|
DataMuseum.dkPresents historical artifacts from the history of: RC4000/8000/9000 |
This is an automatic "excavation" of a thematic subset of
See our Wiki for more about RC4000/8000/9000 Excavated with: AutoArchaeologist - Free & Open Source Software. |
top - metrics - download
Length: 12288 (0x3000) Types: TextFile Names: »microwave«
└─⟦621cfb9a2⟧ Bits:30002817 RC8000 Dump tape fra HCØ. Detaljer om "HC8000" projekt. └─⟦0364f57e3⟧ └─⟦1ad7c9ab2⟧ »diverse« └─⟦this⟧
wave=set 200 wavekorr=set 200 wave=typeset machine.diablo proof.wavekorr *pl 264,6,258,0,0* *pn 0,0* *ld16* *lw 169* *sb!,6* *lg2* *ds#* *ps0,0* *fg21* A MICROCOMPUTER AIDED MICROWAVE SPECTROMETER *nl1* CONTROLLED BY A LARGER COMPUTER *nl2* N.W. LARSEN#h1 and J. OXENBØLL#h2*nl1* #h1Chemical Laboratory V, University of Copenhagen *nl1* the H.C. Ørsted Institute, Universitetsparken 5 *nl1* DK-2100 Copenhagen Ø (Denmark) *nl1* #h2Computer Department, the H.C. Ørsted Institute *nl3* ABSTRACT *np6* *ld12* A Hewlett-Packard 8460A microwave spectrometer has been connected to a microcomputer based on the Motorola MC6800 microprocessor. The microcomputer is programmed to control frequency and Stark voltage and to perform A/D conversion, storage and accumulation of intensities. *np6* The microcomputer is connected to a larger local computer, the RC4000, from which it may be loaded with a program, started and stopped, and from which data transfer between the storages of the two computers is controlled. This transfer may be performed, while the microcomputer is running a spectrum. *ld16* *nl2* INTRODUCTION *np6* Some of the reasons for digitalizing a spectrometer are common to most forms of spectroscopy: The possibility of improving the S/N ratio by accumulation of spectra, and the easy subsequent treatment of the spectra with smoothing procedures and line measuring procedures etc. For a Stark modulated microwave spectrometer it is desirable to control, not only the frequency, but also the Stark voltages from the computer, with the aim of eliminating some of the disadvantages associated with the use of Stark modulation. The HP 8460 microwave spectrometer is contructed in a way that makes it easy to computerize and originally an option equipped with a minicomputer was available. Instead of choosing the traditional set-up with a minicomputer connected to the spectrometer, we found that a better solution, which was moreover given as a service from the computer department, was to connect a microcomputer directly to the spectrometer and to let the microcomputer be controlled from a larger general purpose computer (the RC4000). In this way we benefit from the extensive facilities of the larger computer, although the experiment can run on the microcomputer without interference from other users of the large computer. *nl2* *ps0* THE MICROCOMPUTER *nl1* _«bs»G_«bs»e_«bs»n_«bs»e_«bs»r_«bs»e_«bs»l_«bs» _«bs»d_«bs»e_«bs»s_«bs»c_«bs»r_«bs»i_«bs»p_«bs»t_«bs»i_«bs»o_«bs»n *np6* _«bs»H_«bs»a_«bs»r_«bs»d_«bs»w_«bs»a_«bs»r_«bs»e_«bs». !The microcomputer, called HC6800, was designed in 1975 at the H.C. Ørsted Institute, with the explicit purpose of controlling laboratory equipment (ref.1). It consists of four different types of modules, each placed on a single printed circuit board: 1) a cpu module, based on the M6800 microprocessor from Motorola, with interrupt logic, decoding of address lines for selecting memory and input/output, and with a simple multiprogrammed monitor program placed in a 4 kbyte EPROM and a 2 kbyte RAM for descriptions and variables. 2) a memory card with 32 kbyte of RAM memory, used for application programs and data. 3) a system input/output card, which controls three asyncronous full-duplex serial lines for connection to the RC4000 via the local area network and a 24 bit system timer. 4) a number of user input/output cards, designed for the specific experiment, but using standard Motorola microprocessor peripherals, in a way which is common for all the HC6800 microcomputers. *np6* Only two different kinds of external devices are connected to the microcomputer, namely the RC4000 computer and the equipment to be controlled. This simple configuration makes the hardware and software construction simple and cheap. All other external devices, such as terminals, discs, magnetic tape stations, printers, etc, are connected to the RC4000 computer. *np6* _«bs»S_«bs»o_«bs»f_«bs»t_«bs»w_«bs»a_«bs»r_«bs»e_«bs». ! The monitor in the HC6800 takes care of upstart, communication with the network, and handling of peripherals. The monitor is multiprogrammed and runs several tasks in parallel, eg. the user program and the communication process, and further provides a number of utility procedures used by most application programs. *np6* An Algol program, acting as an operating system for the microcomputer, runs in the RC4000, which has full control over the microcomputer. It starts and stops it, loads it with a user program, and extracts data from it when appropriate. *nl2* _«bs»I_«bs»n_«bs»t_«bs»e_«bs»r_«bs»f_«bs»a_«bs»c_«bs»e_«bs» f_«bs»o_«bs»r_«bs» _«bs»t_«bs»h_«bs»e_«bs» _«bs»m_«bs»i_«bs»c_«bs»r_«bs»o_«bs»w_«bs»a_«bs»v_«bs»e_«bs» _«bs»e_«bs»q_«bs»u_«bs»i_«bs»p_«bs»m_«bs»e_«bs»n_«bs»t *np6* The microcomputer running the HP8460 spectrometer is equipped with *fg140* three user i/o-cards (Fig.1). The first contains a dual 8 bit parallel interface adapter (pia), where one port outputs four 8 bit words into four 8 bit registers, storing 8 bcd digits for the frequency. At the other port two bit are used to control the local/remote inputs on the HP8460 spectrometer for frequency and Stark voltage. The card also contains a programmable timer module (ptm), that controls the experiment by means of a pulse train in which each pulse has the effect of 1) initiating the transfer of the frequency from the four registers to the spectrometer, 2) initiating the A/D conversion (se below), 3) sending an interrupt *nl1* Fig. 1. Diagram showing in a scematic way the HC6800 microcomputer with the interface to the microwave spectrometer and to the RC4000 computer. *nl2* to the microprocessor. The second i/o-card contains a pia and four 8 bit registers, for storing the digital information for two 12 bit D/A converters, also placed on the card. The D/A-converters produce the control voltages for the ground to base voltage and the base to peak voltage respectively. The last i/o-card contains one pia and two 8 bit registers for storing the digital information from a 14 bit A/D-converter, placed on the same card. The A/D-converter is used to convert the intensity signal. *nl2* APPLICATIONS *nl1* _«bs»P_«bs»r_«bs»o_«bs»g_«bs»r_«bs»a_«bs»m_«bs»s_«bs» _«bs»a_«bs»n_«bs»d_«bs» _«bs»p_«bs»a_«bs»r_«bs»a_«bs»m_«bs»e_«bs»t_«bs»e_«bs»r_«bs»s *np6* Two programs are used in the experiment. The first is an Algol program that runs interactively in the RC4000 and controls communication with the microcomputer. It also performs the further treatment of the spectra, such as drawing, smoothing, differentation, line measurement, least squares fitting to Lorenzian and Gaussian line shapes, etc. The second program is written in assembler code for the microcomputer and controls the experiment in accordance with a given parameter set. *np6* The parameter set required by the last program includes the frequency limits (upper and lower frequency), the step length (!!1 kHz), the time between two steps (20 ms - 10 s), the number of spectra to be recorded and added in the microcomputer (!!100), and for each of these spectra the ground to base and the ground to peak values of the Stark voltage. All other parameters defining the experiment must be set manually on the spectrometer. *nl2* _«bs»A_«bs»v_«bs»e_«bs»r_«bs»a_«bs»g_«bs»i_«bs»n_«bs»g_«bs» _«bs»e_«bs»x_«bs»p_«bs»e_«bs»r_«bs»i_«bs»m_«bs»e_«bs»n_«bs»t *np6* When the parameters have been read by RC4000, they are, together with the microcomputer program, transferred to the microcomputer, which is then started. In the most common experiment the microcomputer is now able to work independently of the RC4000 until the spectra have all been recorded and their sum has been stored. Finally the spectrum is transfered to the RC4000 initiated by an order from the terminal. *np6* The Stark voltages are normally chosen according to one of several standard procedures. The most obvious choice is to let the base voltage be zero and let the peak voltage assume the same value for all spectra, with a usual time averaged spectrum as the result. *np6* We have found it very useful however to use Stark averaging instead, which implies that the peak voltage is changed from one scan to the next. The peak voltages may eather be distributed between zero and a chosen maximum value in an equidistant way, or alternatively in such a way that the difference between the squares of two settings is constant. If the first distribution is used whenever the spectral lines have first-order Stark effect and the second distribution is used when the lines have quadratic Stark effect, then each individual Stark component will be shifted uniformly from scan to scan and the result, with an appropriate number of scans and a sufficiently high value of the maximum peak voltage, will be an average spectrum corresponding to zero field, in which the only reminiscence of the Stark components is a shift of the base line. *np6* Stark averaging can also be used with a non zero base voltage, and in this case one ideally obtains an average spectrum corresponding to one particular non zero field, defined by the base voltage. This type of spectrum is very useful for determination of dipole moments (Fig. 2). *ps0* *fg150* *nl1* Fig. 2. Three spectra of the 5#l0#l,#l5!!6#l0#l,#l6 transition of methylene oxalate. (a) Is a usual Stark spectrum at 500 V peak voltage. The other two are Stark averaged spectra with 200 V base voltage (b), and 400 V base voltage (c). The quantum number M, is indicated on (c). *nl2* _«bs»L_«bs»o_«bs»n_«bs»g_«bs» _«bs»s_«bs»p_«bs»e_«bs»c_«bs»t_«bs»r_«bs»a *np6* Since the space available for the spectrum in the storage of the microcomputer is limited to 10000 points, each containing 24 bit, very long spectra must be treated by a different procedure: Only one scan can be recorded and the data must be transferred to the RC4000 during the run. To accomplish this, the storage of the microcomputer is organized as 39 cyclically ordered segments each containing 255 points. ▶17◀Whenever a particular segment is filled, it is marked as ready for transfer, and the filling of the next segment is undertaken. ▶17◀With small time intervals during the run, the Algol program asks the microcomputer whether a segment is ready to be transferred, and whenever the answer is affirmative, a transport is performed. In this way the transfer of data may without problems, be dalayed by the time it takes to fill 38 segments. *np6* The latter type of experiment with direct transfer to RC4000 is mainly being used for survey spectra, but with the automatic line measuring procedure even such spectra may often provide a quite good basis for assignments, calculation of the rotational constants etc. The accuracy of the measurements is mainly limited by the step length of the scan and not by the number of MHz pr. centimeter as for a paper recorded spectrum. *nl2* _«bs»P_«bs»e_«bs»r_«bs»m_«bs»a_«bs»n_«bs»e_«bs»n_«bs»t_«bs» _«bs»s_«bs»t_«bs»o_«bs»r_«bs»a_«bs»g_«bs»e *np6* All kinds of original spectra, whether they have been transferred to the RC4000 during the scan or after the experiment, are as a routine permanently stored on a magnetic tape, before they are altered in any way by smoothing routines, subtraction of background, etc. This means that at any time one can choose an alternative approch to extract the information from the spectrum. *nl2* CONCLUSION *np5* We believe that in general a configuration with several parallel experiments connected to a large central computer offers several advantages to the user. First the setting up of a control system for a new experiment is a relatively simple and cheap task, and secondly the exchange of experiences and programs between the users is easy. With respect to the microwave experiment we feel that the present set-up makes the microwave spectrometer much more flexible and provides several new posibilities, and our experience from the last several months is, that with very few exceptions all of the work on the spectrometer is made by means of the computer system. *nl3* REFERENCES *nl2* *ld12* *sj* 1 A. Lindgård, J. Oxenbøll and E. Sørensen, Hierarchical Multi-Level Computer Network for Laboratory Automation. Software - Practice and Experience. In press. 2. N.W. Larsen and B. Bak, The Microwave Spectrum of Methylene Oxalate. To be published. *ef* scope day wave wavekorr ▶EOF◀