The idea for MeasureNet® arose 10 years ago in this Department. We were using electronic data collection in our honors freshman laboratory section (20 students), but were interested in providing it to all freshman lab courses. Some sections numbered up to 200 students distributed over the four lab rooms. The equipment for each student pair in the honors lab consisted of a PC and an attached SCI-Labworks interface box with measurement probes. While student experiences were basically positive, it quickly became apparent that maintaining hardware and software for an interface-equipped laboratory was both time-consuming and labor-intensive. Furthermore, the relatively short expected lifetime of the PCs (3-5 years) guaranteed substantial ongoing equipment replacement costs. Extrapolating our experience with the honors section to the entire freshman program led to the daunting prospect of maintaining 100 PCs and data collection interfaces in the laboratories. We felt that there had to be a better way.
Bob Voorhees and Paul McKenzie of the Electronics and Instrumentation Facility of this Department came up with a solution. With self-preservation in mind, their idea was to replace the PCs and interface boxes with a new type of reduced-maintenance network system. After considerable research, the type of network chosen was based on CAN (Controller Area Network) protocol. The Robert Bosch Company in Germany developed this network protocol for use in automobiles. It is also widely used in environmental control systems of large buildings and industrial process control systems. It is much simpler than the protocol for an Ethernet network and it provided a solid basis for the development of MeasureNet®. An internal university grant funded the construction of a small, "proof-of-concept" system, which was successfully tested by students in our summer school program. We then applied for a patent and began seeking funding for final development, construction, and installation of the system in our laboratories.
Major funding was obtained in 1996 in the form of matching grants from the National Science Foundation and from the Procter & Gamble Fund (Curriculum Development Grant Program). The system design was finalized and installation in our laboratories was completed by the end of the 1996-97 academic year.
One significant modification was made between the grant proposal writing and actual final design and construction. We originally intended to include a stand-alone colorimeter as one of the available "probes" at each station. Instead, we discovered that it would be feasible to design and construct a shared diode-array spectrometer for inclusion in the network, greatly enhancing the system's measurement capabilities. Integrating a shared spectrometer into a measurement network was a major departure from traditional stand-alone spectrometer design, and it has proven to be of great utility and benefit to our program. It has allowed convenient incorporation of new experiments based on absorption and emission measurements.
The spectrometer was built using a miniature monochromator with an optical resolution of one nanometer and a wavelength range of 350 to 1000 nm. Its integration into the network makes it available to any station, thus providing high-quality spectroscopic measurements to a large number of users. To facilitate efficient sharing of this instrument, the design restricts the activity at the spectrometer itself to the actual measurement process only. Data are graphically displayed at the station of the student making the measurement, making the spectrometer immediately available for the next student. Collection of a typical emission or absorption spectrum requires only a minute or two, so little queuing occurs. After finishing the collection at the spectrometer, students examine the spectrum at leisure at their station.
The defining feature of the MeasureNet® system is that it is a network.
The components of this network are a central "controller", a number of
"dumb terminals" 
(stations),
and the diode-array spectrometer, all much smaller than a PC. The picture
at the left (the devices are not all to scale) shows these components,
with the stations mounted on wooden stands. In our laboratories, the stations
and spectrometer are permanently attached to the vertical surface rising
from the back of the lab bench. Mounting these devices up off the lab bench
puts them out of harm's way and takes up almost no space on the bench.
One of our major objections to PCs in the laboratory was the huge proportion
of available benchtop space they demanded, leaving little room for the
"chemistry".
The network controller contains the system software on a PCMCIA card. This software contains menus to be displayed at the stations and performs calculations to convert measured signals to values to display, etc. As software is changed to add new measurement types or modify the menu structure, simply reprogramming the card immediately updates the entire network (one of the main benefits of using "dumb terminals"). We have done this many times as the system has evolved and grown over the years, and it is much easier than upgrading software at each individual station. Although one network controller can support up to 12 stations, the layout of our labs called for 10 stations per network. Therefore, we built and installed ten 10-station networks, for a total of 100 stations. We upgrade all 100 in just a few minutes by simply reprogramming the 10 PCMCIA cards.
Preparing the system for an experiment consists merely of getting out the appropriate measurement probe(s), such as those for temperature, pressure, pH, voltage, etc. The pair of students performing an experiment configure the station for the desired measurements through selections from intuitive menus. As measurements are made, they are numerically and graphically shown on the station's green backlit display. The display's resolution is similar to that of a typical graphing calculator, although the data are actually measured at very high resolution (two-channel, 24-bit analog-to-digital conversion at each station). The station display serves to provide a preliminary view of the data, adequate to decide whether an experiment needs to be repeated.
Each network has one PC and printer attached to the controller as a peripheral (i.e., not actually on the CAN network). Software running on the PC interacts with the controller to provide the instructor with a great deal of information about the current status of each station. When a student wishes to save or print a graph of a data set they have collected, menu choices at the station cause the data set to be transferred to the PC as a simple text file for storage or printing. If the PC is also connected to a local area network, the files can instead be directed automatically to a server or network printer.
Since data sets are saved as tab-delimited text files, students have a wide range of choices for processing the data. For most experiments in our program we find it adequate for students to simply print graphs of their data and examine them to extract the desired information. The MeasureNet® PC software uses Microsoft Excel in the background to do the actual graphing and printing, so it is possible to use Excel to do some automatic processing of the data before printing. We accomplish this using "custom print codes". When a student prints a data set with a custom print code, the data are placed in a specific Excel workbook sheet with pre-defined manipulations and graphs. Saved files can also be easily imported into a variety of analysis packages, spreadsheet programs (e.g., Excel), on either Mac or Windows platforms. We are now testing a facility for directly storing data sets in a web database. Students will have accounts enabling them to store collected data sets and retrieve them (along with a variety of Excel macros) for later analysis from any computer with web access.
The new measurement capabilities brought to our labs by MeasureNet® have led us to make substantial changes in our lab program. We have been able to introduce new kinds of experiments as well as to modernize others we have had in our program for a long time. The most obvious changes concern spectroscopy, as we now have much more capability than before. We utilize this in experiments involving emission from discharge tubes and flames and in various experiments using absorption measurements.
Less obvious, but equally meaningful, is the modernization of old experiments. As an instructive example, consider a long-standing experiment on freezing-point depression. Before MeasureNet® each student worked entirely alone and, in the course of an afternoon, carried out two tedious cooling curve measurements (solvent and solution) by means of a mercury thermometer and a clock. They would later construct graphs of these data, determine the freezing-point depression of the solution, and calculate the molecular weight of the solute. In the modernized version, using MeasureNet®, a cooling curve is measured and printed in just a few minutes. We prepare a number of solutions before lab, using one or two solvents, two or three solutes, and several concentrations. We divide these up among the 10 pairs of students on one network when they arrive (three or four each), and the students measure cooling curves. In a fairly short time the group determines a large number of freezing points which are combined into one graph of freezing point versus molality. This graph is printed and discussed in the lab, followed by each student preparing an unknown solution and performing a cooling-curve measurement to get a molecular weight (to retain some level of individual accountability in the experiment). The overall result is a much-improved laboratory experience for the student. We follow a similar procedure in two or three other experiments.
One of our main goals in designing MeasureNet® was to reduce the presence of PCs in the lab environment as much as possible, both because of the difficulties associated with hardware and software maintenance (not to mention the tendency of students to use them in ways we didn't intend), and because of their relatively limited lifetimes. After 6-7 years of use (significantly beyond the expected 3-5 years of life) and increasingly numerous failures of drives, monitors, etc., the department finally found it necessary to replace all of the PCs. But because of MeasureNet®, this meant a dramatically reduced financial outlay since we needed just 10 new computers, rather than the 100 which older PC-based systems would have required. The MeasureNet® systems themselves are still going strong with only very few minor problems.
MeasureNet® is continuing to develop at a rapid pace. Under license agreement with the University of Cincinnati, MeasureNet Technology, Ltd. was formed in 1998 to market the system, and a patent was awarded in 1999. Systems are now in use in a number of types of institutions, ranging from high schools through two- and four-year colleges to major universities, and ranging geographically from Puerto Rico to California to North Dakota.
New capabilities and devices are continually being developed and added to the system. A UV-visible version of the spectrometer is now available (200-850 nm), and attachments for simple fluorescence and reflectance measurements are being tested. We have produced a compact optical drop counter, based on an infrared sensor, which also holds both the pH electrode and, when desired, the temperature probe, in a convenient arrangement for carrying out potentiometric titrations. The system software requests the beginning and ending buret readings, so the drop size is automatically calibrated in each experiment. We have also recently added an interface to attach a digital balance to any station, making it possible to record and graph mass vs. time, etc.
Modifications of the system software have also added new capabilities. For example, students can enter their answers (multiple-choice and numeric) to quiz questions at a station. Their score is immediately displayed at the station, and their answers and score are stored for possible later use as part of their grade. We have added a "broadcast" mode of operation whereby one station can take over the network and have its display duplicated in all of the stations on the network. This mode is useful in the initial introduction of system use, eliminating the need for many students to gather around one station as the buttons, menus, etc. are described.
We have also constructed a device consisting essentially of a controller and a station in one box for use in live lecture demonstrations. By connecting a laptop PC to this device and to an LCD projector in an electronically equipped classroom, it is possible to conveniently collect and display data in real time from many kinds of experiments we can do in the laboratory. Even quality spectroscopic measurements can be demonstrated with this device by bringing along one of the networked spectrometers.
Finally, we have returned to our original plan and will soon complete the development of the colorimeter intended to be part of the first system. The possibilities created by the spectrometer have been quite valuable, but the colorimeter is important for a variety of kinetics experiments not conveniently possible on a shared instrument.
URL for the Using Computers in Chemical Education newsletter.