ribbon cable is connected correctly
  • For the analysis software, close and open software GUI after recompiling software

Safety and Warnings

There are some warnings for the sake of the system:

  • Avoid condensation on the board and sensor. This should not be an issue when cooling down as the ambient air will probably always be cooler than the sensor. However, when heating up, the opposite will be true. Ideally we would have 2 separate sensors, one for the ambient air and the other for the sensor. If we assume some reasonable maximum temperature difference between the two, say 7C, we can calculate the relative humidity at which condensation will begin as in RH_dT7C.xlsx a la http://andrew.rsmas.miami.edu/bmcnoldy/Humidity.html. As long as we are below 53% RH, we should avoid condensation.
  • Be careful of damaging wire bonds on Alibava. The bonds are on the sensor and the two Beetle chips on the daughterboard. You should be able to spot them with the naked eye, but a microscope can also be helpful. Never allow anything to touch the wire bonds. They will also corrode when wet, anothe complication related to condensation.
  • Never touch sensor (but you can touch the daughter board when the sensor is on it)
  • Ground yourself before touching any electronics (can use metal plate ground point under laser)
  • Note that condensation or frost will gather in fridge when it's turned on with the door open. When it gets too icy, it is good to turn it off, let the ice melt and drain it.
  • Nitrogen should always be connected to Alibava.
  • Keep Alibava box as airtight as possible (mostly using tape).
  • Before opening fridge, turn off power supplies so that nothing gets a large leakage current from light from the room. Also the insulation on the HV line leaves a bit to be desired, staying away from the cables while biased is probably safest.
  • This experiment involves a radioactive source, please keep your radiation training and the principles of ALARA in mind.

-- Patrick Freeman- 28 Aug 2018

-- LauraGonella - 09 Aug 2018

Topic attachments
I Attachment History Action Size Date Who Comment
Xlsxxlsx RH_dT7C.xlsx r1 manage 38.3 K 28 Aug 2018 - 15:14 UnknownUser Relative humidity at dew point for a 7C temperature difference.
Pptxpptx alibava_schematic.pptx r1 manage 37.4 K 28 Aug 2018 - 16:25 UnknownUser Schematic of Alibava DAQ
Edit | Attach | Watch | Print version | History: r5 < r4 < r3 < r2 < r1 | Backlinks | Raw View | Raw edit | More topic actions...
Topic revision: r3 - 31 Aug 2018 - _47C_61UK_47O_61eScience_47OU_61Birmingham_47L_61ParticlePhysics_47CN_61patrick_32freeman?
 
This site is powered by the TWiki collaboration platform Powered by Perl This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback