APV6 Vienna Manual
Contents
1. CMC Pedestal Mean vs Strip CMC Pedestal Sigma vs Strip 0 i Gated Analysis and File Writing A is CMC Hits vs Strip CMC Hit Sum Histogram CMC Single Strip bose 4 206 llillu Bhtinagestesat naa Update Boundaries Display off Auto Auto Cursors Now Mean 1034 39 a MSE Sigma NaN 11886032 Scale 52177 40 22 0 1140 1150 1160 1170 1180 1190 1200 Status Finished Tempeatures C 1 38 3 2 291 3 243 4 23 8 Detector V l Monitor 100 0 M 646 1 nd Glossary CMS Compact Muon Solenoid Detector at the LHC LHC Large Hadron Collider Future proton proton collider machine at CERN scheduled start of operation 2006 CERN European Laboratory for Particle Physics Geneva CH SNR Signal to noise ratio Strip pitch Distance from the center of a strip to the center of the neighbor strip Implant width Width of the heavily doped thus highly conductive strip implant HEPHY Institute of High Energy Physics Vienna A APV6 Integrated 128 channel preamplifier shaper with 160 cell analog pipeline designed for LHC timing 25 ns bx bx Bunch crossing occurs every 25 ns at LHC VME Standardized crate modules and data bus can be connected to a computer NIM Standardized crate and modules DAQ Data Acquisition Software IEEE 488 Data bus specification for measuring devices SMU Source Measure Unit Providing voltage and measuring2. Hit above In the event data after the pedestal each channel value is compared to its pedestal mean value Once the value exceeds a certain band around the pedestal this is interpreted as a hit Note that peak mode hits are negative lower than the pedestal mean while deconvolution mode hits are positive greater than the pedestal mean The width of the band around the pedestal mean is given in units of the pedestal sigma here Pedestal Gaussian Fit Normally the pedestal mean and sigma values are simply calculated by statistic methods Assuming a Gaussian noise profile a fit can be applied to the channel histograms obtaining Gaussian mean and sigma parameters When this box is checked a fit is applied and the pedestal mean and sigma values are taken from the fit when 0 5 og lt 6g lt 2 os However this does not always work well since APV6 Vienna Manual How to handle DAQ and analysis hard and software 19 there is no reliable control whether the fit was successful or not Thus recommend not to use this option Update Cycles These values specify the interval in seconds between screen updates of the displayed histogram the rate meter and the temperate and Keithley SMU readout Immediately after each temperature and voltage current readout a line with these values the current date time and event number is appended to the information file The update cycle values can be safely adjusted during a run Data file path This
3. APV6 Settings Comparison 30 a eg in re a al ge VLE aoan PME allt il G X VEI a pvolution Mode SNR 7a 1 APV6 Vienna Manual How to handle DAQ and analysis hard and software M Friedl HEPHY Vienna Email Markus Friedl cern ch V0 91 Feb 19 1999 APV6 Vienna Manual How to handle DAQ and analysis hard and software 2 Document Source amp Related Documents This document is electronically available at http Awwwhephy oeaw ac at u3w f fried Awww apv6 where it can be downloaded in PDF and PS formats In the same place our transparencies from talks at the CMS meetings and a more general APV6 info paper can be found Definitely the most important source of knowledge concerning the APV6 chip itself is the Reference Manual which can be found at ftp ftp te rl ac uk apv6 user_manual apv6_user_manual_2 0 ps The blanks are in fact underscore characters Basic knowledge about the APV6 which can be obtained by the manual mentioned is required for understanding this paper Introduction This paper describes the hard and software that is currently available and operable at the HEPHY Vienna Many items may be of principal interest concerning the APV6 others are device specific The intention of the Vienna hardware was not to work towards a final CMS readout design but rather to have a compact stand alone system which is easily transported to test beams or operated in the lab Thus
4. Start DAQ 200 Stop DAQ 80 Help T Quit 130 O 90 y 10 pan Fie io White File Trigger 80 imi willl On v HW 70 A at Sw Max Events a 40 Deni 20000 Current Event 20000 20 Progress a a 900 920 940 960 80 1000 1020 1040 1060 1080 1100 1120 1140 1160 1180 1200 1220 1240 1260 1280 1300 1320 1340 1360 1380 1400 x Trigger Rate Hz 200 400 Graph Boundaries Cursors Freeze Gaussian Fit al a Raw ADC Data Online o T a min 4 900 x mar 3 1400 ge 1029 25 Exo 1234 34 A a Aae Update Boundaries in 4 Displ Off Auto Y_min 5 o Y_max 3 229 gE 113 28 uE 75 68 Pedestal Subtracted CMC Strip Data piesa F ue I i Auto Cursors ho CMC Pedestal Mean vs Strip File Name temp on CMC Pedestal Sigma vs Strip a Now Mean 0 00 creen 0 i ae t 3 tonen Gated Analysis and File Writing 2 ae ea vs Strip a e Sana 300 Print ingle Stip Histogram 12 foes Scale 0 00E 0 eae to fie CMC Hit Sum Histogram M aoee aooe Begin Time 17 02 1999 13 42 26 Status Finished End Time 17 02 1999 13 43 37 APV6 Enor Codes 3 3 3 Tempeatures C 1 335 2 254 3 222 4 225 Source Testbeam Setup The picture below primarily shows the NIM logic for
5. i Delay Logic i VME Delay See Pike Ge z Clock APV6 Read 2 Clock Optional Components ADC Board Harddisk nerag Sae Data PC Storage Serial Interface In the subsequent sections all the hardware devices are described in detail Detector Hybrid Repeater These components mounted together on a plastic support are the very heart of the system The support fits into a green cooling box 12V lt 10A which must be flooded with inert gas N or He when operating to avoid condensation of water on the electronic parts A few months after the testbeam the cooling box was operating for approximately one APV6 Vienna Manual How to handle DAQ and analysis hard and software 4 week without gas flooding for a scan of the bias settings After that the floor of the box was entirely covered with water and the electronics was really wet Also the detector has taken some damage at this unfortunate event some strips now show up with excess noise CAUTION Care must be taken when handling the support as the detector s backplane is uncovered at the bottom and the aluminum window on top is very fragile Repeater APV6 Hybrid Detector Temperature Sensors Detector CAUTION Extreme care must be taken with both detector and APV6 when handling or soldering nearby Never touch these components with bare fingers The wire bonds are extremely fragile The detector currently install
6. ri CMC Pedestal Mean vs Strip File Name Em CMC Pedestal Sigma vs Strip Now Mean 535 creen in il a OO TT S Gated Analysis and File Writing A M Euit vs Strip aire MSE Sigma 14 01 Print o CMC Single Strip Histogram ff 34 54 fo66633 Scale 1324 56 to file CMC Hit Sum Histogram Begin Time 08 02 1999 18 21 18 Status Finished End Time 09 02 1999 09 18 39 APVBEnorCodes 3 3 3 Tempeatues C 1 381 2 288 3 238 4 232 Detector l Monitor 100 0 619 8 nd Write File Turns on or off the writing of both binary data and information text files Trigger In any case trigger signals must be fed into one of the three inputs of the APV6 Read 2 VME module The default setting is hardware HW triggering With the software SW trigger the software generates signal pusles at the OUT NIM terminal which can be used for triggering when connected to one of the three trigger APV6 Vienna Manual How to handle DAQ and analysis hard and software 20 inputs This can help when one quickly wants to test the DAQ system without having a NIM clock dual counter at hand e Max Events When this value is set to 0 the DAQ runs until the Stop DAQ button is pushed On the other hand it is possible to automatically terminate a run after a certain number of events Even then hitting the Stop DAQ button manually overrides the automatic termination The maximum number of events can also be entered or altered during a
7. but also to allow non destroying separation of hybrid and detector Both the detector and the chip bond pads are very small and do not allow more than 1 bond When such an existing bond is removed there is very low probability than another bond will stick there On the glass fanouts however there is ample space for many bonds Consult M Krammer for further information concerning the detector Hybrid CAUTION Extreme care must be taken with both detector and APV6 when handling or soldering nearby Never touch these components with bare fingers The wire bonds are extremely fragile Soldering requires a ground connection to the iron tip The hybrid was manufactured by the Imperial College IC London UK It can support up to 8 APV6 chips but in our setup only holds 3 of them The APV6 chip designed by the Rutherford Appleton Laboratory RAL Chilton UK consists of 128 identical channels with preamplifier shaper stages a 160 cell analog pipeline and a deconvolution network each Surrounding elements are the digital logic including an I C bus the bias generator and the output multiplexer For details concerning the APV6 chip please consult the APV6 user s manual see p 2 The general function flow of the APV6 Operation is the initialization over the C bus after which a reset must be given Each chip has unique I C addresses for writing and reading Then the chip is ready and waiting for triggers According to
8. measurements In order to cut all unwanted triggers the trigger pulse should be as narrow as possible When the trigger pusle width is 25ns all triggers are accepted When it is say 5ns only each fifth trigger is accepted but these triggers are synchronous However there is a certain minimum time around 3ns for the trigger width to be accepted at all Thus the calculation of the effective trigger window is not as easy as diving the width by 25ns The simplest way to measure this window is to plug the trigger into the CAL TRI then into the CAL TRI_Syn terminals and count the accepted rates This can be done either by software or with a NIM counter connected to CAL Out Note that the rates must not exceed the software maximum rate otherwise the veto must be turned off Whenever a calibration signal is generated by triggers at CAL TRI or CAL TRI_Syn a NIM signal with a width of approximately 2us thus seen as a step pulse within the shaping time of the APV6 is pushed out on the CAL Out signal An external delay loop plugged to CAL In on the other end enables the calibration pulse on the APV6 input channel 2 To disable the calibration signal just open the delay loop Consequently shifting the delay allows to record pulse shapes APV6 Vienna Manual How to handle DAQ and analysis hard and software 12 After the module accepted a trigger it sends the calibration pulse to the repeater if enabled then waits for a certai
9. of at least 10 MB s provided that the VME module s can push the data that fast In fact our APV6 Read 2 module cannot push out data at this speed Nevertheless it is quite fast The only interactive element on the front panel is a reset button Note that after hitting this button the Resource Manager ResMan must be re run on the PC to re initialize the driver VME Delay Clock From this module only the 80 MHz NIM Clk output is used The board also provides a VME controlled digital delay however this should not be used for precision measurements APV6 Read 2 APV6 Vienna Manual How to handle DAQ and analysis hard and software 11 This is the center of data acquisition The module provides the control signals and clock for the APV6 Veto LED out of order buffers and digitizes incoming data at 40 MHz and stores f up to 4k words in a FIFO memory NIM IN _ Clock 80 MHz CAUTION Approximately since January 1999 this module intially produces corrupt data when the crate is switched on the driver loaded and the DAQ is started NIMIN CAL TRI The only solution found so far is to simply pull out the NIMIN CAL TRI_Syn card and replug it while the crate remains turned on but 7 the DAQ is not running After that everything works fine Nimin CD CALIn NIMIN TRIG In any case the module must be fed with a usually NIMOUT CAL Out 80Mhz clock which is divided internally to 40MH
10. current or vice versa APV6 Vienna Manual How to handle DAQ and analysis hard and software 23 Hybrid Ceramic support with ICs and SMD devices small PCB PCB Printed circuit board with ICs and other devices C Data bus specification HV High voltage FOXFET Special type of field effect transistors integrated in the detector which act as a resistor used for biasing each strip of a detector Guard ring Grounded metalization around sensitive silicon structure to shield leakage currents LAT Latency time the number of bx or clock cycles the APV6 looks in the past of the analog delay pipeline when a trigger occurs ASPS Switched capacitor deconvolution network SMD Surface Mounted Device MIP Minimum lonizing Particle FIFO First in first out memory A D Analog to digital ADC Analog to digital converter PM Photomultiplier ISA Industry Standard PC bus 1 0 Input Output GUI Graphical User Interface CMC Common Mode Correction removing random shifts of groups of channels
11. eventually false pulses occur mostly in bunches This probably means some electric discharge process At the rates given there is no significant interference lt 0 1 of triggers However experience has shown that the rate of false triggers by electric discharge can increase within a time scale of days or even weeks The signal cables between the 2 PMs and the NIM discriminators are of equal length 21 m The length of this cable again is a critical parameter for the timing Especially longer cables should not be too lossy since a good PM signal shall reach the discriminator Temperature Readout Four temperature sensors are attached to the Detector Hybrid Repeater block Sensor 1 is mounted next to the 3 APV6s sensor 2 on the far end of the detector Sensors 3 and 4 can be placed elsewhere In the June 1998 SPS testbeam sensor 3 was inside the cooling box 4 outside These sensors are supplied and read out by a primitive ISA I O card A 34 pin flat ribbon cable connects the I O card and a small resistor network board which allows to adjust the offset Care must be taken with this connection not to short circuit the PC power lines The sensors are also plugged into the resistor board APV6 Vienna Manual How to handle DAQ and analysis hard and software 14 The temperature readout is optional however the I O board is essential If the sensors are not connected obviously wrong temperature values are displayed by the DAQ soft
12. running data acquisition However one should be careful not to unwantedly terminate the run by entering a number lower than the current event counter If a maximum event number is specified the thermometer bar displays the progress in percent e File Name Enter the name of the file such as run034 without any extension here When file writing is enabled the software will produce a binary data file called run034 dat and an information text file run034 txt Both files are located in the folder specified in the settings panel Note that 1548 bytes of data file are used for each event The behavior is undefined when the disk becomes full during a run e Comments The user may write any comments here The content of this box is also saved within the information text file To produce a line break in this box hit Ctrl Enter e Gated Analysis and File Writing This feature does not make much sense outside a testbeam with spill structure In a testbeam however as mentioned in Source Testbeam Setup section the IN NIM input connected to the spill on off NIM signal accelerates the data taking by a factor of approximately 2 when gating is enabled When the input is not connected the panel switch setting is indifferent After starting a run the error codes of the 3 APV6s are displayed If a 3 appears for all chips everything is OK When starting the DAQ after switching on the VME crate without unplugging and replugging the APV6 Read 2 mod
13. the hardware is limited to the readout of 4 APV6 chips 3 of which are currently attached Certain items of the readout system may be altered in future in that case the logbook and or module schematics should be concerned Hardware General block diagram The following block diagram contains all hardware blocks that can be connected to the system However some devices are not obligatory The Mac488B Bus Controller serial to IEEE 488 interface is only necessary when detector voltage current monitoring by the DAQ is desired The VME CTR module which controls delay and attenuation modules in the NIM crate is only required for special pulse scans over time Also one can do without the temperature monitoring APV6 Vienna Manual How to handle DAQ and analysis hard and software 3 Vienna APV6 Readout System Detector Hybrid Repeater Silicon 3APV6s Repeater trip Detector n Hybri Electroni Mac488B Keithley 237 sMU IP on Hybrid Electronics oS Voltage yr Regulators ee a Cal Pulse SS Generator 12 different geometry zones 32 strips each Si 384 strips in total Temperature Sensors Scintillator Photomultipliers NIM Crate VME Crate HV Supply and S fe Discr Coinc Logic Trigger and
14. anual How to handle DAQ and analysis hard and software 10 Please refer to the Keithley 237 and the OTech Mac488B manuals for further information VME Modules The slot position of any VME module is irrelevant except for the Crate Controller which must use the leftmost slot 1 Each module uses a certain address range which in most cases can be selected using DIP switches The T amp M Explorer which is sort of a VME registry for the driver must know all valid address ranges However it does not matter if a module specified in the T amp M Explorer does not really exist in the crate Refer to Helmut Steininger for further information on this topic The table below summarizes the address ranges used for the APV6 setup The VME Delay Clock is used as a NIM clock source only and thus needs no communication VME Device VME Address Range Crate Controller VME MXI 2 0x000000 Ox00FFFF APV6 Read 2 0xB02000 OxBO2FFF VME oC OxAC2000 OxAC2FFF OxAC3000 OxAC3FFF Delay CTR 0xA10000 OxA1OFFF After turning on PC and VME crate the Resource Manager ResMan must be started in order to initialize the communication On our current PC it always hangs up after everything is done nevertheless everything works fine Crate Controller The VME Crate Controller is manufactured by National Instruments and connects to a PCI card inside the PC via the MXI bus two thick cables It is specified for a 32 bit trasfer rate
15. convolution mode signals are positive while peak mode signals are negative APV6 calibration pulse shapes peak and deconvolution modes default bias settings 150 Peak 109 Decon signal ADC 50 50 When there is no trigger to the APV6 it sends alive ticks every 1 75 us A few us after a trigger input a 4 bit header an 8 bit pipeline address and then 128 analog channel values are serially pushed out on the output line The channel values do not appear in the natural strip order but they are sent over a multiplexer stage and come in a mixed up order Natural channel order 2 31415 6 7 8 9410 11 12 13 14 15 Position in output data O 32 64 96 8 40 72 104 16 48 80 112 24 56 88 120 When two or more triggers appear within a few bx the corresponding pipeline cells are marked against being overwritten and a number of header address signal blocks are subsequently pushed out after the initial delay of a few us In peak mode up to 18 such events may occur before the data is pushed out without loss of data In _ inv deconvolution mode where 3 aes ee BRD wns cells are must be saved per J JED a a E a a E tas sample only up to 6 events are possible This leads to little signal loss with high trigger rates see http pcvisi5 cern ch cmstcontrol documents APV_pipeline_inneficiencies padf and ht
16. current If the current remains in the order of or below 1nA one can be sure that the bias line is interrupted somewhere When the final voltage probably 100V or 150V is reached hit ENTER to remove the cursor The voltage current monitoring is optional and can be switched on and off with the DAQ software It only works if there is a Mac488B interface present which is wired both to the PC with a DB9 Mac8 serial cable and to the Keithley 237 IEEE Address 16 with an IEEE 488 cable The parameters for the serial connection are Serial Port COM2 Baud rate 9600 Data bits 8 Stop bits 2 Parity None Handshaking RTS CTS These settings are the defaults of the Mac488B device Inside the Mac488B case there are 3 rows of DIP switches with which also the serial parameters can be altered Moreover when facing the front panel DIP switch 2 in the middle row is pressed down in the top CLOSED position MacDriver488 mode by default This mode only works with a special Mac driver which is not available on the PC Therefore this switch must be pushed down in the bottom OPEN position System Controller mode for the operation within the APV6 environment When you doubt about the connection to the Mac488B connect with any terminal program such as Hyperterminal which comes together with Windows NT and send hello CR to the Mac488B which should answer with its name and version number APV6 Vienna M
17. disk and squeezed through the online analysis This accelerates the data acquisition approximately by a factor of 2 with the current PC If gating is disabled or no beam on signal is present each event data is immediately written to disk and analyzed Note that the spill on off signal in no case poses a veto on incoming triggers The smart solution for this problem simply is to additionally plug the spill on off signal into the coincidence Software The main program simply referred to as DAQ software is the apv6 prj project It has a GUI and quite sophisticated online analysis features APV6 Vienna Manual How to handle DAQ and analysis hard and software 17 Data recorded with this program can be analyzed with a UNIX C program called apv6_oa c Basically it performs the same analysis as the DAQ software but more elaborate in detail A special PAW macro called apv6 kumac displays the analyzed data and applies Landau fits A few other DAQ programs exist which are poorly documented and do not have a GUI Often such programs have been adapted for current issues by commenting unwanted lines Nevertheless will try to briefly describe these programs PC CVI Programs CVI is a visual C manufactured by National Instruments http www natinst com Before any CVI program can communicate with the VME crate the specific drivers must be installed and configured Note that the driver needs the address information of each VME module se
18. e Hardware section Provided the drivers are OK one must activate them by running the Resource Manager ResMan once the VME crate is turned on An administrator account is necessary to run this program On our current PC ResMan hangs up after everything is done Ignore this fact and click on the OK box to terminate the program Note that ResMan cannot be started once a CVI program is running You must exit any running program not CVI itself before activating ResMan Now any CVI program can be executed by loading the project and running it The ResMan initialisation need not be repeated when quitting and restarting CVI programs or the CVI environment itself The only actions when ResMan must be executed are e Resetting the VME Crate Controller by pushing the RESET button or turning it off and on again e User logging off ending the session or restarting the PC There are two possible modes to run a project which can be set in the Options Run Options popup To test a new or adapted program set the Debugging Level to Standard to allow breakpoints and such stuff For the final program always set None which accelerates the execution by a factor of up to 10 DAQ software apv6 prj This software runs with all components shown in the general block diagram see Hardware section except for the VME Delay CTR The DAQ software can handle all setups described in the previos section It has powerful online anaysis routines perfor
19. ed is a 6 25 x 6 25 cm multi region sillicon detector 12 manufactured by the Hamamatsu corporation It consists of 12 zones with different strip geometries Each zone has 32 strips making 384 strips altogether matching the number of inputs of 3 APV6 chips 128 input channels each The detector zones have different strip pitches and implant widths Viewed from the side with the APV6 chips the zones and strips are numbered from left to right zoemer o 1 2 3 4 sie 7iea le jo a APV6 Vienna Manual How to handle DAQ and analysis hard and software 5 ali pitch 6o so 240 120 60 so 240 120 60 80 240 120 a width 55 40 70 50 20 25 50 35 15 15 30 20 Strip numbers 0 31 32 63 64 95 96 127 128 159 160 191 192 223 224 255 256 287 288 319 320 351 352 383 Zone limits mm Note that the various zones are not of equal widths due to the different pitches This causes some complication when e g plotting a beam profile hits vs strips The usually Gaussian shape is not linearly projected onto the strips for two reasons First the various strip pitches cause geometric distortions in the shape and second the hit probability or collection depends on the area which again depends on the pitch Furthermore a fraction of hits is lost in the inactive regions between the zones On the side opposite to the APV6 chi
20. en of the NIM modules as we use only standard components except for the VME controlled delay and attenuator modules which are described in the previous section VME Modules APV6 Vienna Manual How to handle DAQ and analysis hard and software 13 For the source testbeam setup a HV supply for the 2 photomultipliers is necessary This need not be a NIM module but probably is the easiest to integrate into the setup Moreover a dual discriminator a coincidence unit a shaper and a delay unit are required at the minimum For the pedestal calibration setup a dual timer a shaper and a delay unit are essential Refer to the Setups section for detailed schematics Note that different instrumentation and or cable lengths within the NIM setup changes the timing properties The scintillator 1 x 1 cm with two photomultipliers PMs is mounted on an aluminum plate There is no lightguide but only air between these components The two PMs are labelled LEFT PM and RIGHT PM and both need negative high voltage The required HV and discriminator parameters optimized for a Sr source are summed in the table below In a testbeam these values principally should be fine as well 90Sr 1mCi source LEFT PM RIGHT PM Coincidence HV V WA Max Signal Amplitude mV Discriminator THR mV Discriminator WID ns 0 28 Source rate s 40000 40000 35000 Background rate s The dark count rates are very small but
21. ge which leads to a reverse biased pn junction CAUTION Make sure never to apply even the smallest negative voltage to the detector as the diode may easily take damage from forward biasing Bias voltage 100 V should be entirely secure The dark current drawn is normally lt 1 uA thus a compliance of a few microamps should be fine However the 240 um pitch zones may not yet be completey depleted for them 150 V is better However the detector may show some kind of breakthrough at this voltage after which it must rest for at least several hours Thus be careful when applying the bias voltage When minimum ionizing particles MIP traverse a fully depleted 300 um silicon detector the resulting charge spectrum after pedestal contamination is removed has a most probable MP maximum value of 22400 e and a mean value of around 31000 e The spectrum can be approximated by a Landau distribution APV6 Vienna Manual How to handle DAQ and analysis hard and software 6 The final detector however should withstand at least 500 V With increasing radiation damage the depletion voltage decreases in the beginning until the point of type inversion then it increases continuously At the end of the scheduled 10 years of LHC operation a voltage in the order of 500 V will be applied in order to achieve depletion There are 2 glass fanouts between the detector strips and the APV6 inputs Their task is not only to match the different geometries
22. her Data file path e apyv6data values Note ila pa Note Use instead of Latency Selects the I Keithley 237 V I Monitoring latency LAT register value for all 3 APV6s This Cancel value times 25ns specifies the pipeline delay VADJ These values select the offset of the analog channel data relative to the baseline individually for each APV6 This does not affect the overall DC offset which is adjusted by potentiometers on the VME module Note that the analog data levels are quite low after power on and continously increase and settle after a few hours of warming up IPRE ISHA IPSP ISFB VPRE VSHA VCAS These settings have an effect on the APV6 internal bias voltages and currents for preamplifier and shaper The pulse shape depends more or less on all of these settings The pulse shape with the standard settings are quite close to the optimum Judging the effect of these parameters is difficult and can only be done when looking at the entire pulse shape Furthermore the peaking time depends on some of these parameters which requires to adjust the timing NIM delay Refer to the APV6 User Manual for details CLVL CSKW CDRV These settings only affect the internal calibration Consult the APV6 User Manual for further information Pedestal Events In the online analysis the first N events are taken for pedestal mean and sigma spread calculation Only subsequent events are analyzed for hits N is specified here
23. ming a CMC calculating pedestal mean and sigma values a hit profile single channel histograms and a hit histogram However only the raw data is written to disk if enabled An information file with settings and temperatures over time is written in parallel Before starting a DAQ run you must at least once open the Settings panel This pop up window allows to adjust the APV6 bias settings as well as analysis and display settings When opening for the first time standard settings appear which should work fine in a testbeam All parameters can be adjusted while the DAQ is running However not all settings have an immediate effect Especially the update cycle settings can be changed at all times having an immediate effect on the display refresh rate APV6 Vienna Manual How to handle DAQ and analysis hard and software 18 The controls will now be explained in detail Settings APY 6 Settings Online Analysis Settings Mode Selects Peak 7 or Mode gt IE Latency 31 Pedestal Events 2 200 Deconvolution 5 modes J Hit above 2 4 00 sigma 7 for all 3 APV6s Note that eer ns ib z i I Pedestal Gaussian Fit the l analog l signal is PRE mi vsHa g o E positive F In the ISHA SIBE E a ol Histogram update cycle 10 0 deconvolution mode but IPSP 84 CLVl 255 e negative in the peak Ee mine ite L 10 mode Refer to the APV6 Temperature I update cycle 60 0 VPRE 150 cDRV 40 User Manual for ot
24. n number of clock cycles default value is approximately 1 8us can be adjusted with DIP switches and then sends the trigger to the repeater The repeater echoes clock and trigger and with these echoes a certain number 1024 is default can be adjusted with DIP switches of analog signal values are digitized with 40MHz and stored in the FIFO memories which are then read out by the DAQ software The used FIFOs allow writing to and being read out at the same time The internal delay time must match with the APV6 latency value in order to get the correct output The A D conversion of the VME module does not depend on the cable length since clock and trigger signals are echoed by the repeater The only thing to adjust is the exact timing of the ADC sampling relative to the echoed clock which is adjustable by DIP switches The analog signal output of the APV6 is 20MHz Since the A D conversion is clocked with 40MHz there are two samples for each analog value In most cases a mean value is calculated in the analysis The module also provides a NIM input and a NIM output for general purpose The use of these terminals are software specific In the APV6 DAQ software the input is used to gate the analysis and file writing while the output when plugged to the trigger input allows software triggers see software description Furthermore there are 4 potentiometers for adjusting the analog signal baseline offsets Refer to the schematics and to M Per
25. nicka for details on the DIP switches VME amp I C This module provides a VME to C interface using standard bus master and current amplifier components Clock and data signals are decoupled by an optical link A flat ribbon cable 30 m is used to connect this module to the repeater Not only clock and data signals are transferred through this cable but also the power for the receiver on the far side This galvanic isolation eliminates possible sources of noise The length of the cable is irrelevant for timing issues However distances larger than 30 m may require a reduction of the transmission speed 400 kbit s Delay CTR This module can be connected to one or several special NIM delay and or attenuator units A unique two digit number must be selected on the front panel of each unit It is possible then to individually set the delay and attenuation parameters by software This comes in very handy when delay curves must be recorded As the NIM delay boxes are equipped with cable delay lines these devices are accurate and fit for both digital and analog signals The VME Delay CTR together with 2 or 3 NIM delay units and a NIM attenuator unit currently are not used in the normal DAQ software but only with the recording of signal shapes The delays are used for shifting the through the calibration pulse while the attenuator simply turns it in or off NIM Modules and Scintillator Photomultipliers No specific description will be giv
26. of the divider the top resistor although only a fraction of pF together with the 1000 times higher voltage step feeds much more charge into the amplifier than the real capacitor Furthermore the final division stage should be terminated and as close as possible to the amplifier input Considering all these items the realized network seems to be a good choice The actual layout of the calibration pulse injection circuit is really critical in terms of noise The final divider stage and the capacitor are therefore SMD components placed closely together In order to calculate the injected charge Q CV one has to measure the voltage step for the Q CV relation The nominal capacitance value is 8 2 pF While the real actual value of such capacitors usually is somewhat higher stray capacitance against ground decreases the value thus the nominal value seems to be a good deal It is impossible to measure the voltage step which is in the order of 1 mV directly at the capacitor terminal However it may be possible to probe the voltage at the indicated point selecting 20 MHz bandwith on the scope If this is still too low the MAX435 output must be measured In any case the actual voltage step must be calculated according to the division ratio The linear range of the APV6 is limited to 5 MIPs Then the calibration capacitance is increased the voltage division ratio must also be adapted in order to get the same charge Consult M Pernicka for furthe
27. p 5 MSE Sigma 28 68 Print Off CMC Single Strip Histogram 1960321 Scale 3769 49 to file CMC Hit Sum Histogram Sa Begin Time 08 021999 16 21 18 End Time 09 02 1999 09 18 39 APV6 Eror Codes 3 3 3 Status Finished Tempeatures C 1 29 1 3 243 4 23 8 Detector V I Monitor 100 0 V 646 1 nd APV6 Vienna Manual How to handle DAQ and analysis hard and software Settings Start DAQ Stop DAQ Help Quit uu pag Write File On Off Max Events 0 inf I T Curent Event 2135430 Progress Pieepeeepecepeespeens 0 20 40 60 80 100 Trigger Rate Hz 200 400 o 700 1000 950 9300 850 800 750 700 650 600 550 500 450 400 350 300 250 200 150 100 50 ELUENT 1000 1010 1020 1030 1040 1050 1060 1070 1080 1090 1100 1110 1120 113 _min s Graph Boundaries 1000 x max 1200 Y_min 0 Y_max 4 1000 i ll HHI AA IAA aaa TTT E STRAT UAT WATT HAA T Umar TTT i pam HA AAA na ae ATTA wiht x Cursors Freeze Raw ADC Data a a fz 1043 52 xi 1136 92 Re ordered Strip Data E 533 05 vE 162 05 Pedestal Subtracted CMC Strip Data File Name temp o a Begin Time 08 02 1999 18 21 18 End Time 09 02 1999 03 18 39 Print Screen Print to file APV6 Eror Codes 3 3 3
28. ps the guard ring of each detector zone is connected to the ground rail Each strip is equipped with an integrated FOXFET bias resistor between the strip and the guard ring Due to the guard rings the first and last strips of each zone strip numbers 04 32z 31 32z z 0 1 2 11 have a higher ground capacity than mid zone strips causing excess noise Thus only strips 1 30 in each zone can collect hits Taking the inactive regions into account one expects a collection efficiency in the order of 70 In practice the fraction depends on the hit cut parameters since signal and pedestal are not clearly separated with a SNR 12 which is roughly achieved with the APV6 The strip capacity of the current multi region detector is between 8 and 12 pF depending on the geometry In the final design of the CMS tracker two detectors are connected in series giving a 6 25 x 12 5 cm area In that case the detector strip capacity is twice that of the single detector causing more noise The strip pitches in the CMS tracker will be 60 80 120 and 240 um from the inner to the outer layers giving a tradeoff between spatial resolution and the number of channels The current multi region detector was designed to examine the various implant widths The thickness of the detector is 300 um it is based on an n type bulk and backplane with p implants for the strips The cable for the bias voltage ends with a BNC connector which must be fed with positive volta
29. r details on the APV6 and other electronic parts APV6 Vienna Manual How to handle DAQ and analysis hard and software 9 Repeater The repeater is a rather simple circuitry performing a few tasks It buffers both incoming and outgoing signals and it provides the voltages for the APV6 2V the digital and analog buffers 5V and the I C buffer 3V 2V The clock and trigger signals sent to the APV6 are not only buffered but also sent back to the APV6 Read 2 VME card This feature allows some independence on the timing In fact the VME board does not need to know the exact timing which is only specified by the LAT register value coarse and the NIM delay fine Note that the repeater especially the voltage regulators dissipate a lot of thermal power In the current setup whole repeater is included in the cooling box However would rather suggest to place them outside of the cooled environment in future designs Keithley 237 SMU and Mac488B The Keithley 237 Source Measure Unit is used to provide the detector bias voltage s Detector It must be used with a special BNC TRIAX interface box with the TRIAX cable connected to the OUTPUT HI terminal at the rear of the device First set the compliance value Set the output voltage to O and push the OPERATE button Always remember that only positive voltages are allowed Choose the 1V digit with the SELECT buttons and ramp up the voltage with the knob keeping an eye on the
30. relevant Thus the logic could be even simplified However it is more consistent to use a standard setup even in this case APV6 Vienna Manual How to handle DAQ and analysis hard and software 15 With the calibration run the timing is essential With the default internal trigger delay of the VME module the latency register setting the coarse timing must be set to the values shown in the graph The fine tuning is done by the NIM delay and depends on the bias settings The graph shows a NIM delay value which is close to the optimum but need not be exactly at the peak Also there could be a slight variation of the optimum delay with time or temperature Thus recommend to check the optimum after power off periods To calibrate the setup one has to find a relation between charge electrons and ADC counts This is done by measuring the injected voltage step as described in the Hybrid section On the other end one has to look at the pedestal and calibration pulse histograms to calculate the difference in ADC counts The DAQ software is able to plot histograms of single channels In our case the calibration circuit is connected to channel 2 The easiest way to get a plot with both pedestal and calibration pulse distributions is to start a run with N events at open delay loop and close the delay loop approximately after N 2 events With the cursors in the centers of each peak one can easily calculate the ADC difference Settings
31. source and testbeam runs APV6 Vienna Manual How to handle DAQ and analysis hard and software 16 Source and Testbeam Setup ogee 80MHz NIM D lay _2ns_ NIMShaper __ ens CaLTriSyn DG Trigger isis width 5ns CLK 2ns VME APV6 Read2 IN_NIM 4 Beam on NIM Coinc Deconvolution Mode 5 LAT 91 Default Bias Settings ons ons For Peak Mode 7 LAT 88 and 21m NIM Delay must be optimized APV6 gt lt HV supply NIM D ance 21m Sc PMs WID 20ns Cable propagation times are irrelevant when omitted With this setup almost every cable propagation time contributes to the timing Moreover different NIM devices may have different timing so stick to a standard set of units Principally the source and testbeam setups are exactly identical however the exact timing and the discriminator settings may slightly vary Remember to set the coarse timing with the latency register and the fine tune it with the NIM delay Often testbeams have a spill structure i e there is periodically beam for some seconds and then no beam for another couple of seconds In this case the IN_NIM terminal can be fed with the spill on off signal The DAQ software then if gating is enabled entirely concentrates on data taking during spill on The data is temporarily buffered in memory and only at spill off it is written to
32. specifies the path where to save data and information file Use normal slashes instead of backslashes here Keithley 237 V I Monitoring Enables or disables the detector voltage current readout of the Keithley SMU This feature requires a Mac488B interface After closing the settings panel the program is ready to begin a DAQ run However the user may want to enter some additional settings on the main panel All items to be entered or pag Write File On A Off Mai 0 Cur 0 Wu accepted before starting a DAQ run are colored in a light green 5 32777 Settings 30000 Start DAQ 28000 26000 Stop DAQ 24000 Help 22000 We 20000 Quit 18000 16000 File 14000 12000 Trigger T SW x Events h A d ment Event 2135430 10000 8000 6000 4000 2000 Progress a a LLL l l l l Tce ee O O A S 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 x 20 40 60 80 100 Trigger Rate Hz 200 400 Graph Boundaries Cursors Freeze Gaussiz 3 Raw ADC Data Online S 0 Ju pen o Bima 250 x g 80 98 Re ordered Strip Data Analysis Dorsk Besse Y_min o Ymax 3 32777 fe 21874 63 98 ye 10902 37 Pedestal Subtracted CMC Stip Data OSPY pee
33. the pipeline architecture the triggers may appear up to 160 x 25 ns 4 us after the corresponding event This complicated feature becomes a necessity in the LHC design as the trigger decision is not possible within two bx moreover not even signals from opposite ends of the CMS detector come together within this time After receiving a trigger the APV6 puts out an element of the pipeline which is a certain number of bx or clock cycles in the past This value is called latency LAT and is configured during initialization The shaper ouput is sampled and stored in the pipeline ring buffer with every clock cycle In the peak mode these values are sent to the output directly while in the deconvolution mode 3 subsequent pipeline values are mixed together in a switched capacitor filter APV6 Vienna Manual How to handle DAQ and analysis hard and software 7 network ASPS resulting in a narrower signal shape which allows the separation of signals from subsequent bx The signal output shape cannot be monitored directly as it is sampled when entering the pipeline 300 However scanning curves 250 can be recorded either using 200 the internal calibration delay see APV6 user manual or by sequentially shifting the external calibration delay The output shape depends on the seven bias parameters and of course the mode Sa a aT PT Note that the output polarities t Ins differ in peak and deconvolution modes In our setup de
34. tics it simply refuses the fit In these cases a NaN not a number value is displayed for one or more of the fit parameters and the fit curve is not plotted APV6 Vienna Manual How to handle DAQ and analysis hard and software 21 A Print Screen button allows to send the current main panel to a printer With the Print to File button a PostScript file is saved containing the current main panel provided that the default printer is a PostScript printer Settings Start DAG Stop DAQ Help du Quit pag Write File Trigger On v Hw off SW Max Events o m a l Curent Event 2135430 Progress Peevpeeepecegeerqereg 0 20 40 60 80 100 Trigger Rate Hz 200 400 0 700 7 File Name temp o en Ss 28342 1 T 26000 L 24000 22000 20000 J 18000 16000 t 12000 8000 E TT i all DA OA O O ee fee 0 20 40 60 80 100 120 140 200 220 240 260 280 300 320 340 360 383 Graph Boundaries Cursors Freeze Raw ADC Data Online 3 a a 5 emin 3 xmag 383 fe 60eg fxj 20 35 Re ordered Stip Data i Analysis oe prac Tie Arj v mas 28342 uE 15129 62 uE 4632 12 Pedestal Subtracted CMC Strip Data ey T i I 2 Auto Cursors CMC Pedestal Mean vs Strip En CMC Pedestal Sigma vs Strip Now Mean 203 72 ashe Gated Analysis and File Writing A ga Eei vs Stri
35. tp www hep ph ic ac uk leb97 APV6java html The 3 cell mixing in deconvolution mode APV6 Vienna Manual How to handle DAQ and analysis hard and software 8 also implies that the latency time for equivalent timing must be set 3 counts larger than in peak mode looking farther back in the past In our setup the three chips are numbered with channels 0 127 128 255 and 256 383 from left to right corresponding to the detector strips when looking at the hybrid facing the repeater All channels of the 3 APV6 chips are connected to a detector strip except for channels 0 2 on the first chip While O and 1 are left open channel 2 is used for calibration It is bonded to an 8 2 pF capacitor with a voltage divider to which a small voltage step is applied when calibration is enabled Point to measure AV MAX435 meets 4 22k0 y APYE gain AN 56Q 390 RY current source L J i Dg output ma a 9 6809 cLcen4 L The output of APV6 0 is terminated with 308Q WHY The capacitor should have a value similar to the detector capacitance Using Q CV one can derive really small voltage steps needed for charge inputs in the order of 1 MIP 22400 e To achieve this an extended voltage divider lies in front of the capacitor Experience has shown that a voltage division ratio of say 1 1000 cannot simply be realized by a 2 resistor network because the stray capacitance between input and output
36. ule see Hardware section no data is acquired and the error codes are 2 Once a run is started and also after it is finished a lot of histograms can be viewed In plots which display data from all 3 APV6 chips these 3 data blocks are simply concatenated corresponding to the total 384 detector strips The displayed plots either refer to the current event Raw ADC data Re ordered strip data Pedestal subtracted CMC strip data or represent cumulated values The plots are automatically refreshed in intervals specified in the settings except if the Display is set to Freeze which does not affect the data acquisition X and y axes can be manually set at the Graph Boundaries However as the y axis of the plots is automatically scaled with each display refresh the manual settings are quickly gone unless the plot is freezed 2 Cursors can be moved to any position by either entering the coordinates in the x and y boxes or simply moving the cursors around with the mouse pointers A Gaussian fit can be applied to the CMC hits vs strip CMC single strip histogram and CMC hit sum histogram plots The fit curve can be updated manually or automatically with each plot refresh Moreover the fit boundaries can either be calculated automatically or set manually with the 2 cursors Mean sigma and scale parameters and the mean standardized error are displayed However sometimes the fit routine does not work correctly Especially with very high statis
37. ware The temperature is calculated from DC voltage levels with a parabolic fit function which has been calibrated at three temperature points in the interesting range two points inside a refrigerator and one at room temperature Setups Generally each setup must warm up for at least a few hours after power on For precision measurements recommend to run all components of the system overnight to settle at operating temperatures The APV6 internal bias generator is not enabled at power on It is switched on every time a DAQ run is started Although it is not necessary recommend to run the DAQ overnight without writing the data to disc to accelerate the warming up procedure Especially when using the cooling box it takes quite a while to approach stable temperatures Pedestal Calibration Setup The picture below primarily shows the NIM logic for pedestal and calibration runs Pedestal and Calibration Setup Clock 80MHz NIM Clock NIM Shaper eee Da Trigger 60kHz width 5ns CLK A 4ns NIM D lay Cal Out2 VME APV6 Read2 or Nee ans Cal In2 A Open delay loop to turn off Cal Pulse Deconvolution Mode 5 LAT 76 Default Bias Settings A 8 2pF 21m CF For Peak Mode 7 LAT 73 and Se NIM Delay must be optimized APV6 1 Cable propagation times are irrelevant when omitted For the pedestal run recording pedestals and noise the timing is ir
38. z and sent to the APV6 Furthermore the connection to the repeater with a special cable orange DELPHI cable nimout C OUT NIM which is 21m long is essential Do not use other cables i since the termination resistors of the signal lines are n VETO matched with the cable s impedance Z 103 5Q Whenever the length of this cable is amended the timing values specified in this paper must be adapted The 34 pin 3M propagation time of 1m of cable is approximately 5ns connector NIMIN IN NIM Repeater The module has a veto switch which is usually turned on This feature ensures that an incoming trigger sets a veto for further triggers until it is reset by the software after reading out the data However with DIP switches on the board the number of subsequent triggers allowed before vetoing can be selected to be 1 15 Ch1 DC offset Ch2 DC offset Ch3 DC offset Ch4 DC offset Oo o There are 3 different trigger inputs on the front panel TRIG accepts a trigger at any time while CAL TRI does the same but additionally produces a calibration pulse which need not reach the APV6 With CAL TRI_Syn a trigger is only accepted at the clock edge others are rejected This feature emulated the synchronized behavior of the final CMS operation where bx occur synchronously to the 40MHz clock This mode is essential for all measurements where a signal is measured however not necessary for noise
Download Pdf Manuals
Related Search