Macs Prove Their Worth as High-End Lab Assistants
Abundance of software, ease of use make Macs the tool of choice for researchers
By Bruce Schneier
Macs are used extensively in the sciences. Not just for writing research papers and creating presentation graphics but also for instrument control, data acquisition and analysis, and scientific simulation. There are many scientific applications available on the Macintosh - commercial, free and custom-built - and scientists all over the world are taking advantage of them.
The Mac is also commonly used in the sciences as a front end to high-end workstations. Data collected and numbers crunched on workstations are often brought over to the Mac for final manipulation and presentation.
Why the Macintosh?
Users we interviewed cited ease of use and availability of hardware and software options as primary reasons for using Macs.
Michelle Limon, assistant research physicist with the Space Sciences Laboratory at Lawrence Berkeley (Calif.) Laboratory brought four Macs to the Amundsen-Scott South Pole station. "I was able to set up an AppleTalk network in five minutes inside a shack on the ice," she said. "Try doing that with NetWare."
In addition to software designed specifically for scientific applications, a wide range of established programs make the Mac indispensable for related tasks such as document processing, presentations, lecture aids and graphical visualization of data. Users said they found it easy to transfer data from scientific programs to graphics, drawing or word processing programs.
The Mac's use is widespread in the scientific community as a data-collection device, monitoring, controlling and collecting data from various recording devices. The most popular Mac instrument-control software is National Instruments Corp.'s LabVIEW, a complete software system for building instrumentation and analysis applications. Using various hardware boards, scientists can collect data from GPIB (General Purpose Interface Bus), VXI and RS-232 ports, or directly from laboratory equipment.
LabVIEW's icon-based graphical programming language lets users design an interface that includes instrument control, data collection, data analysis and display functions. The language is platform-independent; applications written using LabVIEW can be compiled to run on Macs, Windows machines and Sun SPARCstations. LabVIEW also includes instrument libraries, which contain controls for more than 130 common laboratory instruments.
Derek England, metallurgical development technician at Falconbridge Ltd. in Timmins, Ontario, used LabVIEW to design a temperature- and pressure-monitoring system for an experimental heat pipe. He said he wired a thermocouple and a pressure sensor to the pipe and used a customized National Instruments data-acquisition board and LabVIEW to analyze that data and display the results. "I recorded a few sounds on a Mac LC and incorporated them with alarm settings on the LabVIEW front end," England said. "Everyone could sit and do other work while the computer monitored the heat pipe."
England also designed an instrumentation system controlled by LabVIEW that can acoustically measure the length of a pipe. "We used a Mac IIfx with a National Instruments data-acquisition board. Most of the system was built in LabVIEW; we developed the DSP (digital signal processor) routines in [Wolfram Research Inc.'s] Mathematica and translated them to LabVIEW," he said.
Users who find LabVIEW's capabilities overpowered for their needs have other options. "LabVIEW tries to be everything for everybody," said Drew Weisenberger, associate coordinator for the detector group in the physics division of the Continuous Electron Beam Accelerator Facility in Newport News, Va. He said although LabVIEW would work for his needs, he can get faster results using Kmax from Sparrow Corp. "Kmax is mainly for electronic-module-based data acquisition and control," Weisenberger said. "With Kmax, you can have a data-acquisition system up and running in five minutes." Weisenberger's group tests detectors for high-energy physics research.
AxoData from Axon Instruments Inc. also provides instrument control and data acquisition. It allows researchers to use icons to specify instrument connections, create control features, and acquire and save data.
National Instruments makes data-collection boards for NuBus slots, including basic boards for Mac SE, LC and LC II machines. Each board has different numbers of channels, sampling rates, resolutions, ranges and other options. The boards move data directly between external test equipment and system memory, so real-time data-acquisition applications can use the Mac for routine functions while the board acquires data in parallel.
National Instruments has a family of DSP boards for the Mac that range in speed from 27 Mflops (millions of floating-point operations per second) to 40 Mflops. Additional hardware lets users connect thermocouples, strain gauges, thermistors and voltage sources, as well as process currents to the Mac.
Roger Ray, senior research associate at Northwestern University Medical School in Chicago, uses Macs equipped with National Instruments and custom-designed data-acquisition cards for electrophysiology research. "We needed to collect neuron-firing times at high accuracy and could not find any commercial hardware we thought could meet our needs," Ray said. "We designed and built a NuBus board based on Creative Solutions [Inc.]'s Hurdler prototype card." Northwestern wrote its in-house software using Symantec Corp.'s Think C.
Holly Gorton, associate professor in the Biology Department at St. Mary's College of Maryland in St. Mary's City uses a Mac IIci as a laboratory automation tool in her plant ecophysiology research. She uses an analog-to-digital and digital-to-analog board from Strawberry Tree Inc. connected to sensors and controllers for monitoring and controlling temperature and flow rates. A several-hundred-line Pascal program runs the system. Gorton said she chose the Strawberry Tree board over the National Instruments board "because [National Instruments] doesn't sell boards with built-in thermocouple compensation, which I needed."
Remote Measurement Systems Inc. sells ADC-1, an external instrument control and measurement device that communicates with the Mac via the serial port. Its software, ADControl, is designed to run the devices. Ron Tubby, graduate student in the Department of Environmental Health at the University of Washington in Seattle, used this system to examine how well rubber gloves protect the wearer from chemical solvents.
There are a number of Mac software options for manipulating, analyzing and presenting data. Among the most popular applications for this task is WaveMetrics Inc.'s Igor, an interactive environment for experimenting with scientific and engineering data. Igor allows researchers to analyze complex data using its powerful mathematical tools and display it using a variety of built-in graph and table formats. Igor includes a powerful macro language that allows users to automate such steps as importing data, analysis and printing results.
"Igor lets me control my data acquisition and analyze and graph the results in real time," said Harold Herchen, research assistant in the Department of Mechanical Engineering at Stanford (Calif.) University. "I can then export the figures for camera-ready copy in a few minutes."
Synergy Software's KaleidaGraph is a data-analysis application that supports mathematical data analysis, statistical modeling and graphing. "KaleidaGraph is easy to learn and use, is fairly quick, and creates excellent output," said David Legler, research associate and associate director for the Mesoscale Air-Sea Interaction Group at Florida State University in Tallahassee. His group does research on the role of oceans in Earth's climate.
Axon Instruments Inc. sells AxoGraph, a data-analysis and graphics program. Other data-analysis and visualization tools include Transform 3.0 and Dicer 2.0, both from Spyglass Inc. Transform assists in the visualization of 2-D data, while Dicer works with 3-D data.
"We do image acquisitions on a SPARC-based image-processing system, transfer the data over Ethernet to Quadras in the lab and analyze the results using Igor and Spyglass Transform," said Richard Bookman, assistant professor of molecular and cellular pharmacology at the University of Miami School of Medicine. Bookman's lab is studying how nerve cells release neurotransmitters by examining the relationship between electrical measurements of the surface area of a single cell and fluorescent images of intracellular calcium ions.
Simulink from The Math Works Inc. is an interactive system for analyzing, modeling and simulating dynamic nonlinear systems. It is built on Matlab, the company's numeric computation and visualization environment. Werner Truoel, research assistant in the Automatic Control Laboratory at the Swiss Federal Institute of Technology in Zurich, Switzerland, uses Matlab and Simulink for the analysis of dynamical systems with polytypic parameter uncertainties.
Scientists who are capable programmers sometimes get the best results by writing their own applications. Derrick Mancini, postdoctoral fellow at Uppsala (Sweden) Universitet, studies the energy that is emitted from solids and gases when they are bombarded with high-energy electrons and X-rays. One of Mancini's graduate students, Tomas Wiell, wrote a Mac program in Think Pascal to do this task. "A Mac takes data from our detector in real time and sums it over time into a two-dimensional array," Mancini said. "In the foreground, the program simultaneously does the calculations to map the array into an energy spectrum and periodically updates the results on a screen."
Researchers can work within the program or switch to other applications while data acquisition goes on in the background. "We considered running a program supplied by the hardware vendor on a PC, but since more and more people were using a Mac for other applications, it made sense to use a common platform," Mancini said. Other users of the same detector have expressed interest in acquiring the Mac software developed by Mancini's group.
At the Department of Molecular and Cell Biology at the University of Connecticut in Storrs, Conn., Professor David Knecht uses a variety of shareware, commercial and custom programs to design and analyze gene sequences and to track manipulations of DNA molecules.
At Iowa State University in Ames, Kenneth Jolls and his students are producing 3-D graphics of thermodynamics equations that have never been graphed before. The research started on a Digital 5000/200, migrated to the Macintosh and will move to the PowerPC.
Many researchers still find the Macintosh unacceptable for complicated scientific research. Northwestern's Ray uses workstations for high-end data analysis. "Because of performance issues, we use the Macs as the laboratory front end, workstations as powerful analysis and data-presentation nodes, and Macs again for integration into finished publication," he said.
For Jamshed Ghandi, project engineer at Becton Dickenson Research Center in Research Triangle Park, N.C., the Mac is just a front end to a more powerful processor. He does numerical modeling of plastic disposable medical products for structural integrity and uses a Quadra 900 to display X Window System applications running on workstations and on a VAX 9000-320.
Schneier.com is a personal website. Opinions expressed are not necessarily those of Co3 Systems, Inc.