Description of the Test of the D-Zero Dual FET PreAmps -------------------------------------------------------- Original Rev. 19-JAN-2006 Current Rev. 10-Nov-2008 Intent ------ The intent was to test the D-Zero dual FET preamps and estimate how well they would work with LArTPC signal levels. For this test I tried to simulate the signal level and waveform at the middle induction plane in the LArTPC detector. The performance of the ICARUS V791Q electronics with signals from the middle induction plane can be seen in their document titled, "The ICARUS Readout System", page 14. This is the ICARUS document that has the spiral coil along the left hand side. The heading on page 14 is, "Performance of the V791Q Boards". Your can find gif images of pages 13 and 14 from this ICARUS document in this web directory with the filenames, icarus_readout_book_pg_13/14.gif Test Signal ----------- The test signal is one cycle of a 150 kHz sin wave. As used for this test the single cycle of 150 kHz sin wave repeats every 10 msec. The first half of the sin wave simulates the cloud of electrons appearing as it comes through the first induction plane and drifts towards the middle induction plane. The second half cycle simulates the cloud of electrons drifting away from the middle induction plane and "disappearing" as it is absorbed by the collector plane. Each half cycle of 150 kHz takes 3.3 usec which should be about the length of time required for the cloud of electrons to drift from one plane to the next. The signal is scaled so that first 22k electrons go into the preamp and then 22k electrons are pulled back out of the preamp. Test Setup ---------- An HP arbitrary waveform generator is used to make the single cycles of 150 kHz at the level of a couple of Volts. This signal goes through a coaxial choke as it enters the preamp enclosure where it is attenuated and then goes through a set of high value resistors (to make the test signal into a current source signal) and is then connected to the preamp input. Connected between the preamp input and ground is a silver mica capacitor to simulate the detector capacitance. Three different values of detector capacitance were used in this test: 0 pFd, 330 pFd, and 660 pFd. The preamp enclosure also contains filters on the power supplies for the preamp and it contains a stage of additional amplification that is used between the output of the D-Zero dual FET preamp and the input to the D-Zero ADF-2 card (the ADC plus circular buffer memory card). The analog output signal cable from the preamp enclosure runs about 10 feet over to a VME crate that holds the ADF-2 card. The VME crate is connected to and controlled by a PC. The VME to PC interface is made with standard commercial hardware from Bit-3. The software on the PC that was used for this test is derived from the Run 2B L1 Cal Trig control software. The firmware in the ADF-2 card that was used for this test is derived from its L1 Cal Trig firmware. The frequency response of the analog section may be shaped at a number of points both in the preamp itself and in the amplifier between the preamp and the ADF-2 card. The high frequency and low frequency roll off in gain that were used for this test are "reasonable" but have not been optimized. Optimizing the response of the analog section as a function of frequency requires knowing: the real signal shape, the characteristics of the noise sources, and how the signal will be analyzed in the raw ADC data. The response of the ADF-2 card is basically flat across the frequencies of interest for these tests. The ADCs on the ADF-2 card are "sampling converters", i.e. when they are told to convert they sample their analog input voltage at that instant in time and then provide a digital output value that is proportional to this voltage. Differences between this test and ICARUS ---------------------------------------- The signals at ICARUS are scaled so that a muon as seen by the middle induction plane results in about 10 ADC counts in their system. Refer to page 14 of their publication mentioned above. They say that they use such small Physics signals because they need a lot of dynamic range. I assume that they need this dynamic range to handle large low frequency noise. They have 10 bits ADC (1023 counts full scale). For this test I have scaled the amplification in front of the ADC so that a 22k electron signal makes about 60 to 70 ADC counts. I wanted a larger digital output from the ADC to reduce the quantization noise and to have a better idea what the signal looks like. The ADF-2 card also has 10 bit ADCs. ICARUS samples their signals once every 400 nsec (see pg 14 again). For this test I wanted to see more detail so I sampled about twice as often. The ADC sampling in the ADF-2 card is related to the Tevatron RF frequency. For this test the ADF-2's ADCs are sampling every 197 nsec and the circular buffer is 2048 locations long. Signal to Noise Measurement --------------------------- So far I have only collected a number of "output records", i.e. records of 2048 consecutive ADC samples from the ADF-2 card when the test signals described above was sent into the preamp. I have plotted this data and put it on the web as described in the next section. I have not tried to quantify the signal to noise ratio. I have not tried to quantify this because I don't know what to measure as the signal. During real Physics analysis the signal would be a set of consecutive ADC samples that have the correct shape and amplitude to be a real signal. If for now we want to quantify the signal by just its amplitude then we could measure the signal to noise ratio by taking the RMS value (in ADC counts) of the ADC samples well away from the test pulse and comparing that with the amplitude (in ADC counts) of the test pulse. The bulk of each 2048 ADC sample record is just noise (the test pulse is narrow compared to the record length) so there is plenty of data to measure RMS noise level. Files from this Test -------------------- All of the files are on the web at: www.pa.msu.edu/~edmunds/LArTPC/Preamp_Tests/ The filenames are of the form: ind2_xpf_y.tdj or ind2_xpf_y.pdf The "ind2" is for 2nd induction plane. The "_xpf_" in the filename is either: 0pf 330pf 660pf i.e. the amount of detector capacitance that was used in this run. The "_y" in the filename is either: 1 2 3 4 i.e. I recorded 4 files at each setting of the detector capacitance. The ".pdf" files are just normal pdf file plots of this data. The horizontal axis is the ADC sample number: 1 through 2048. The vertical axis the the ADC value. Note that for this test the ADF-2 cards were set to have a 500 count pedestal, i.e. with zero input signal the digital output of the ADC was about mid scale. The "bump" due to the test pulse is rather narrow on the time scale of these plots so don't miss it. In all cases the peak of the test pulse should be at about sample number 1715. The actual record length of each file is about 405 usec, i.e. 2048 x 198 nsec. The ".tdj" files are normal ascii files that you can look at on the web with no problem. Each file has a short header and then it has the raw ADC data from that record. This data is in "sample number, ADC value format". At the end of the file there is a two line trailer. The intent of providing the raw data in this simple ascii decimal format is that it would be easy for other people to import it if they want to work with it. For example you should be able to clip it and put it right into a spread sheet. Some things people might want to do with the data are: measure the RMS noise away from the test pulse, look at just every other point (i.e. 400 nsec sampling), plot just the section around the test pulse. Noise Spectrum -------------- After the initial tests, that are described above, were made on these D-Zero Dual FET Preamps it was decided to collect some 2048 long records of just the noise. These records were then transformed into frequency space to get an idea of what the noise spectrum looks like. This data was collected with both 0 pFd and 660 pFd simulated detector capacitance. Two data records were collected at each detector capacitance. This noise data is in the files in this web directory with filenames of the form: spectrum_noise_150kHz_ijkpf_n.data The "_ijkpf" in the filename is either 0pf or 660pf, i.e. the detector capacitance. The "_n" is either 1 or 2 for the first or second example of each noise record. Each of these files is just 2048 consecutive ADC sample. The data is in ascii 4 digit decimal numbers in the range 0:1023. The same filenames with the ".pdf" suffix are the plots of the transform of this data. The information at the top of each of these plots describes the data. The horizontal axis is calibrated in frequency in MHz. The vertical axis is just a relative scale. The vertical scale is signal voltage, not power. Note that both scales are log. To give some idea of the vertical scale calibration and to verify the horizontal frequency calibration, I had a very low level 150 kHz continuous sin-wave signal injected into the input of the Preamp while this data was collected. This 150 kHz continuous sin-wave input was at a low enough level that when the output of the Preamp was viewed on a scope it was just visible in the noise. As expected the noise increases at lower frequencies. This emphasizes the importance of selecting the low frequency cutoff in this signal path. The cutoff frequency has not be optimized for these tests. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Signal Response Curves -------------------------- Rev. 24-SEPT-2008 The following set of tests was made to record the waveform at various points along the signal processing path of the PMB-16 card with semi- realistic input waveforms to the preamplifier. The HP generator was setup to provide either a semi-realistic induction plane waveform (one cycle of a 150 kHz sine wave that starts by swinging negative) or a semi-realistic collection plane waveform (a negative going trapezoid pulse with the HP generator set to make a 3.33 usec wide pulse with 1.5 usec rise and fall times which on the scope looks like 5.2 usec wide at the top, 1.3 usec wide at the bottom). These waveforms are described in more detail on pages 187 and 189 of notebook #8. The voltage output of the HP generator is carried on a coax cable through two coax chokes and then through the wall of the outer shield box. Immediately inside the wall of the outer shield box the HP generator signal reaches a Pomona box where it is terminated, AC coupled, and converted into a current signal by a high value resistor. From the Pomona box the current signal passes through about 5 feet of miniature coax cable which includes another coax choke and then into the inner shield box where it plugs into a channel on the PMB-16 board. The preamplifier uses in the test to record the first 8 waveforms in this series was fully modified for standard Bo operation as described in the following file on the web: www.pa.msu.edu/~edmunds/LArTPC/Bo_LArTPC_DAQ/PMB_16/ pre_amplifier_modifications.txt Recall the steps in the PMB-16 signal processing including the coupling network to the ADC on the ADF-2 card: 1. DC Blocking Capacitor and Bias Voltage Distribution Resistor 2. Preamplifier section of the Preamp Hybrid 3. Driver section of the Preamp Hybrid 4. Low Pass Filter (LPF) rejects signals above some frequency 5. High Pass Filter (HPF) rejects signals below some frequency 6. Coupling network to the ADC on the ADF-2 card Waveforms were recorded at the output of the preamplifier section of the preamp hybrid, at the output of the LPF, and either at the output of the HPF or else at the output of the coupling network to the ADC, i.e. the ADC's view of the signal. Waveforms were recorded with both the induction plane and collection plane input signals and at both a 4 usec/div and at a 10 usec/div sweep rate The first 8 scope pictures in this series have the following filename format: ind hpf 4us fc_1_ or - or - or . tif col adc 10us "fc_1_" --> family of curves series 1 "ind or col" --> Induction or Collection input signal "hpf or ADC" --> The bottom yellow trace is the output signal from either the High Pass Filter or from the coupling network to the ADC on the ADF-2 card. "4us or 10us" --> the horizontal sweep rate In the "fc_1_" series of scope pictures the: top Pink trace is the induction or collection plane input signal Blue trace is the output of the preamplifier section of the preamp hybrid Green trace is the output of the Low Pass Filter bottom Yellow trace is the output of either the High Pass Filter or of the coupling network to the ADC on the ADF-2 card Note this yellow trace on the scope shows up as dark blue on some displays and printouts. All scope pictures in the "fc_1_" series were made using a 20kHz channel (Channel Number 10) on the PMB-16 card SN#1. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Signal Response Curves -------------------------- Rev. 25-SEPT-2008 The following set of tests continues the study of the signal processing of the Bo electronics. To try to improve the signal shape of the Collection plane electronics do the following steps: 1. Return to the original value of the capacitor at the input to the Driver stage of the D-Zero preamp hybrid. That is, C6 is 150 nFd which gives a time constant at the input to the Driver stage of the preamp hybrid of 150 nFd x 3.3k Ohm = 495 usec. 2. Increase the decay time of the preamplifier integrator stage from 40 usec to 200 usec. This will now match the preamp integrator decay time of the T962 electronics. This is done by increasing the feedback resistor in the preamplifier input stage from 20 Meg Ohm to 100 Meg Ohm. 3. At this point take a set of scope pictures using the simulated collection plane input signal. 4. The time constant of the AC coupling between the output of the preamp Driver stage and the input of the first filter stage, i.e. Low Pass Filter, on the stock Bo electronics is 50 usec i.e. C20 is 150 nFd x 335 Ohm = 50 usec. Increase C20 to 620 nFd to give a time constant of 200 usec. 5. At this point take a set of scope pictures using the simulated collection plane input signal. 6. Finally move the second filter, i.e. the High Pass Filter, from a cutoff of 20 kHz to a cutoff of 50 kHz. This was done by changing C31 from 22 nFd to 10 nFd, changing C33 from 22 nFd to 10 nFd, and changing C34 from 4.7 nFd to 2 nFd. 7. At this point take a set of scope pictures using the simulated collection plane input signal and then a set of scope pictures using the standard simulated induction plane input signal. All of this work was done on channel number 10 of the PMB-16 card serial number #1. All of the scope pictures from this series of tests are in the files with filenames that start, "fc_2_". All waveforms in this series of the output of the PMB-16 card were taken at the output of the second filter stage. The scope lead connections in this series of pictures is the same as in the fc_1_ series which is: In the "fc_2_" series of scope pictures the: top Pink trace is the induction or collection plane input signal Blue trace is the output of the preamplifier section of the preamp hybrid Green trace is the output of the Low Pass Filter bottom Yellow trace is the output of the 2nd stage filter, the High Pass Filter. Note this yellow trace on the scope shows up as dark blue on some displays and printouts. The filenames are of the form: ind int 4us fc_2_ or - cpl - or . tif col or 10us flt "fc_2_" --> family of curves series 2 "ind or col" --> Induction or Collection input signal "int" --> Only the integrator had been changed, i.e. step #3 scope pictures. "cpl" --> Both the integrator and the coupling at the output of the Driver stage had been changed, i.e. step #5 scope pictures. "flt" --> The integrator, driver output coupling, and the 2nd filter, the HPF, had all been changed. These scope pictures are from step #7 above. "4us or 10us" --> the horizontal sweep rate =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Unipolar and Bipolar Shape Curves ------------------------------------- Rev. 7-OCT-2008 The two files in this directory: shape_unipolar_7oct08.tif and shape_bipolar_7oct08.tif show the "official" response curves to the standard collection plane test pulse in the test setup at MSU. top Pink trace is the collection plane test pulse to the V_to_I box. Blue trace is the output of the Preamplifier section of the preamp hybrid Green trace is the output of the Driver section of the preamp hybrid (before the series output R and C). bottom Yellow trace is the output of the PMB-16 card as seen by the ADC on the ADF-2 card, i.e. the PMB-16 output has been terminated by 470 nFd and 80 Ohms to ground and the scope is looking across the bottom 50 Ohms of the 80 Ohms. - The test pulse is the standard 3.33 usec wide, 1.5 usec edge time single pulse. - The preamps have been "flattened", i.e. 100 Meg Ohm feedback resistor and 4.7 uFd input cap to the Driver section. - The unipolar section uses 15 nFd coupling cap into the first filter. This is the only differentiation. The first and second filters are both the same, the standard design 200 kHz 2 pole LPF. - The bipolar section uses 150 nFd in parallel with 4.7 uFd coupling capacitor into the first filter which is the standard 200 kHz 2 pole LPF. The second filter is the standard 20 kHz 2 pole HPF. - PMB-16 SN#8 Channels 0:7 are setup for unipolar and channels 8:15 are setup for bipolar. - All channels on PMB-16 SN#8 have 60 Ohm series output resistors and the output capacitor has been removed. - This PMB-16 card will be tested in the Bo Collection Plane. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Narrow Unipolar Shape Curves ---------------------------------- Rev. 31-OCT-2008 The six files in this directory: narrow_unipolar_pict_*.tif show the "official" response curves of the narrow unipolar Gaussian filter to a step function input and to the "flattened" Bo preamp with a simulated collector plane input signal. Filter response curves to step function input: narrow_unipolar_pict_1.tif 4 usec/div step function input narrow_unipolar_pict_2.tif 10 usec/div step function input narrow_unipolar_pict_3.tif 4 usec/div 1.5 usec edge input narrow_unipolar_pict_4.tif 4 usec/div 3.0 usec edge input top Ch #3 Pink trace Input signal to the filter (Preamp-Drive hybrid output) Ch #2 Blue trace is the output of the first stage of the filter Ch #4 Green trace is the output of the 2nd stage of the filter taken before the series terminator output resistor bottom Ch #1 Yellow trace is the output of the PMB-16 card as seen by the ADC on the ADF-2 card, i.e. the PMB-16 output has been terminated by 470 nFd and 80 Ohms to ground and the scope is looking across the bottom 50 Ohms of the 80 Ohms. Filter response curves to the Preamp output with a simulated Collector plane input signal narrow_unipolar_pict_5.tif 4 usec/div preamp Collector plane signal narrow_unipolar_pict_6.tif 10 usec/div preamp Collector plane signal top Ch #3 Pink trace Input signal to the preamp (simulated Collector plane signal) Ch #2 Blue trace is the output of the preamp (preamp to driver node) Ch #4 Green trace is the output of the 2nd stage of the filter taken before the series terminator output resistor bottom Ch #1 Yellow trace is the output of the PMB-16 card as seen by the ADC on the ADF-2 card, i.e. the PMB-16 output has been terminated by 470 nFd and 80 Ohms to ground and the scope is looking across the bottom 50 Ohms of the 80 Ohms. The file narrow_unipolar_daq_plot.gif is a plot of the daq readout with the standard 3.5 femto-Coul Collection plane test pulse at the preamp input. - The step function test pulse is 330 mVpp fast edge going to the input of a "Driver Only" hybrid where it is terminated in 50 Ohms and coupled to the Driver Only hybrid input with 4.7 uFd Tantalum capacitor in parallel with a 15 nFd ceramic capacitor. - The collector plane test pulse going to the preamp input through a compensated 22 Meg Ohm resistor is the standard 3.33 usec wide, 1.5 usec edge time, 23.8 mVpp single pulse which gives 3.5 femto-Coul. - The preamps have been "flattened", i.e. 100 Meg Ohm feedback resistor and 4.7 uFd in parallel with the normal input cap to the Driver section. - There is a 10 nFd coupling cap between the Driver output and the input to the first filter stage. This is the only differentiation. - The first and second filter stages are both part of a 1 usec Gaussian 4 pole LPF. - This PMB-16 test channels has 60 Ohm series output resistors and the output capacitor has been removed. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Narrow Bipolar Shape Curves --------------------------------- Rev. 4-NOV-2008 The four files in this directory: narrow_bipolar_pict_*.tif show the "official" response curves of the narrow bipolar Gaussian filter to a step function input and to the "flattened" Bo preamp with a simulated collector plane input signal. Filter response curves to step function input: narrow_bipolar_pict_1.tif 4 usec/div step function input narrow_bipolar_pict_2.tif 2 usec/div step function input top Ch #3 Pink trace Input signal to the filter (Preamp-Drive hybrid output) Ch #2 Blue trace is the output of the first stage of the filter Ch #4 Green trace is the output of the 2nd stage of the filter taken before the series terminator output resistor bottom Ch #1 Yellow trace is the output of the PMB-16 card as seen by the ADC on the ADF-2 card, i.e. the PMB-16 output has been terminated by 470 nFd and 80 Ohms to ground and the scope is looking across the bottom 50 Ohms of the 80 Ohms. Filter response curves to the Preamp output with a simulated Collector plane input signal narrow_bipolar_pict_3.tif 4 usec/div preamp Collector plane signal narrow_bipolar_pict_4.tif 10 usec/div preamp Collector plane signal top Ch #3 Pink trace Input signal to the preamp (simulated Collector plane signal) Ch #2 Blue trace is the output of the preamp (preamp to driver node) Ch #4 Green trace is the output of the 2nd stage of the filter taken before the series terminator output resistor bottom Ch #1 Yellow trace is the output of the PMB-16 card as seen by the ADC on the ADF-2 card, i.e. the PMB-16 output has been terminated by 470 nFd and 80 Ohms to ground and the scope is looking across the bottom 50 Ohms of the 80 Ohms. The file narrow_bipolar_daq_plot.gif is a plot of the daq readout with the standard 3.5 femto-Coul Collection plane test pulse at the preamp input. - The step function test pulse is 330 mVpp fast edge going to the input of a "Driver Only" hybrid where it is terminated in 50 Ohms and coupled to the Driver Only hybrid input with 4.7 uFd Tantalum capacitor in parallel with a 15 nFd ceramic capacitor. - The collector plane test pulse going to the preamp input through a compensated 22 Meg Ohm resistor is the standard 3.33 usec wide, 1.5 usec edge time, 23.8 mVpp single pulse which gives 3.5 femto-Coul. - The preamps have been "flattened", i.e. 100 Meg Ohm feedback resistor and 4.7 uFd in parallel with the normal input cap to the Driver section. - There is a 10 nFd coupling cap between the Driver output and the input to the first filter stage. This is the first differentiation. - There is a 4.7 nFd coupling cap between the first stage filter output and the second stage filter input. This is the second differentiation. - The first and second filter stages are both part of a 1 usec Gaussian 4 pole LPF. - This PMB-16 test channels has 60 Ohm series output resistors and the output capacitor has been removed. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Narrow Unipolar Filter with Induction Plane Signals --------------------------------------------------------- Rev. 10-NOV-2008 The two files in this section show the expected performance of the Unipolar Gaussian filter (the same filter that was described in the "Narrow Unipolar Shape Curves" section above) but now operating with a test signal that models the 2nd Induction Plane signal. narrow_uni_fltr_ind_pln.tif 4 usec/div preamp Induction plane signal top Ch #3 Pink trace Input signal to the preamp (simulated Induction plane signal) Ch #2 Blue trace is the output of the preamp (preamp to driver node) Ch #1 Yellow trace is the output of the PMB-16 card as seen by the ADC on the ADF-2 card, i.e. the PMB-16 output has been terminated by 470 nFd and 80 Ohms to ground and the scope is looking across the bottom 50 Ohms of the 80 Ohms. bottom Ch #4 Green trace is the output of the 2nd stage of the filter taken before the series terminator output resistor. The file narrow_uni_fltr_ind_pln_daq_plot.gif is a plot of the daq readout with the standard 3.5 femto-Coul Induction plane test pulse at the preamp input. - The Induction plane test pulse going to the preamp input through a compensated 22 Meg Ohm resistor is 1 cycle of a 150 kHz sin wave (the negative half cycle comes first) with a peak to peak amplitude if 73.2 mV. - The preamps have been "flattened", i.e. 100 Meg Ohm feedback resistor and 4.7 uFd in parallel with the normal input cap to the Driver section. - There is a 10 nFd coupling cap between the Driver output and the input to the first filter stage. This is the only differentiation. - The first and second filter stages are both part of a 1 usec Gaussian 4 pole LPF. - This PMB-16 test channels has 60 Ohm series output resistors and the output capacitor has been removed. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-