From tigwiki

[edit] Introduction

Each TIGRESS Clover has an associated Compton Suppression Shield. Each suppression shield set is comprised of three subsystems;

• CsI(Tl) Back Catcher Two components each with 2 optically separated crystals. Two Photomultiplier tubes per crystal.

• BGO Side Catcher One large component consisting of 4 parts, each with 2 optically separated crystals. Two Photomultiplier tubes per crystal.

• BGO Front Shield Four separate components each with 2 optically separated crystals. Two Photomultiplier tubes per crystal.

Various tests are carried out on the components to measure the performance and consistency of the set.

[edit] Performance and Acceptance Tests

[edit] Measuring the Gain of each Photomultiplier tube Signal

This test measures the amplitude of the raw signal coming out of the detector at two different Bias voltages. The measurements are done with a ¹³⁷Cs source and an oscilloscope.

Example of a good Gain test signal‎

• Apply a bias of 850V to each of the four photomultiplier tubes in one unit. See #How to Apply Bias to Suppression Shield PMTs.
• Measure each signal individually by plugging the cable directly into the oscilloscope. You will need a button connector.
• Set the oscilloscope to have a 1MΩ termination and to trigger on the correct channel with a negative slope in Normal mode. The x and y ranges should be around 100mV and 10μs respectively. Adjust the trigger level so the trigger is only on the full photo-peak trace and not Compton scattering.
• Check the signal is not too noisy before putting the source by the detector. Black tape must be used to cover the mounting screw holes on the BGO Side Catcher or the signals are very noisy.
• Place a 1μCi ¹³⁷Cs source in view of the detector and quite close so signals from 662keV gamma rays dominate the output.
• A good, clean signal should look like the image to the right. See also #Sources of Noise.
• Record the Amplitude of the brightest trace on the scope. Adjusting the Intensity knob can help. Also record the trigger level.
• Repeat this for all tubes at 850V bias and then at 1200V bias.
• Note also the Gain which is equal to the amplitude at 1200V divided by the amplitude at 850V.
• A test sheet is available here Media:BGO_Gains_Test.pdf

[edit] Voltage Matching

It is necessary for all further tests to have the output signal from each tube to be equal. Therefore the signals must be gain-matched by varying the bias applied to the tube. This is done roughly with an oscilloscope and then checked/refined using spectra generated by using the data acquisition or MCA.

Example of a good Voltage Matching signal‎

• Investigate one signal at a time by plugging the signal cable into a preamp. Never unplug/plug-in cables while preamp power or tube Bias is ON or the preamp is likely to blow! Easiest is to turn off tube bias and the camac crate from which the preamp power is supplied while changing cables. BGO Side Catchers and CsI(Tl) Back Catchers use 25kΩ preamps and the BGO Front Shields use 50kΩ preamps.
• Apply an initial bias of 850V to each of the four photomultiplier tubes in one unit. See #How to Apply Bias to Suppression Shield PMTs.
• Plug the preamp output signal into the oscilloscope.
• Apply power to the preamp unit.
• Set the oscilloscope to have a 50Ω termination and to trigger on the correct channel with a positive slope in Normal mode. The x and y ranges should be around 100mV and 2μs respectively. Adjust the trigger level so the trigger is only on the full photo-peak trace and not Compton scattering.
• The preamp will induce a DC offset of around -138mV for 25k Ohm and -100mV for 50kΩ.
• Place a 1μCi ¹³⁷Cs source in view of the detector and quite close so signals from 662keV gamma rays dominate the output.
• A good, clean signal should look like the image to the right.
• Examine the Amplitude of the brightest trace on the scope. Adjusting the Intensity knob can help.
• Adjust the bias voltage until the signal amplitude is +330mV.
• Record the Required Bias Voltage for each tube. Use this voltage for all further tests. The gain-matching will be checked/refined during the next test, #Energy Resolution
• Alternatively, to match the bias voltages connect the preamp output to the ADC. Use the 25kΩ preamps for the Back Catchers and BGO Side Catchers, and use the 50kΩ preamps for the BGO Front Shields. Adjust the bias voltages of each photomultiplier tube in turn until the centroid of the ¹³⁷Cs 662keV peak is aligned to a common channel number on Maestro. Set course gain to 5 and shaping time to 3μs then the required bias voltages to align the peak to channel 670 will be around 1000V. Record the Required Bias Voltage for each tube.
• A test sheet is available here Media:BGO_Voltage_Match.pdf

[edit] Energy Resolution

• Investigate each crystal by plugging the signals from both tubes on a crystal into the input for the same preamp. Never unplug/plug-in cables while preamp power or tube Bias is ON or the preamp is likely to blow! Easiest is to turn off tube bias and the camac crate from which the preamp power is supplied while changing cables. BGO Side Catchers and CsI(Tl) Back Catchers use 25kΩ preamps and the BGO Front Shields use 50kΩ preamps.
• Apply power to the preamp unit.
• Apply the gain-matched bias voltage to each of the four photomultiplier tubes in one unit. See #Voltage Matching
• Place a 1μCi ¹³⁷Cs source in sight of the detector at a distance so that the count rate is less than (but close to) 1000Hz for each PMT (also 1000Hz combined).
• Plug the preamp output into an Ortec 855 Dual Spec Amp with 3μs shaping time, positive input, course gain set to 4 and fine gain set to 2.5.
• The threshold level of the discriminator can be set quite high for this test.
• In the web browser start a run.
• Count until there are more than 10⁵ counts in the 662keV peak.
• Record the Centroid and FWHM for the 662keV peak. See #Analyzing a Spectrum
• Calculate the Resolution of the detector. The acceptable resolution is less than 150keV for 662keV gamma rays which is less than 22.7%.
• A test sheet is available here Media:BGO_Resolution_Test.pdf

[edit] Noise Level

Example of a noise test BGO spectrum with no source‎

Closer view of same spectrum at noise level. 70 counts/channel noise level is specific to this example‎

• Investigate two crystals at a time by plugging each set of two signal cables from both PMTs into the same preamp (so two preamps are used, one for each crystal). Never unplug/plug-in cables while preamp power or tube Bias is ON or the preamp is likely to blow! Easiest is to turn off tube bias and the camac crate from which the preamp power is supplied while changing cables. BGO Side Catchers and CsI(Tl) Back Catchers use 25kΩ preamps and the BGO Front Shields use 50kΩ preamps.
• Plug the preamp output signals into an Ortec 855 Dual Spec Amp with 3μs shaping time, positive input, course gain set to 4 and fine gain set to 12.5. These will be the first two spectra, E00 and E01.
• It is important to use the higher gain setting for the noise test or the binning effects in the spectrum changes the noise level measured!!
• Use a BH-1 Tail Pulse Generator in the third adc channel which can also generate triggers. This pulser will be used to measure the live time of the system. Set the pulser to: single pulse mode, 10Hz, 0.1μs risetime, 0.1μs fall time, x10 and x100 attenuators in the down position, amplitude set to 1.5. 10Hz is set when the outer black dial points to 100Hz and the inner dial is rotated anti-clockwise as far as possible. The pulser peak will appear around channel 300 of the third energy spectrum, E02.
• Apply power to the preamp unit.
• Apply the gain-matched bias voltage to each of the four photomultiplier tubes in one unit. See #Voltage Matching
• Place a 1μCi ²⁴¹Am source in view of the detector.
• Lower the discriminator threshold level as low as it will go. Do this by rotating the threshold screw anti-clockwise until there is a clicking sound.
• In the web browser start a run.
• Count until there are more than 10⁵ counts in the 59.5keV peak then stop the run.
• Use the centroid of this peak (See #Analyzing a Spectrum) to calibrate the spectrum. Assume linear relationship between gamma energy and channel number with the channel 21 noise peak representing the zero-energy offset (this peak appears because the adc is read out when a trigger is generated by a signal in the other crystal.) The keV per channel is required to determine the noise level.
• Remove the source from sight of the detector - Do not change anything else!.
• In the web browser start a run to measure the noise level. Count until there are over 1000 counts in the pulser peak and then stop the run.
• Calculate the number of counts per channel that corresponds to a noise level of 10 counts/s/keV using:

• Looking from the left of the spectrum determine the channel where the noise level has dropped to this level, below 10 counts per second per keV.
• Record this Noise level in keV. The required noise level should be less than 15keV.

[edit] Position Sensitivity

• Investigate one crystal at a time by plugging the signals from both tubes of that crystal into the input for the same preamp. Never unplug/plug-in cables while preamp power or tube Bias is ON or the preamp is likely to blow! Easiest is to turn off tube bias and the camac crate from which the preamp power is supplied while changing cables. BGO Side Catchers and CsI(Tl) Back Catchers use 25kΩ preamps and the BGO Front Shields use 50kΩ preamps.

Positions of the source for each crystal

• Apply power to the preamp unit.
• Apply the gain-matched bias voltage to each of the four photomultiplier tubes in one unit. See #Voltage Matching
• Plug the preamp output into an Ortec 855 Dual Spec Amp with 3μs shaping time, positive input, course gain set to 4 and fine gain set to 2.5.
• Place a 1μCi ¹³⁷Cs source in the first location shown in the diagram. Use a 1cm thick piece of foam to lift the source away from the detector surface and reduce the count rate a little.
• The threshold level of the discriminator can be set quite high for this test - we are only interested in the 662keV peak.
• In the web browser start a run.
• Count until there are more than 10⁵ counts in the 662keV peak.
• Record the Centroid of the 662keV peak. See #Analyzing a Spectrum
• Now repeat for all the other locations shown on the diagram. There are 5 positions for each back catcher crystal, 12 positions for each side catcher crystal and 10 positions for each front shield crystal.
• If the spectra are to be analyzed in gf3 using the BGO_test_analysis.c program then the runs should be performed in the order: segment A then B for Back Catchers, Side Catchers then Front Shields, in the order of the serials. All readings should be made in the E00 spectrum (cable BGO1.) It should be neccessary to input the first run number only into the BGO_test_analysis.c code.
• The "reference position" is position 1, 7 and 4 for the back catcher, side catcher and front shield respectively.
• The results reported will be the average and maximum percentage difference of the 662keV photopeak centroid of the different positions when compared to the centroid of the reference position.
• The average percentage difference is equal to where μ_i and μ_ref are the centroids from the source placed at the i^th position and the reference position respectively.
• The centroid should not shift by more than 10% for any position.

[edit] Count Rate Dependence

• Investigate one crystal at a time by plugging the signal cable from both PMTs into the same preamp. Never unplug/plug-in cables while preamp power or tube Bias is ON or the preamp is likely to blow! Easiest is to turn off tube bias and the camac crate from which the preamp power is supplied while changing cables. BGO Side Catchers and CsI(Tl) Back Catchers use 25kΩ preamps and the BGO Front Shields use 50kΩ preamps.
• Plug the preamp output into an Ortec 855 Dual Spec Amp with 3μs shaping time, positive input, course gain set to 4 and fine gain set to 2.5.
• Apply power to the preamp unit.
• Apply the gain-matched bias voltage to each of the four photomultiplier tubes in one unit. See #Voltage Matching
• Place a 10μCi ¹³⁷Cs source in sight of the detector at a distance so that the count rate is close to 20 kHz.
• In the web browser start a run.
• Count until there are more than 5x10⁴ counts in the 662keV peak then stop the run.
• Record the Centroid of the peak. See #Analyzing a Spectrum.
• The 1kHz data for comparison was taken during the #Energy Resolution test.
• The percentage difference between the centroid of the 1kHz (μ_1kHz) run and the 20kHz (μ_20kHz) run is equal to
• The centroid should not shift by more than 10% for any position.

[edit] Bleeding Resistance

• Take a Digital VoltMeter (DVM), turn it on and set it to measure resistance (The Ω symbol).
• No cable detectors should be connected to anything during this test.
• Hold the black probe of the DVM to the outer casing of one HV connector.
• Touch the central pin of the same HV connector with the red probe and record the restance displayed on the screen.
• The values will be around 18-20MΩ for the SideCatchers and BackCathers and around 11MΩ for the Front Shields.
• Repeat for all PMT tubes.

[edit] Other Useful Stuff

[edit] Setting up the Electronics for Suppression Testing

Basic configuration of electronics for BGO testing.‎

• Before connecting or disconnecting any PMT signal cables to the preamps make sure no voltage is applied to the PMTs and no power is supplied to the preamps. The preamp supply is from the back of the carmac crate (2nd from top) so make sure the crate is turned off at the front. Never unplug/plug-in cables while preamp power or tube Bias is ON or the preamp is likely to blow!
• The detector cables go into a preamp box. 50kHz preamp for the BGO front shields and 25kHz for the BGO Sidecatcher and CSI(Tl) Back Catcher. The cables from both the central and lateral photomultiplier tubes of a single segment (AC amd AL or BC and BL) go through the same preamp.
• Each preamp signal goes into a Spec Amp. The output from the Spec Amp goes to the Phillips 7164 ADC. The signal from the Spec Amp for the 'BGO1' cable goes to channel 1 of the ADC and the amplified signal from cable 'BGO2' goes to channel 2.
• For the ADC gate, each preamp signal also goes into a Quad TFA, the first stage in shaping the gate pulse.
• The Quad TFA output goes into a discriminator. The width and threshold of the output pulse can be ajusted on the front panel with a pole zero screwdriver. Connect the discriminator output and the Spec Amp output to a scope and use the discriminator pulse as a trigger to see the low energy pulses (noise) accepted by the threshold level.
• The discriminator output pulse goes into a GG8000 Octalgate Generator for further shaping. Use the output as the trigger on a scope with the Spec Amp output to check the gate fits the signal pulse and will be recognised by the ADC. The gate for the Phillips 7164 ADC should be 50ns to 50μs wide and should precede the voltage peak by 20ns minimum. If needed the delay and width of the gate can be adjusted on the front panel of the Octalgate Generator using a pole zero screwdriver.
• Each gate output (BGO1 and BGO2 and pulse gate) goes into (LeCroy 428F) linear fan-in/fan-out unit. The output is used as the gate for the ADC.
• The BH-1 Tail Pulse Generator can be used to measure the live time of the system. Set the pulser to: single pulse mode, 10Hz, 0.1μs risetime, 0.1μs fall time, x10 and x100 attenuators in the down position, amplitude set to 1.5. 10Hz is set when the outer black dial points to 100Hz and the inner dial is rotated anti-clockwise as far as possible.
• The pulse generator has two outputs. The Pulse Output goes through a Spec Amp into the ADC 3rd channel. The Trigger Output goes through the Quad TFA, discriminator, Octalgate Generator shaping process described above into the fan-in/fan-out unit as the ADC gate. The pulse generator spectrum should consist of a sharp peak at around channel 300 in spec# E02. You might need to adjust the gain on the Spec Amp to achieve this.
• Use a ratemeter (LeCroy 4604) to measure the frequency of the signals accepted by the discriminator (use discriminator output as the ratemeter input.) Set the timer to 1ms. For a quick indication of the rate use a preset of 10³ (1s) and for a more reliable rate value use a preset if 10⁴ or 10⁵.
• Make sure any loose connections are terminated.

[edit] Starting the Acquisition for Suppression Testing

1. Log into midtig01 as bgo_fe with the usual password.
2. Open a terminal with 5 tabs. Name the tabs with the following names and perform the following operations in the appropriate tab in the following order.
   1. analyzer tab:
      1. cd bgo_fe
      2. ./analyzer
   2. daemons tab:
      1. mlogger -D
      2. mhttpd -p 8081 -D
      3. mserver -D
   3. frontend tab:
      1. minicom
      2. Some text will appear in the terminal
   4. roody tab:
      1. roody
      2. Right-click on the online folder which appears in the roody window.
      3. Enter midtig01:9090 as the hostname
   5. HV supply tab:
      1. telnet tighv03 1527
      2. Login as admin
3. Open a web browser
   1. Enter the path midtig01:8081
   2. Select the bgotest link
   3. Click on ODB at the top
   4. Click on Logger
   5. Check that the Data dir path exists and is writable otherwise run files will not be saved. If the files are not being saved the only warning is an error message in the analyzer tab and can easily be missed!

[edit] Data-Taking Proceedure and Streamlining the Process

[edit] Analyzing a Spectrum

[edit] Fitting a Peak in Roody

• These results are not trustworthy
• Open the spectrum in Roody.
• Increase the size of the window so that it is at least half the height of the screen. This way you will be able to read the small numbers of the fit before Roody crashes.
• Zoom in around the peak to be fitted. Zoom in by moving the curser over the numbers along the x-axis until the curser changes to a hand. Click and hold the left mouse button for the region you want to zoom in on.
• In the main Roody window open the Refresh menu at the top and ensure refresh is set to off or roody will crash.
• Move the curser over the data in the spectrum window so that the curser changes from a cross to an arrow. Right-Click to bring up a menu.
• Select FitPanel on the menu to bring up a new window.
• Click the Fit button in the new window. A blue fit line should appear in the spectrum.
• Click the Close button in the fit window (or Roody will crash!!).
• The parameters of the fit are displayed in the top right of the spectrum window. Alternatively they are printed in the roody terminal tab
• Mean is the centroid of the fit.
• The FWHM is related to sigma by: FWHM=sigma*(2*sqrt(2*ln(2))) = sigma*2.3548
• The Resolution is equal to the FWHM divided by the Centroid.
• Close the spectrum display window before opening another spectrum or calling the fit program again, or roody will crash.
• Roody will probably now crash.

[edit] Fitting a Peak in ROOT

• These results are not trustworthy
• First, you need to determine the name of the spectrum you want to analyze and the name of the spectrum within that file. The spectrum will have a name like Sum00.
• The file name will look something like /tig/midtig01_data2/tigress/his00059.root.
• The path name is set by a variable in the MIDAS ODB, /Logger/Data dir.
• Note that this is the path name on he logging computer. If you are trying to analyze the data on a different computer you may need to use a full NFS path; e.g., /data2/tigress/*.root will work on midtig01, but you will have to use /tig/midtig01_data2/tigress/*.root if you are on mother8pi. The file with the spectra is the ROOT histogram file,

his00059.root. The 00059 part is the run number. The method described below cannot work on the run in progress, so you must end the run and then analyze the data.

• Next, in a terminal window, go to the BGO testing directory, cd ~/bgo_fe.
• Start up root. Load in the fitting routine by typing .L fittest.C.
• Declare a histogram and fit-function object by typing TH1F* h and TF1* g .
• Open the file of interest by typing in a command like TFile* F = TFile::Open("/tig/midtig01_data2/tigress/his00059.root"). Replace the part in quotes with the name of the file with the spectrum you want to analyze. Load all the files into memory by typing F->ReadAll() .
• Next, to pull out and display a specific spectrum, type in something like h=(TH1F*)gROOT->FindObject("Sum00"). Replace Sum00 with the name of the spectrum you want. Display the spectrum by typing h->Draw().
• You are now ready to fit. Zoom in on the fitting range by click-and-drag on the horizontal axis. Then type g=fittest(h).
• The fit will automatically overlay on the spectrum and the parameters will be printed out in the terminal window. The parameters, in order, are the height, centroid and sigma of the best-fit Gaussian, followed by the abc of a polynomial in a background. The fit is overlaid on the spectrum, showing the Gaussian peak in green, the polynomial background in blue, and the sum in red.

[edit] Fitting a Peak in Radware's GF3

• First we need to convert the root spectrum into a radware spectrum. There is a program to do this called root2spe and if you are logged onto midtig01 as tigbgo then the source code is located at /home/tigress/root2spe
• Each root tree (file with extension .root) contains many spectra. root2spe will extract one spectrum from the root tree file and output it as a 1d radware spectrum (extension .spe)
• The command to extract the spectrum E00 from the root tree example.root into a radware spectrum called example.spe would be: root2spe example.root E00 example.spe
• To launch GF3: gf3
• Press enter once
• To open a spectrum: sp example.spe
• To zoom in: ex then with the mouse click the left limit and right limit.
• To make a fit: nf then with the mouse click the left limit and right limit and the centroid location. Press enter twice.
• To exit: st
• For a more detailed guide of gf3 see http://radware.phy.ornl.gov/gf3/gf3.html

[edit] How to Apply Bias to Suppression Shield PMTs

This guide is for the TIGRESS Suppression Shield Testing Setup in room 156.

1. Ensure the large red HV cable is connected to both the HV crate in the blue rack and the BGO HV distribution box.
2. Connect the four HV cables from the detector to the channels in the BGO HV dictribution box.
3. Open a terminal or go to the HV Supply tab in the main aquisition terminal.
4. Type telnet tighv03 1527
5. Login as admin
6. Press the down arrow to bring up the main menu
7. Select Channels
8. The slot for the BGOs starts in channel 13 with the Primary voltage for the slot. The primary voltage should be 1500V.
9. The following channels are associated with the outputs on the BGO HV distribution box.
10. To change a voltage value move the black selection curser with the arrow keys to the voltage in the V0Set column you want to change. Type the new value and press return. The new value should be displayed.
11. To turn the voltages on move to the Primary row and to the Pw column. Press the spacebar and the status should change from off to up and then ramp to the voltage set and display on.
12. Now move down to a channel and press the spacebar again. All the channels should then begin to power up.
13. To power down press the spacebar again.
14. To move to the main menu press tab.

[edit] Sources of Noise

• A Ground Loop can cause noise. In the #Measuring the Gain of each Photomultiplier tube Signal test a ground loop noisy signal looks like

Example of a signal with ground loop noise

• It is caused by things being plugged into different power supplies and the noise itself looks like this image

Example of ground loop noise