UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An on-line computer assisted mass spectrometer Mitchell, David Laurie 1971-04-29

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
831-UBC_1971_A6_7 M58.pdf [ 3.43MB ]
Metadata
JSON: 831-1.0053430.json
JSON-LD: 831-1.0053430-ld.json
RDF/XML (Pretty): 831-1.0053430-rdf.xml
RDF/JSON: 831-1.0053430-rdf.json
Turtle: 831-1.0053430-turtle.txt
N-Triples: 831-1.0053430-rdf-ntriples.txt
Original Record: 831-1.0053430-source.json
Full Text
831-1.0053430-fulltext.txt
Citation
831-1.0053430.ris

Full Text

AN ON-LINE COMPUTER ASSISTED MASS SPECTROMETER by DAVID LAURIE MITCHELL B.Sc, King's College London, 1968 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in the Department of GEOPHYSICS We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1971 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver 8, Canada 0 ABSTRACT An analog data acquisition system incorporating an Interdata Model 4 digital computer has been designed and built for a mass spectrometer. This system has been conceived with the primary objectives of improving analytical precision and production. Automated mass spectrometer operation allows for the collection of larger quantities of data while decreasing operator involvement and consequently diminishing operator bias and fatigue. The analog signal from the mass spectrometer measur ing system is digitized using a digital voltmeter and trans mitted, via an interface, to the processor where the digital information is manipulated in accordance with a computer pro gram. An additional facility is provided whereby digital data from the processor can be displayed, if desired, on a 5 decade numerical readout situated at the mass spectrometer console. Hardware is also available in the interface to provide control of the magnetic field scan rate. A function control switch at the mass spectrometer console allows the operator to convey a variety of predetermined instructions to the computer at any time during the course of a run. Thus, the system provides for on-line (real time) data processing of mass spectra as well as limited computer control over the mass spectrometer. This thesis is primarily concerned with the design and construction of the logic hardware for this system together with a demonstration of its operating ability. iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iiLIST OF FIGURES v LIST OF TABLES vi ACKNOWLEDGEMENTS viCHAPTER 1. INTRODUCTION 1.1 General Background 1.2 Design Objectives CHAPTER 2 GENERAL DESCRIPTION 2.1 The Mass Spectrometer 2.2 The Computer 2.3 Programming the Interdata Model 4 CHAPTER 3 THE INTERFACE HARDWARE DESIGN 3.1 Introduction 3.2 Device Addressing 3.3 Data and Status Input 3.4 Data and Command Output 3.5 Interrupt Control 3.6 Read/Write Sequencing 3.6.1 Read Operation 3.6.2 Write Operation 3.7 The Analog-to-Digital Converter 3.8 The Numerical Display 3.9 Construction of the Interface CHAPTER 4 ON-LINE FILTERING OF DATA CHAPTER 5 SYSTEM PERFORMANCE 5.1 Introduction 5.2 Preparation of Strontium Samples 5.3 Analysis of Eimer § Amend SrC03 Interlaboratory Standard 5.4 Analysis of Strontium in a Rock Sample CHAPTER 6 CONCLUSIONS BIBLIOGRAPHY APPENDIX I Ion Current Amplifier and Output Attenuator Circuit II Scan Drive Circuit III Summary of Interdata Programming Instructions IV Mass Spectrometer Interface Programming Guide V LIST OF FIGURES Page FIGURE 1. COMPUTER TO MASS SPECTROMETER INTERFACE 4 FIGURE 2. ANALOG DATA ACQUISITION SYSTEM BLOCK DIAGRAM 8 FIGURE 3. THE MULTIPLEXOR CHANNEL 10 FIGURE 4. BASIC LOGIC FUNCTIONS AND TRUTH TABLES 16 FIGURE 5. DEVICE ADDRESSING, LOGIC DIAGRAM 21 FIGURE 6. DATA AND STATUS INPUT, LOGIC DIAGRAM 23 FIGURE 7. DATA AND COMMAND OUTPUT, LOGIC DIAGRAM 24 FIGURE 8. INTERRUPT CONTROL, LOGIC DIAGRAM 26 FIGURE 9. COMPUTER/MASS SPECTROMETER INTERFACE READ/WRITE LOGIC 29 FIGURE 10. EXTERNAL CONNECTIONS TO READ/WRITE SEQUENCING LOGIC 30 FIGURE 11. READ SEQUENCING INSTRUCTIONS 32 FIGURE 12. WRITE SEQUENCING INSTRUCTIONS 32 FIGURE 13. ANALOG-TO-DIGITAL CONVERTER BLOCK DIAGRAM 35 FIGURE 14. DUAL SLOPE WAVEFORMS FOR DIGITAL VOLTMETER 6 FIGURE 15. THE NUMERICAL DISPLAY 38 FIGURE 16. MASS SPECTROMETER FILTER -DISPLAY PROGRAM 43 FIGURE 17. FREQUENCY RESPONSE OF DIGITAL FILTER 46 vi LIST OF TABLES TABLE I DATA PROCESSING EQUIPMENT PURCHASED TABLE II SPECIFICATIONS OF QUAD 2-INPUT NAND FUNCTION FOR THE FIVE LOGIC FAMILIES TABLE III INTEGRATED CIRCUIT PACKAGES USED IN CONSTRUCTION OF INTERFACE LOGIC CIRCUITRY TABLE IV RESULTS OF ANALYSES OF EIMER AND AMEND INTERLABORATORY STANDARD SrCOj TABLE V RESULTS OF ANALYSES OF GREY ARGILLITE Page 7 18 40 54 54 ACKNOWLEDGEMENTS The success of this project owes a great deal to the various people,who assisted the writer. The project was initiated by Dr. R. D. Russell whose thoughtful guidance, enthusiasm and encouragement throughout were always appreciated. Special thanks go to J. Blenkinsop who wrote the necessary computer programs, imparted much useful advice, and was responsible for the well-functioning of the mass spectrom eter. The writer is indebted to the technical staff of the Department of Geophysics, U.B.C, in particular C. Croucher, R. D. Meldrum, and E. J. Bellis whose specialized skills made a large contribution to the project. Thanks must also go to B. D. Ryan who provided the writer with much needed geological enlightenment and to Dixie Pidgeon for typing this thesis. The Data Technology Corporation and Interdata Incor porated, very kindly gave permission for the reproduction of several pages from their equipment manuals. The entire computer system was purchased from funds granted to Dr. R. D. Russell by the National Research Council of Canada, Mobil Oil Canada, Ltd., and the Standard Oil (Indi ana) Foundation, Incorporated. Finally, the writer wishes to thank Dr. T. J. Ulrych whose critical appraisal of the manuscript in its final stages made for a greatly improved thesis. CHAPTER 1 INTRODUCTION 1.1 General Background During the past few years a number of important advances have been made in the field of isotope geophysics. New analytical techniques, such as the double spiking pro cedures of Compston and Oversby (1969) and the triple-filament technique for lead (Catanzaro, 1967), benefit from a high order of precision in mass spectrometer analysis. In order to do full justice to these improved analytical techniques it is necessary to increase the measurement precision of mass spectrometer ion-currents. One way in which this can be achieved is by means of digital data collection and analysis procedures. Using an on-line digital computer, Wasserburg and his associates at the California Institute of Technology have demonstrated the value of these procedures by achieving results of outstanding precision on analyses of strontium (Papanastassiou and Wasserburg, 1969) and gadolinium (Albee et al, 1970) . At the University of British Columbia, research of digital data collection systems commenced in 1963, using a gas-source mass spectrometer. A servo-voltmeter ion-current amplifier (Stacey et al, 1965) was utilized which had, as a primary output, the shaft rotation of a motor-driven potent-2. iometer. Thus analog —to-digital conversion was easily achieved using a shaft position encoder (Weichert et al, 1967) . Digital data was initially stored on punched paper tape but, as the system was developed, the paper tape punch was replaced by an incremental magnetic tape recorder. In both cases, the data was processed using the facilities of the University's Computing Centre which presently include a duplex IBM 360/67 computer. To maintain completely independent data acquisition systems for the three mass spectrometers operated by the Department would have required the purchase of two more tape recorders, and the construction of suitable interfaces. A small digital computer was therefore purchased in 1969 and work began on the design of a suitable digital logic interface to a mass spectometer. The on-line data acquisition system that was subsequently built is capable of immediately and sim ultaneously supervising the operation and processing the data from the three mass spectrometers, which are used at different times for isotopic analysis of Pb, Th, U, Rb, Sr, Gd, Eu and Sm. The system description presented in this thesis is intended as an example of what is possible and no claim is made that the design is optimum in any sense. 3. 1.2 Design Objectives The system was conceived with three main objectives in mind: 1. A decrease in the overall time required to process the data from the mass spectrometer. 2. A reduction in operator involvement and hence a reduction of operator bias and operator strain. 3. An increase in analytical precision. These objectives have been realized by using an on-line digital computer interfaced to the mass spectrometer. The computer should be able to filter the incoming digital data in real time and display the filtered data points at the mass spectrometer console for the operator's convenience. In addition, it should be possible to reduce the mass spectra from the input data and display the results at his request. The computing system should have the capability of automatically adjusting the operating conditions of the mass spectrometer. The system built has the facility for controlling the magnetic scan rate but additional control features may be added at a later time. The writer's research was primarily concerned with the design and construction of the necessary digital electronics for the interface together with a demonstration of its operating ability. A simplified block diagram of the computer to mass spectrometer interface is shown in Figure 1. SCAN READ PARAMETERS FUNCTION SELECT SWITCH READ SCAN RATE CONTROL WRITE ION EEATT ION CURRENT AMPLIFIER 4-DIGIT DIGITAL VOLTMETER DATA READY ± CHART RECORDER DISPLAY^ WRITE INTERFACE "HsECTION-2 4 3 feet CABLE INTERFACE ION-1 ^SECT] INTERDATA MODEL-4 DIGITAL OMPUTER % V PEC 7-TRACK SYNCHR. TAPE RECORDER FIGURE 1. COMPUTER TO MASS SPECTROMETER INTERFACE 5. CHAPTER 2  GENERAL DESCRIPTION 2.1 The Mass Spectrometer The mass spectrometer used in this research project was designed principally by R. D. Russell in cooperation with F. Kollar, J. S. Stacey and T. J. Ulrych, and was brought to its present state of operating efficiency by J. Blenkinsop. It is a 90-degree sector, 30 cm radius, solid source machine, with single order focussing and a variable magnetic field scan. In the past and at present this machine has been used for rubidium-strontium analyses but it is in no way limited to these elements. The analog section of the mass spectrometer measuring system incorporates a hybrid amplifier using an electrometer vacuum tube and integrated circuit components. The voltage output from this amplifier can be attenuated by factors of 1, 1/3, 1/9, 1/27, 1/81 using a shunt switching network. The schematic circuits for the ion current amplifier and output attenuator, both designed by R. D. Russell, can be found in Appendix I. The analog output of the measuring system, which is generally of the order of one volt for most isotope peaks, is fed to a digital voltmeter which displays the input voltages and also functions as an analog -to-digital converter for the computer interface. 6. 2.2 The Computer An Interdata Model 4 computer was chosen chiefly because of its versatility at a reasonable cost. In addition, it uses a halfword length of 16 bits and a programming language that is similar to the University of B.C. IBM 360/67 computer with which it can be easily interfaced if required. The 8k 8-bit bytes of memory supplied with the Model 4 Pro cessor are marginally adequate for sophisticated mass spectrom eter control and data reduction programs, however additional memory modules may be added, up to a maximum of 65k bytes, at any time. A Teletype Model ASR33 and Peripheral Equipment Corporation PEC 3520-72 synchronous tape transport were pur chased with the Model 4 Processor (see Table I). The teletype was supplied ready interfaced to the computer and the tape transport was interfaced in our laboratory by R. D. Meldrum using logic circuitry of his own design. Figure 2 shows the complete analog data acquisition system for one mass spectrometer. The processor can communicate with the teletype and mass spectrometer via a multiplexor channel which has the capacity to handle a total of 256 devices. Each peripheral device is connected via a device controller to the multiplexor bus, it is this controller that provides the interface between the computer and an external device. The only essential requirement of the device is that it can supply data to the device controller in an acceptable digital format. In 7. TABLE I DATA PROCESSING EQUIPMENT PURCHASED ITEM MODEL NUMBER MANUFACTURER APPROXIMATE COST (1969) PROCESSOR INTERDATA INC., $7,800 * 8,192 BYTE MEMORY INTERDATA INC., $6,000 * SELECTOR CHANNEL INTERDATA INC., $2,900 * HIGH SPEED ARITHMETIC § INPUT/OUTPUT INTERDATA INC. , $1 ,500 * TYPEWRITER $ PAPER TAPE READER/PUNCH ASR33 TELETYPE CORPORATION $1,900 MAGNETIC TAPE TRANSPORT PEC PERIPHERAL EQUIP- $4,000 3520-72 MENT CORPORATION DIGITAL VOLTMETER NUMERICAL DISPLAY POWER SUPPLY DT-344-2 DATA TECHNOLOGY $ 500 CORPORATION DS-103-5 DISPLAY GENERAL $ 300 INCORPORATED PS-200 DISPLAY GENERAL $ 60 INCORPORATED * - 1971 PRICES ARE ABOUT 401 LOWER THAN 1969 PRICES 8. CORE MEMORY HIGH SPEED MEMORY BUS PROCESSOR MULTIPLEXOR CHANNEL ~2W MULTIPLEXOR BUS SE TS 7K DEVICE CONTROLLER 32. TELETYPE DISPLAY SCAIM RATE CONTROL J DEVICE"^ CONTROLLER*1 DATA READY DVM SHUNT SELECTOR ION CURRENT AMPLIFIER MASS SPECTR-OMETER. SELECTOR CHANNEL SELECTOR BUS DEVICE ^ONTROLLEF; READ M.S. PARAMETERS MAGNETIC TAPE Figure 2. Analog Data Acquisition System Block Diagram the case of the mass spectrometer this is achieved by using the binary-coded-decimal (BCD) outputs of a digital voltmeter. Figure 3 shows the 27 lines that constitute the multiplexor bus; 16 lines are reserved for data input and output which is matched to the 8-bit byte. The 8 control lines, with one exception, are used to enable data onto the data lines. The control designated ACKO, together with the attention line (ATNO) and the system clear line (SCLRO), are associated with the processor interrupt facility which is an optional feature of every device controller. The synchronize line (SYNO) is used to notify the processor when a signal on one of the control lines is accepted by a device controller. A typical sequence of operations to service the mass spectrometer over the multiplexor channel would be: 1. A pulse from the digital voltmeter, signifying that data is available at the mass spectrometer, causes a signal to be sent along the attention line (ATNO) which interrupts the processor. The processor acknowledges the interrupt and sends a signal over the ACKO line which initiates a hardware scan cycle to determine which device caused the ATNO signal. The mass spectrometer automatically returns its device number to the processor over the data request lines (DRL's). The processor is now ready to service the mass spectrometer. 2. The mass spectrometer is addressed by the processor over the 8 data available lines (DAL's). This address appears on the bus to all device controllers. INTERDATA PROCESSOR V INIT. 8 DAL 1s -8 DRL's MULTIPLEXOR BUS •— 8 CL's • ACKO ^ o o SYNO ^ o < EH RACKO TACKO RACKO TACKO ATNO SCLRO ' INI FAC1 INr FAC1 2-2 [NT! FACE - 3 M.S.I M.S. 2 M.S.3 FIGURE 3. THE MULTIPLEXOR CHANNEL 3. The processor then activates the address control line which signifies that the DAL's now provide an address (as opposed to data). 4. The mass spectrometer interface decodes its address, sets a flip-flop memory, and sends a signal back to the proc essor along the SYNO line. The mass spectrometer remains addressed until another device controller is addressed or until a system clear signal (SCLRO) is received. 5. The processor places an "output command" on the DAL's. 6. The processor then activates the command control line; this causes the data from (for instance) one particular decade of the digital voltmeter, specified by the output command, to be made available. A SYNO signal is sent back to the processor to indicate the command has been stored in the device controller. 7. The mass spectrometer is again addressed by the proc essor, as described in steps 2, 3 and 4. 8. The processor then activates the data request (DR) control line which enables the byte of data, made available in step 6, from the digital voltmeter to the processor, along the DRL's. A synchronize signal (SYNO) is generated to indicate that the data has arrived at the processor. 12. 2.3 Programming the Interdata Model 4 The Interdata Model 4 computer possesses a repertoire of 75 basic instructions which can manipulate data between core memory, 16 general registers, and up to 256 external devices. In addition, the Interdata system also provides for the direct transfer of a block of data between core memory and a peripheral device under control of an optional selector channel. Once initiated by the processor, this direct transfer takes place in visibly without interruption to normal processing. Data of three different word lengths; the 8-bit byte, 16-bit halfword, and 32-bit fullword, can be manipulated by the instruction set. In the Interdata system hexadecimal notation (base 16) is used to express binary information, so that, for example, a byte of data can be represented by two hexadecimal digits. Three instruction formats are available in the Inter data system: register to register (RR), register to indexed memory (RX), and register to storage (RS). A total of sixteen 16-bit general registers, numbered Oto F in hexadecimal notation, function as accumulators or index registers in arithmetic and logical operations. In all three instruction formats, bits 0-7 specify the machine operation to be performed (th 8-bit OP code); bits 8-11 specify the address of the first operand, which is normally a general register. In the RR format the address of the second operand is specified by bits 12-15 and is always a general register. In the RX instruction formats, bits 12-15 always specify the address of a general register whose content is used as an index value. The remaining 16 bits (bits 16-31) specify a memory address in the RX format, and, in the case of the RS format, an integer value for use as an immediate operand. A summary of the Interdata Model 4 programming instruc tions is given in Appendix III. The information necessary for program execution is contained in the 32-bit program status word (PSW). Bits 0-11 of the PSW define the status of the current user program; bits 12-15 constitute the 4-bit condition code (CC) which is set after execution of input/output, logical, shift or arithmetic instructions. The memory address of the next instruction to be executed is specified by the 16-bit instruction address field (bits 16-31 of the PSW). In instances of machine malfunctions, divide faults, illegal instructions, and external device service requests, system interrupts are generated. When an interrupt is recognised, the current PSW, which defines the present operating status of the processor, is placed in a reserved storage area (the old PSW) and a new PSW re-defines the status of the machine. On com pletion of the interrupt service sub-program the previous machine status, stored in the old PSW, is restored. Input/output data transfer in the Interdata system can be either program controlled or interrupt controlled. The former method interrogates the device to ascertain if it is ready to transfer data, and waits if necessary until transfer can occur. The interrupt method allows the device to demand service when it is ready for the transfer of data. This latter method is the one employed for the mass spectrometer interface. CHAPTER 3 THE INTERFACE HARDWARE DESIGN 3.1 Introduction. The construction of a successful computer interface involves the design of suitable logic circuitry which will enable the computer to communicate with the particular peri pheral device. The hardware design is evolved using the standard logic elements available, namely gates, inverters and flip-flops. Figure 4 summarises the most commonly used logic elements together with their respective truth tables. In general, gates are used to direct the signals to and from the computer and flip-flops are used to store information, acting as one-bit memories. Logic circuit design is in general easier than most analog circuit designs because only two voltage levels are employed, and the only major concern is the duration and timing of these voltage signals. There are two classes of signal, the steady voltage level and the pulse, the latter having a duration of a few tens of nanoseconds (typically). The timing of both types of signal are controlled by the processor and the peripheral device. 16. INVERTER &MC834P) AND GATE &MC1806P) NAND GATE &MC849P) OR GATE &MC1809P) A_ B NOR GATE A_ B CLOCKED FLIP-FLOP (MC845P) C_ J-K FLIP-FLOP &MC855P) t J Q C K TRUTH TABLES A B f 0 1 1 0 0 0 0 0 1 0 1 0 0 1 1 1 0 0 1 0 1 1 1 0 1 1 1 0 0 0 0 0 1 1 1 0 1 1 1 1 0 0 1 0 1 0 1 0 0 1 1 0 S-n, R* C> 0 0 0 1 0 1 0 1 1 1 0 0 Cv 0 1 0 1 0 1 I 1 FIGURE 4. BASIC LOGIC FUNCTIONS $ TRUTH TABLES Once the interface logic circuitry has been designed, by suitably matching the computer and peripheral device spec ifications to perform the required functions, it is then only necessary to select appropriate integrated circuit logic packages (chips) to perform the required task consistently. When choosing logic packages there is normally a choice between five different logic families namely, resistor-transistor logic (RTL), diode-transistor logic (DTL) , transistor-transistor logic (TTL), high-threshold logic (HTL), and emitter-coupled logic (ECL); Table II compares the 2-input NAND function for the five logic families. ECL is a non-saturating form of logic which eliminates transistor storage time as a speed limiting characteristic, it is used where extremely high speed operation is required. HTL was developed for applications such as in industry requiring a higher degree of inherent electrical noise immunity than is available with the more standard integrated circuit logic families. Disadvantages of HTL are a relatively high supply voltage (15±1 volts) and slow speed. TTL is a medium speed, high noise-immunity family of saturating integrated logic circuits, it is presently the most commonly employed logic family. DTL offers moderate speed and good noise immunity, it is somewhat inferior to TTL and used to be less expensive. RTL is slow, has poor noise immunity and a large power dissipation, it is no longer used commercially in high quality digital electronics. TABLE II SPECIFICATIONS OF QUAD 2-INPUT NAND FUNCTION FOR THE FIVE LOGIC FAMILIES LOGIC FAMILY TYPICAL MOTOROLA NUMBER SUPPLY VOLTAGE VOLTS OUTPUT LOADING FACTOR RELATIVE NOISE IMMUNITY PROPAGATION DELAY ns TYPICAL TOTAL POWER DISSIPATION mW TYPICAL/PKG. APPROX. COST/ PKG. RTL MC9714P +3±.3 WORST 55 145 $1.55 oo DTL MC849P +5±l/2 7 MODERATE 25 66 $1.65 TTL MC7400P +5±l/2 10 GOOD 13 40 $1.29 HTL MC672P +15±1 10 BEST 110 114 $1.40 ECL MC1048P -5.2±l/2 25 BAD 130 $3.90 Diode-transistor logic was chosen for the mass spectrometer interface because its specifications are more than adequate for the relatively slow speeds involved and because, at the time of purchase, it was slightly less ex pensive than the equivalent transistor - transistor logic, it is also the logic family most used in the Interdata Model 4 Processor. Where a particular logic function was not avail able in DTL then the appropriate TTL function was employed. An aggravating problem often encountered with digital logic circuitry is electrical noise pickup, either, from adja cent lines running in close proximity (interline crosstalk) or, from other sources. The best way of eliminating it is to use high-threshold logic (HTL) which operates at a 7.5 volt threshold level. This logic family is somewhat inconvenient to use, un fortunately, since a separate 15 volt regulated power supply is required together with HTL/DTL level converters on all lines to and from the interface. In addition, HTL is about five times slower than DTL. Using diode-transistor logic, with careful circuit design, it is generally possible to keep noise pickup at a negligible level. Where a circuit element is required to drive a long line, for instance between the processor and a per ipheral device, a power gate and appropriate pull-up resistor are always used. This helps to provide good noise immunity, primarily by lowering the impedance level of the lines. In addi tion, all long lines are made false-active, i.e., a line is active 20. when it is at ground potential (zero volts), to further reduce the possibility of noise pickup. The following pages describe the logic circuitry in some detail. It has been found convenient, for circuit descrip tion, to divide the interface into five sections, namely, device addressing, data and status input, data and command output, interrupt control, and read/write sequencing circuitry. 3.2 Device Addressing The Interdata device addressing logic diagram is shown in Figure 5. The mass spectrometer address (hexadecimal D) is wired into the device number selection board. When this address appears on the data available lines (DALOO through DAL07) the decoded device output (DDO) goes low. A signal on the address control line (ADRSO), in conjunction with DDO, sets the address flip-flop so that its output (DENB1) is made high. During the presence of ADRS1, a synchronize signal is returned to the processor via the address synchronize line (ADSYO). A delay of about 200 nano seconds is produced by capacitor Cl on the synchronize line (SYNO). This prevents the processor from lowering the ADRSO line before the address flip-flop has been set. The device en able line (DENB1) gates all other input/output control lines to the interface. When the address of another device appears on the data available lines, DDI goes low, causing ADRS1 to reset the address DALO 0[> ATS Y NO 0-(r-70-ADRSO o SYNO <r 0 DEVICE NUMBER SELECTION -O -o o o -o DDO DDI AP_RS_1~ :470pF S Q T R ATSYN1 GO G7 0 DENB1 -DAO -D> -i> 4> -0 -0A7 -0 ADDRESS FLIP-FLIP ADSYO ATSYNO -DRSYO ;470pF CDSYO SRSYO —DASYO Figure 5. Device Addressing, Logic Diagram flip-flop, and disable the mass spectrometer interface. The eight NAND gates, GO through G7, together with the lines ATSYNO and ATSYN1, are part of the interrupt control circuitry described in section 3.5 3.3 Data and Status Input Figure 6 shows the Interdata input gating logic diagrams. When the mass spectrometer is addressed, DENB1 is high, enabling the data request (DRO) or status request (SRO) control line. The data or status byte is thus enabled onto the data request lines (DRLOO through DRL07). A return syn chronize signal, DRSYO or SRSYO, is automatically generated when either of the control lines is enabled. The lines AO through A7 connect to the eight gates, GO through G7, shown in Figure 5 and form part of the interrupt controller described in section 3.5. 3.4 Data and Command Output The circuit shown in Figure 7 is used to control the flow of data and commands from the processor. When the mass spectrometer is addressed, DENB1 is high, enabling the data available (DAO) or command (CMDO) control line. DAO or CMDO, in turn, enable the data or command byte onto the data available lines (DALOPO through DAL0P7). In addi tion, control pulses are sent to the read/write logic (Figure 9) 23. DENB1 D SRO 0— DRO 0— <r <r <r <r 7<r SRSYO DRLO oo a p +5v DRSYO <3 +5v 1470.0. AO A T A A A Q 9-® » 4 A <& ^ A7 —© & SIIMOP —Oo DINOP —<]o -0 -0 -0 •0 -0 •0 -0 -0 -<3 -0 -0 -0 -07 -07 Figure 6. Data and Status Input, Logic Diagram Figure7 Data and Command Output, Logic Diagram via the DAGOP or CMGOP line. The duration of these pulses is shortened, from about 800 ns to about 400 ns, by the use of a one-shot multivibrator in each line; this ensures that the control pulses are removed before the data disappears from the DALOP lines. The lines DASYO and CDSYO return the respective synchronize signals to the processor. The data available and command control lines are OR ed so that there cannot be any data or commands on the DALOP lines except when one of the two control lines is active. Thus this OR gate eliminates extraneous noise on the data transmission cable, an advantage when several interface cables are run in close proximity. 3.5 Interrupt Control The logic circuit for the Interdata interrupt control ler is shown in Figure 8. The detailed operation of this circuit is described in the Interdata"Systems Interface Manual" and only a brief explanation will be given here. The enable/disable switch is set in the enable position to activate the interrupt controller. When a byte of data is available at the digital voltmeter a data ready pulse is gener ated which causes the queue flip-flop to be direct set. The output from the queue flop-flop sends an attention signal (ATNO) to the processor via G12. The processor responds by returning 26. TACKO RACKO +5v4 | V 47Cvn| P G13 +5v d*70ii 0-ATSYNO ATIMO G12 * * +5v< +5v 470pF Q J T> K — >470.n. Gil 0 & C2 QUEUE FLIP-FLOP I 470pF .+5v DATA -0 READY SCLROA <j ATSYN1 0 Figure 8. Interrupt Control, Logic Diagram 27. a receive acknowledge signal (RACKO) to the controller. The gate G8 disables G13 and prevents the transmit acknowledge signal (TACKO) from being sent to the next device. Gates G9 and G10 generate attention synchronize (ATSYNO) which sends a synchronize signal (SYNO) to the processor as well as causing the mass spectrometer address to appear on the inputs of GO through G7 (see Figure 5). This address is then enabled onto the data request lines by the ATSYNI output from Gil. On receiving the SYNO signal, the processor raises RACKO, causing the output of Gil to drop and the queue flip-flop to reset. If the processor is busy servicing another device interrupt, when a data ready pulse is generated, RACKO is low and the mass spectrometer interrupt is disabled. However, this latter interrupt is stored in the queue flip-flop and is serviced immediately after the previous interrupt has been serviced. A push-button switch situated at the processor (the initialize switch) is connected via the system clear line (SCLRO) to each interrupt controller such that all queue flip-flops can be direct reset simultaneously. The interrupt acknowledge control line (ACKO) shown in Figure 3 is divided up into a series of short lines to form the daisy-chain priority system. Clearly the acknowledge signal must pass through every interface equipped with an interrupt controller, and the device situated closest (electrically) to the processor, along the daisy-chain, has highest priority. 28. 3.6 Read/Write Sequencing The logic circuit shown in Figure 9 is used to control the sequencing of the read and write operations called for by the computer program. Data and command bytes from the processor arrive along the DALOP lines and are fed to the output command (0C) memory and display logic (Figure 15) via the four DALIP lines. Three of the outputs from the 0C memory are used to drive a l-of-8 decoder which provides the sequencing signals for both read and write operations. The remaining output is used to provide a read or write enable signal to a series of AND gates. Both the data bytes received from the digital voltmeter and the data bytes to be written upon the display have to be sequenced in a particular order. This is accomplished partly by the software and partly by the sequencing logic. The external connections to the read-write sequencing logic (interface section 2) are shown in Figure 10. 3.6.1 Read Operation Binary-coded-decimal (BCD) information is available on four lines from each decade of the digital voltmeter. This data, together with the overrange digit, is gated onto the lines DIN0P4 through DIN0P7; the lines DINOP0 through DIN0P3 being unused. Additional information pertaining to shunt number, scan direction and display function switch position, are also gated onto the DINOP lines. p L>n FIGURE 9. COMPUTER/MASS SPECTROMETER INTERFACE READ/WRITE LOGIC 4DAL0P-5 : 6 : 7 : DAGOP: CMGor: 4 5 6 7 <D.U. DIGITAL VOLTMETER 17 LINES INTERFACE SECTION-2 READ/WRITE SEQUENCING LOGIC SHUNT NUMBER SCAN DIRECTION 1 FUNCTION 8 LINES SELECT i SWITCH 4 LINES 14 LINES DISPLAY AUTO RANGED TO SHUNT 'SELECT SWITCH STOP. ,GANGED TO SCAN UP /DIRECTION SWITCH SCAN STOP UP DOWN SEP. SEP. ~ REJ. ODOWN II it II >*<REJ, MAN, SCAN RATE CONTROL 2 LINES A SCAN V DRIVE FIGURE 10. EXTERNAL CONNECTIONS TO READ/WRITE SEQUENCING LOGIC Figure 11 shows a typical sequence of instructions necessary to read the digital voltmeter, shunt number, scan direction, and display function switch position. It is assumed here that the digital voltmeter has generated an interrupt and the processor is now ready to service the mass spectrometer. An "output command (0C)" is sent from the processor, arrives at the flC memory (Figure 9), and is promptly stored on" receipt of a command strobe pulse (CMGOP). This first 0C contains the coded information requesting that the overrange digit of the digital voltmeter be placed on the DINOP lines. The next instruction is "read data (RD)" which causes the data on the DINOP lines to be read into a specified location in core memory. The "add halfword register (AHR)", "compare logical halfword immediate (CLHI)", and "branch on low (BL)", instructions cause the 0C index register to be repetitively incremented by one until all available information has been read, one byte at a time, into separate locations in core memory. 3.6.2 Write Operation The sequence of instructions necessary to write (out put) data from core memory (Figure 12) resemble those used to read (input) data from the mass spectrometer (Figure 11) . The "output command (0C)" sequences the output bytes in a similar manner to that described in Section 3.6.1, however bit 4 of the OOOOR 0002R 0004R 0006R 0008R OOIOR 0012R 001 6R 001AR 001ER 0022R 0024R 0026R 0809 OAOB OCOD OEOF 0755 C820 0001 C830 OOOD DE35 OOOOR DB35 0008R 0A52 C550 0008 4280 001AR RCMD DC DC DC DC RDATA DS XHR LHI LHI RSTART OC RD AHR X'0809' X'OAOB' X'OCOD' X'OEOF' 8 5, 5 2, H' 1 ' 3>RCMD<5) 3. » RDATAC 5 > 5* 2 CLHI 5*8 BL RSTART DEFINE CONSTANTS DEFINE STORAGE ZERO REG#5 LOAD 1 INTO REG#2 LOAD 13 INTO REG# 3 OUTPUT COMMANDC READ) READ ONE BYTE DATA INCREMENT REG#5 BY 1 CONTENTS REG#5<8? YES; BRANCH TO RSTART FIGURE 11. READ SEQUENCING INSTRUCTIONS OOOOR 0002R 0004R 0006R OOOCR OOOER 0012R 001 6R 001AR 001 ER 0020R 0024R 0001 0203 0400 0755 C820 0001 C830 OOOD DE3 5 OOOOR DA35 0006R 0A52 C550 0005 4280 001 6R WCMD DC DC DC WDATA DS XHR LHI LHI WSTART OC WD AHR CLHI BL X'OOOl • X '0203 ' X '0400 * 6 5> 5 2#H'1 ' 3*H"1 3' 3*WCMD(5) 3>WDATAC5) 5* 2 5> 5 WSTART DEFINE CONSTANTS DEFINE STORAGE ZERO REG#5 LOAD 1 INTO REG#2 LOAD 13 INTO REG#3 OUTPUT COMMANDCWRITE) WRITE ONE BYTE DATA INCREMENT REG#5 BY 1 CONTENTS REG#5<5? YES; BRANCH TO WSTART FIGURE 12. WRITE SEQUENCING INSTRUCTIONS 0C is now a zero causing the write line to be made active. When a "write data (WD)" instruction is executed, a byte of data is fetched from a specified location in core memory and placed on the DAL's. A pulse on the DAGOP control line strobes this data byte into one decade of the display desig nated by the previous 0C. The location of the decimal point on the display is controlled by a separate memory and decoder, and can be up dated when required by a suitably coded 0C and WD instruction. Similarly the mass spectrometer magnetic scan rate can be varied using another memory. Each output of this latter memory is connected to a simple transistor switching circuit which controls the frequency of a unijunction transistor oscillator. This, in turn, determines the magnetic scan rate via a stepping motor and potentiometer. The complete scan drive circuit, designed by R. D. Russell, is given in Appendix II. In the present system there are four possible scan speeds which are considered ample for the programmed control envisaged A programming guide for the mass spectrometer inter face is given in Appendix IV. 3.7 The Analog-To-Digital Converter A Model 344-2 digital voltmeter (DVM) manufactured by the Data Technology Corporation is used as an analog-to-digital (A/D) converter in the data acquisition system. This 34. instrument, utilizes the dual slope integration technique for A/D conversion. Figure 13 shows a block diagram of the DVM and Figure 14 illustrates some typical dual slope waveforms. The analog voltage output from the mass spectrometer measuring system is applied to the input amplifier of the DVM. A pulse from the reset oscillator (frequency: 5Hz) initiates a 10,000 count such that the input signal is integrated for a period of 50ms. This integration time is controlled by a200KHz oscillator (clock). The integrating capacitor (Cl) is dis charged until the 10,000 count is completed (full scale), leav ing a voltage on Cl which is proportional to the input signal. Upon reaching full scale the input current (Ijjyj) is switched off and a constant current source (I^p) is switched to Cl. The capacitor is then charged at a constant rate while the counter continues to run. When the voltage across the capac itor reaches the start voltage (15 volts), a zero detect (ZD) pulse is generated which resets the ZD flip-flop and disables the clock. The number of counts (N), accumulated by the counter, is proportional to the input voltage. The output of the ZD flip-flop triggers a one-shot pulse (data ready) which enables the quatch-latch memories to accept the new BCD numbers into storage. The BCD outputs from these memories are used to provide the digital information for the computer interface. +200V VlN UN ^POS NECi) ( "1000" ~^ (" "100" ~~^) ^ "10" ^ "1" """^ 5 r JK F/F JK F/F FULL SCALE START-FULL SCALE FLIP-FLOP ZERO DETECT(ZDI QUAD LATCH 8 4 2 1 DECODER DRIVER QUAD LATCH 4 2 1 COUNTER 200 kHr OSC RESET OSC ONE SHOT EXT START DATA READY FIGURE 13. ANALOG-TO-DIGITAL CONVERTER BLOCK DIAGRAM (Reproduced, by permission, from Data Technology Corp., Manual#18915-10) RESET to tfg tzo 1 1 { START FULL SCALE I FLIP-FLOP ^ 50ms ^ c 200KHz CLOCK ZERO DETECT DATA READY FIGURE 14. DUAL SLOPE WAVEFORMS-FOR DIGITAL VOLTMETER An important feature of the dual slope technique is that the accuracy of the A/D conversion is not dependent on the drift of the 200KHz oscillator. Since the 10,000 count remains constant, the integrating time varies in accordance with any oscillator drift such that the total count, N, remains constant for a given input voltage. The Model 344-2 DVM has a full scale voltage range of 0-1.0000 volts with a 40% overrange capability. The manu facturers quoted accuracy is ±(.011 reading + .0001) volts. 3.8 The Numerical Display The numerical output display for the mass spectrom eter interface consists of 5 Datecon DS-103 display modules. Each module consists of a quad-latch memory which stores the input data (in BCD format) on application of a strobe pulse. The output from the memory is fed to a BCD-to-Decimal decoder which drives a cold-cathode decade display tube. Each display tube incorporates a decimal point which can be controlled by a separate quad-latch memory and decoder (Figure 9). Power requirements, in addition to the +5 volts logic supply, are a +200 volts supply for the display tube anodes. Figure 15 shows the complete display logic circuit. 3.9 Construction of the Interface The interface was constructed in two sections to try 38. O-1 O-2 O-3 t>-4 o-DAL IP 2 8 MEMORY 1 2 4 8 MEMORY BCD/DEC DECODER DRIVER COLD-CATHODE A+200V DECADE DISPLAY TUBE 0 MEMORY BCD/DEC DECODER DRIVER MEMORY BCD/DEC DECODER DRIVER BCD/DEC DECODER DRIVER MEMORY A A A A A DECIMAL POINTS FIGURE 15. THE NUMERICAL DISPLAY 39. and minimize the number of communication lines between the computer and mass spectrometer. Section 1 consists of the address logic, the data and status input logic, the data and command output logic, and the interrupt circuitry built on two double-sided printed circuit boards which slide into a rack situated inside the computer cabinet. Section 2 contains the read/write logic circuitry illustrated in Figure 9; the prototype unit constructed by the writer was built on 6 single-sided printed circuit boards which could communicate with each other through 22 pin connectors. The 6 printed circuit boards, numerical display, digital voltmeter, and power supply are mounted on an aluminum chassis which fits into the mass spectrometer console. An aluminum front panel incorporates the display and voltmeter bezels and all the necessary switches. The integrated circuit (i.e.) logic packages (chips) used in the construction of the interface are mostly Motorola DTL plastic types with the exception of the complex functions (decoders and quad latch memories) which are of the TTL family (see Table III). Each 14 or 16 pin plastic package contains from 1 to 6 logic elements depending upon the particular logic function desired. All the i.e. packages employed in the mass spectrometer interface were designed for operation from a 5±1/2 volts regulated power supply (±5%). The power supply (PS-200) in section 2 supplies the 200 volts for the cold-cathode decade display tubes and a stabilized 5 volts for the read/write logic circuitry. The i.e. packages in section 1 utilize the Interdata Processor 5 volt TABLE III MOTOROLA NUMBER MC834P MC845P MC849P MC855P MC858P MC1803P MC1806P MC1809P MC4038P MC7442P MC7475P INTEGRATED CIRCUIT PACKAGES USED IN CONSTRUCTION OF INTERFACE LOGIC CIRCUITRY FUNCTION HEX INVERTER CLOCKED FLIP-FLOP QUAD-2 INPUT NAND J-K FLIP-FLOP QUAD-2 INPUT POWER 8-INPUT NAND QUAD-2 INPUT AND QUAD-2 INPUT OR l-OF-8 DECODER BCD/DEC DECODER QUAD LATCH MEMORY FAMILY DTL DTL DTL DTL DTL DTL DTL DTL TTL TTL TTL OUTPUT LOAD ING FACTOR /OUTPUT 8 12 7 11 27 7 8 7 11 11? 10 PROPAGA- TOTAL POW- COST QUANTITY TION DELAY ER DISSIPA- /PKG USED IN ns TYPICAL TION mW TYP/PKG $ INTERFACE 30 40 25 40 30 25 35 30 45 45? 30 66 60 66 140 130 16.5 72 115 240 105? 160 2.00 1.62 1.65 2.10 2.70 1.55 1.90 1.90 6.65 6.75 4.50 8 1 17 1 17 2 6 1 1 1 3 o supply which serves to eliminate one line between sections 1 and 2. An unshielded cable, approximately 43 feet in length, is used to transfer information on 13 lines between section 1 and 2; this cable should be kept as short as possible. No noise or crosstalk problems were experienced provided all the lines were made false-active and power gates, with a suitable pull-up resistor, were used at the transmission end of each line. In addition great care was taken to avoid ground loops which can easily arise during the construction of complex electronic equipment. After the prototype interface had proved itself reliable over several months of testing, a double-sided, one piece printed circuit board was designed (by E. J. Bellis) for the section 2 logic circuitry. Using this new printed circuit board, two more complete interfaces were constructed and installed in the remaining two 30 cm radius mass spectrometers at the University of B.C., isotope geophysics laboratory. All three interfaces are now operating and are being used for isotope analyses, and it is intended that the one Interdata computer will service all three mass spectrometers on a time sharing basis. CHAPTER 4 4.1 On-Line Filtering of Data The simplest mode of operation of the data acquisition system is the on-line filtering of data from the digital volt meter and the display of the filtered data on the numerical readout. Previous computer off-line programs designed to pro cess mass spectral data have always incorporated some form of digital filtering to reduce higher frequency noise. It therefore seemed reasonable to design a digital filter program for the Interdata computer which would display the filtered data points at the mass spectrometer console immediately, as well as storing a smoothed version of the mass spectrum in a memory buffer, from which it could be transferred to magnetic tape via the Selector Channel. A suitable low-pass digital filter had been designed by R. D. Russell and J. Blenkinsop for an IBM 360/67 and this program was rewritten (by J. Blenkinsop) in the Interdata programming language. A flow-diagram of the filter program is shown in Figure 16. The digital voltmeter produces 5 data points/second which corresponds to a Nyquist frequency of 2.5Hz. The fund amental sampling theorem requires that in order to completely recover the original signal, the sampling frequency must be 43. READ 5 CHARACTERS FROM DIGITAL VOLTMETER \ / CONVERT DECIMAL TO BINARY 7 POINT AVERAGE SKIP EVERY 2ND POINT 3 POINT AVERAGE 5 POINT AVERAGE > SKIP 2 OUT OF 3 POINTS STORE FILTERED POINT CONVERT BIN . TO DEC. WRITE 5 CHARACTERS ON DISPLAY FIGURE 16. MASS SPECTROMETER FILTER-DISPLAY PROGRAM at least twice the highest occurring frequency in the sampled signal. Physical signals, however, do not have a finite frequency content. The part of the signal spectrum lying above the Nyquist frequency will be reflected and superimposed (folded back) onto lower frequencies, and the original signal can only be recovered approximately. The best one can do is to choose the Nyquist (folding) frequency (and therefore the sampling rate) high enough to include all frequencies lying in the passband of the ion current measuring system. However, fre quencies above 2.5Hz do not contribute significantly to the mass spectral records encountered and a sampling rate of 5 point second should therefore be quite adequate. When selecting a filter for mass spectral data it is important that the width of the total averaging function (the data window) is less than or equal to the mininum width of the peak tops. For the filter used, the width of the data window is 3.8 seconds which necessitates dwelling on a peak top for a period of at least 3.8 seconds. Digital filtering is accomplished in the Processor by adding together points to form a moving average by sevens and applying every other averaged point to a three point moving average, followed by a five point moving average. One out of every three points from the five point average is stored in a memory buffer, converted from binary to decimal notation, and then written on the display. The 7-point, 3-point, and 5-point averages all have tapered endpoints, which is to say that the end-points have weighting coefficients of one half (in our case) 45. This has the effect of reducing the amplitude of side lobes on the filter response (Figure 17) . It should be evident from the above description that there is one filtered point available, at the output of the filter, for every six raw input data points, hence the display is updated once every 1.2 seconds. The Nyquist frequency is lowered to 2.5/6 Hz after filtering, but since there is little signal or noise still present above 0.1Hz, it is clear that no information is being discarded in this way. a FREQUENCY-CPS FIGURE 17. FREQUENCY RESPONSE OF DIGITAL FILTER CHAPTER 5 SYSTEM PERFORMANCE 5.1 Introduction i In order to test the accuracy and reliability of the computer interface two strontium isotope analyses were performed using both an interlaboratory standard and a rock sample of unknown composition. The mass spectrometer which has been interfaced to the computer is principally used for the rubidium-strontium age dating of rock samples. Natural rubidium has two isotopes, Rb85 which is stable, and Rb87 which is radioactive and decays to Sr87 with a half-life of approximately 5x1010 years. There are four stable isotopes of strontium, namely Sr81*, Sr86, Sr87 and Sr88 , and of these, in a closed chemical system, only the abundance of Sr87 increases with time. If the initial amount of Sr87 in the sample can be determined, and the present abundances of Sr87 and Rb87 in the sample are measured, the age of the sample can be determined. 5 .2 Preparation of Strontium Samples The method devised by B. D. Ryan (1971) for the chemical spearation of strontium from a rock sample is used. The rock sample is crushed to a fine powder and 48. approximately 0.25 gms is \tfeighed out into a teflon beaker. The sample is completely dissolved in H2S0it and HF, heated, and evaporated to dryness. The residue is dissolved by warming with HC1 and allowed to cool. The sample is centri fuged and the solution transferred to the top of a cation exchange column, of length 21 cms, containing Dowex 50W-X8 Resin (200-400 Mesh). The column is eluted with 6NHC1 and the eluate collected in a measuring cylinder. The first 25 ml of eluate are discarded and the next 40 ml collected. The cut of 40 ml is evaporated dry and taken up in 2 ml 2NHC1. Mean while, the columns are back aspirated with 2NHC1. The sample (2 ml) is added to the column and eluted with2NHCl. The first 90 ml of eluate are discarded and the next 40 ml collected. This cut of 40 ml is evaporated to dryness and the residue dissolved in 3 drops 2NHC1. 5.3 Analysis of Eimer and Amend SrC03 Interlaboratory Standard The Eimer and Amend SrC03 was chosen because it has been analysed at a large number of different laboratories, in cluding The University of British Columbia,where it has been analysed a number of times previously using the same mass spectrometer. For the purposes of interlaboratory comparison the Sr87/Sr86 ratio is taken as 0.70800. The SrC03 is dissolved in 2NHC1 to produce a solution containing approximately 400ug strontium per ml of 2NHC1. Two drops of the solution are deposited on each of the outgassed rhenium side-filaments* and evaporated to dryness by passing a current of about one Amp through each filament while they are exposed to the atmosphere. The two side-filaments and a rhenium centre-filament are mounted in a stainless steel block which, in turn, is positioned inside the mass spectrometer. The whole system is evacuated to a low pressure (< 2xl0"7 mm Hg). The centre-filament is heated by passing a current of about 4.0 Amps through it. Collision with the hot centre-filament produces efficient ionization of the sample that is being gently evaporated from the relatively cool side-filaments. The positive ions produced are accelerated by a potential of 5000 volts and can be focussed into a Faraday cup using a variable field strength electromagnet. The charges that collect in the cup, constitute the ion current which is measured, the magnitude of this current is directly proportional to the abundance of the particular isotope ion beam focussed into the Faraday cup. The centre-filament current is slowly increased and a search is made for a Rb85 contamination peak. If any rubidium is present, it is burnt off the side-filaments at a filament temperature just below that required to produce an appreciable strontium spectrum. When the height of the Rb85 *The triple-filament technique of solid-source mass spectrometry was used for all the analyses. 50. peak is negligible, which implies that the height of the Rb87 peak is also negligible (since the ratio Rb85/Rb87 = 2.593 is constant), the centre-filament current is increased to a value such that a Sr88 peak height of about 1 Volt is obtained with the output attenuator set on shunt 3. A check is made for rubidium contamination at this new centre-filament temper ature and if a Rb85 peak is still detectable it is burnt away before the strontium spectrum is scanned. The strontium 86, 87 and 88 peaks are scanned in sequence about 12 times by varying the magnet field intensity using a peak-hopping technique. The peak heights and baselines are read from the filter display and are written by hand on the chart recorder which provides a visual record of the spec-trum. In order to calculate the Sr87/Sr86 ratio for each scan it is necessary to correct the measured ratio for the growth or decay of the peak heights, then this ratio is corrected for mass discrimination at the ion source. Fortunately the peak heights grow or decay almost linearly and hence a correction is easily applied. Discrepancies in the isotope abundances due to mass discrimination (fractionation) can be simply correc ted for in the case of strontium since the ratio Sr87/Sr86 is constant (=0.1194). The Sr86/Sr88 ratio is calculated for each scan and the discrimination per unit mass determined. An average value for the corrected (normalised) Sr87/Sr86 ratio over all the scans is calculated, together with a value for the standard deviation of the mean. The result, given in Table IV, indicates about a twofold increase in precision over previous values obtained by J. Blenkinsop, using the same mass spectrometer, without the computer interface. 5.4 Analysis of Strontium in a Rock Sample As a further test of the performance of the computer interface two analyses were performed on a rock sample. The sample .selected was a light grey argillite (slightly metamor phosed claystone) from the Creston formation outcropping in S.E. British Columbia. This formation is part of the Purcell Series which is stratigraphically equivalent to the Belt Series which outcrops in western Montana and northern Idaho. Smith and Barnes (1966) refer to these two series as the Belt-Purcell Supergroup. The Supergroup, which crops out over an area of more than 50,000 square miles, consists largely of metamorphosed sediments which have not undergone intense deformation. Thick nesses of greater than 40,000 feet have been attained. The sediments contain no useful datable fossils and therefore pro vide excellent material for isotope dating studies of Precambrian rocks (1400 - 900 m.y. [million years]). Obradovich and Peterman (1968) have dated rocks of the Belt Series using Rb-Sr and K-Ar techniques. Their deter minations yield ages ranging from about 900 m.y. to around 1300 m.y. Rb-Sr isotope measurements performed by Ryan and Blenkinsop (1971) on the Hellroaring Creek Stock in the Purcell Mountains indicate an approximate age of 1260 m.y. This stock, which intrudes the lowest known formation of the Purcell Series (the Aldridge Formation) , is the oldest recognized in British Columbia. The sample of argillite was collected by Dr. W. C. Barnes of the Geology Department, University of B.C., and was made available to the writer by Mr. B. D. Ryan. X-ray fluorescence measurements, performed by C. Croucher, have shown that this sample contains 80 ± 2% ppm (parts per million) rubidium and 248 ± 2% ppm strontium. The chemical spearation described in Section 5.2 was carried out twice to provide two strontium samples (A and B) from the one specimen of argillite. This duplicate separa tion was used to provide a check on the reproducibility of the chemistry since, ideally, samples A and B should yield identical isotope abundance ratios. Mass spectrometric analyses of samples A and B were performed in a similar manner to that described in Section 5.3 for the Eimer and Amend standard. The much lower concentration of strontium in the argillite samples necessitates much greater care in the filament preparation. Consequently it is important to ensure that all the sample is deposited on the side-filaments, which normally entails pipetting three or more drops of solution onto each side-filament. In order to perform this task, and still maintain all the sample within the middle one-third upper-surface of each side-filament ribbon, the following technique has been found satisfactory. Place one drop of solution in the centre of each side-filament, slowly evaporate the drop to dryness by passing a current of about one Amp through each filament, cool completely, add another drop to the middle of each ribbon, evaporate to dryness, and so on until all the sample solution is shared equally between the side-filament. The current through each filament is then increased until the filaments glow with a dull red color and are then left to "cook" for about one minute. This latter procedure helps produce a more stable ion beam. Table V gives the results of the two mass spectrom eter runs for the samples A and B and the concentrations of the four isotopes of strontium in the argillite sample. These concentrations were calculated using the X-ray fluorescence data and the ratio Sr87/Sr86 derived from the mass spectrometer runs. The percentage deviation from the mean of the ratio SR87/Sr86 at a 95% confidence level is about 0.05% which is as good or better than previous analyses performed on rock samples using the same chemical separation procedure and mass spectrom eter without the digital filtering of data. OH . TABLE IV RESULTS OF ANALYSES OF EIMER S AMEND INTERLABORATORY STANDARD SrCQ3 LABORATORY UBC (RUSSELL ET AL, 1971) UBC (RYAN § BLENKINSOP, 1971) USGS (STACEY ET AL, 1971) USGS (STACEY ET AL, 1971) MIT (SPOONER $ FAIRBAIRN, 1970) YALE (DASCH, 1969) U of T (PURCY § YORK, 1968) RATIO (Sr87/Sr86)n COMMENTS 0.7080±0.0002* 0.7082±0.0004 0.7080±0.0002 0.7079±0.0006 0.7082±0.0008 0.7075±0.0012 0.7080±0.0012 ON-LINE FILTER, HAND CALCULATED (THIS THESIS) HAND CALCULATED DIGITALLY RECORD ED $ COMPUTED HAND CALCULATED HAND CALCULATED HAND CALCULATED HAND CALCULATED ACCEPTED VALUE 0.70800 TABLE V RESULTS OF ANALYSES OF GREY ARGILLITE SAMPLE A B MEAN OF A 5 B CONSTANT RATIOS: Sr8VSr86 - 0.0568 RATIO (Sr87/Sr86)n 0.7419±0.0003* 0.7420±0.0005 0.7419±0.0004 Sr86/Sr88 = 0.1194 ISOTOPE CONCENTRATIONS (parts per million) ±2% RK 1 QrS1' 9T86 <!r87 ^r88 (TOTAL) (TOTAL-' br &r br bT 80 248 1.3 23.9 17.9 204.8 * - ALL UNCERTAINTIES QUOTED ARE TWO STANDARD DEVIATIONS (95% CONFIDENCE LEVEL) CHAPTER 6 6.1 Conclusions The analyses described in Chapter 5 were not expected to show any appreciable increase in precision over previous analyses, performed on the same mass spectrometer, without the computer interface, but only to test the design and operating reliability of the system. With this end in mind one can only be very satisfied with the results so far obtained. The full potential of this on-line data acquisition system will only be realized when the processor is programmed to determine peak heights and baselines, control scan rates and apply the various analytical corrections. J. Blenkinsop has completed a prototype program to provide on-line processing of data from the mass spectrometer and his initial results, from several trial runs, indicate about a threefold increase in precision over previous off-line analyses. No further improvements in the hardware are envisaged in the immediate future, although possible areas of investiga tion may lie with the computer control of source conditions and the direct reading of the magnetic field intensity for each peak. There remains great scope for improvements in the software and it is in this field that much effort is being channelled. BIBLIOGRAPHY Albee, A.L., Burnett, D.S., Chodos, A.A., Eugster, O.J., Hun^eke,' J.C. , Papanastassiou, D.A., Podosek, F.A., Russ, G.P., Sanz, H.G., Tera, F., and Wasserburg, G.J. (1970). Ages, irradiation history, and chemical composition of lunar rocks from the Sea of Tran quility. Science 167, 463. Catanzaro, E.J. (1967). Triple filament method for solid-sample lead isotope analysis. J. Geophys. Res., 72, 1325. Compston, W., and Oversby, V.M. (1969). Lead isotopic analysis using a double spike. J. Geophys. Res., 74_, 4338. Dasch, E.J. (1969). Sr isotopes in weathering profiles, deep-sea sediments , and sedimentary rocks . Geochim.. Cosmochim., 3_3, 1521. Data Technology Corp. Instruction Manual 344 4 Digit/Systems Meter (part No. 18915-10). Interdata, Inc. (1969). Reference Manual (publication No. 29-004R02). Systems Interface Manual (publication No. 29-003R02). Motorola Semiconductor Products Inc. (1968). The Integrated Circuit Data Book. Obradovich, J.D., and Peterman, Z.E. (1968). Geochronology of the Belt Series, Montana. Can. J. Earth Sci., 5_, #3, part 2, 737. Papanastassiou, D.A., and Wasserburg, G.J. (1969). Initial Sr isotopic abundances and the resolution of small time differences in the formation of planetary objects. Earth § Planet. Sci. Letters, 5_, 361. Purdy, J.W., and York, D. (1968). Rb-Sr whole rock and K-Ar mineral ages of rocks from the Superior Province near Kirkland Lake, northeastern Ontario, Canada. Can. J. Earth Sci., 5_, 699. Ryan, B.D., and Blenkinsop J. (1971). Geology and geochronology of the Hellroaring Creek stock, B.C. Can. J. Earth Sci., 8, 85. 57. Russell, R.D., Blenkinsop, J., Meldrum, R.D., and Mitchell, D.L. (1971). On-line computer assisted mass spectrometry for geological research. Journal of the Physics of the Earth (in press). Smith, A.G., and Barnes, W.C. (1966). Correlations of and facies changes in the carbonaceous , calcareous, and dolomitic formations of the Pre-cambrian Belt-Purcell Supergroup. Bull. Geol. Soc. Amer.,77, 1399. Spooner, CM., and Fairbairn, H.W. (1970). Sr87/Sr86 initial ratios in pyroxene granulite terranes. J. Geophys. Res., 75, No. 32 , 6706. Stacey, J.S., Russell, R.D., and Kollar, F. (1965). Servo-am plifiers for ion current measurement in mass spectrom etry. J. Sci. Inst., 42, 390. Stacey, J.S., Wilson, E.E., Peterman, Z.E., and Terrazas, R. (1971). Digital recording of mass spectra in geologic studies, I. Can. J. Earth Sci., 8_, 371. Wasserburg, G.J., Papanastassiou, D.A., Nenow, E.V., and Bauman, CA. (1969). A programmable magnetic field mass spectrometer with on-line data processing. Rev. Sci. Inst. , 40 , 288. Weichert, D.H. (1965). Digital analysis of mass spectra (Ph.D. Thesis, University of British Columbia). Weichert, D.H., Russell, R.D., and Blenkinsop J. (1967). A method for digital recording for mass spectra. Can. J. Physics, 45, 2609. APPENDIX I. ION CURRENT AMPLIFIER $ OUTPUT ATTENUATOR CIRCUIT. ( Designed by R.D.Russell, June 1969. ) STEPPING MOTOR -28v L O nrrYTYn. APPENDIX II. SCAN DRIVE CIRCUIT (designed by R.D.Russel1 , June 1966) 60. APPENDIX III. SUMMARY OF INTERDATA PROGRAMMING INSTRUCTIONS* • CODE TYPE MNEMONIC INSTRUCTION 01 RR BALR Branch and Link 02 RR BTCR Branch on True Condition 03 RR BFCR Branch on False Condition 04 RR NHR AND Halfword 05 RR CLHR Compare Halfword 06 RR OHR OR Halfword 07 RR XHR Exclusive OR Halfword 08 RR LHR Load Halfword OA RR AHR Add Halfword OB RR SHR Subtract Halfword OC RR MHR Multiply Halfword OD RR DHR Divide Halfword OE RR ACHR Add with Carry Halfword OF RR SCHR Subtract with Carry Halfword 28 RR LER Floating-Point Load 29 RR CER Floating-Point Compare 2A ,RR AER Floating-Point Add 2B RR SER Floating-Point Subtract 2C RR MER Floating-Point Multiply 2D RR DER Floating-Point Divide 40 RX STH Store Halfword 41 RX BAL Branch and Link 42 RX BTC Branch on True Condition 43 RX BFC Branch on False Condition 44 RX NH AND Halfword 45 RX CLH Compare Logical Halfword 46 RX OH OR Halfword 47 RX XH Exclusive OR Halfword 48 RX LH Load Halfword 4A RX AH Add Halfword 4B RX SH Subtract Halfword 4C RX MH Multiply Halfword 4D RX DH ' Divide Halfword 4E RX ACH Add with Carry Halfword 4F RX SCH Subtract with Carry Halfword 60 RX STE Floating-Point Store OP CODE - TYPE MNEMONIC INSTRUCTION 68 RX LE Floating-Point Load 69 RX CE Floating-Point Compare 6A RX AE Floating-Point Add 6B RX SE Floating-Point Subtract 6C RX ME Floating-Point Multiply 6D RX DE Floating-Point Divide 90 RR UNCH Unchain 92 RR STBR Store Byte 93 RR LBR Load Byte 96 RR WBR Write Block 97 RR RBR Read Block 9A RR WDR Write Data 9B RR RDR Read Data 9D RR SSR Sense Status 9E RR OCR Output Command 9F RR AIR Acknowledge Interrupt CO RS BXH . Branch on Index High Cl RS BXLE Branch on Index Low or Equal C2 RX LPSW Load Program Status Word C4 RS NHI AND Halfword Immediate C5 RS CLHI Compare Logical Halfword Immediate C6 RS OHI OR Halfword Immediate C7 RS XHI Exclusive OR Halfword Immediate C8 RS LHI Load Halfword Immediate CA RS AHI Add Halfword Immediate CB RS SHI Subtract Halfword Immediate CC RS SRHL Shift Right Logical CD RS SLHL Shift Left Logical CE RS SRHA Shift Right Arithmetic CF RS SLHA Shift Left Arithmetic DO RX STM Store Multiple Dl RX LM Load Multiple D2 RX STB ' Store Byte D3 RX LB Load Byte D5 RX AL Autoload D6 RX WB Write Block D7 RX RB Read Block DA RX WD Write Data DB RX RD Read Data DD RX SS Sense Status DE RX OC Output Command DF RX Al Acknowledge Interrupt * Reproduced, by permission, from Interdata Reference Manual #29-004R02. 62. APPENDIX IV MASS SPECTROMETER INTERFACE PROGRAMMING GUIDE ADDRESSES HEX D - M.S.2 HEX E - M.S.I HEX F - M.S.3 STATUS AND COMMAND BYTE DATA BIT NUMBER 0 1 2 3 4 5 6 7 STATUS BYTE DU COMMAND BYTE ( .-REAl )-WRIr 3 rE si •QUENi :ING DU - The device unavailable power supply is off. READ - This command sets the data can be read from bit is set when the M.S. 5v output command memory so that the M.S. WRITE - This command sets the OC memory so that data can be written on the display. The scan rate is also controlled in the write mode. SEQUENCING Bits 5,6 and 7 are used for sequencing the data in both the read and write modes. READ SEQUENCING - COMMAND BYTE BIT NUMBER 4 5 6 7 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 •Read DVM overrange •Read DVM decades in sequence -Read shunt number -Read scan direction code -Read function select switch Data is read using the OC instruction followed by the read data (RD) instruction. WRITE SEQUENCING COMMAND BYTE BIT NUMBER 4 S 6 7 1 1 1 1 1 1 1 1 1 1 1 1 •Write on display decades in sequence -Write decimal point -Set scan rate -Not used Data is written using the OC instruction followed by the write data (WD) instruction. 64. DVM DECADE NUMBERING (Front view). OVER-RANGE 1 2 3 4 DISPLAY DECADE NUMBERING (Front view) 0 1 2 3 4 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0053430/manifest

Comment

Related Items