Spectrometer G is our newest high resolution instrument. The magnet for this system is a medium bore 750 MHz magnet supplied by MAGNEX. The magnet has just recently been delivered. It has been successfully brought to field and is now in the process of being shimmed. As is our usual practice, the spectrometer for the system has been constructed in-house. This spectrometer is now complete. It has two channels of new transmitter electronics, a new receiver, and a new averager and pulse programmer. A temperature controller and waveform generator unit have also been constructed. The lock channel is a duplicate of those on our other high resolution machines. All of this electronics has been sucessfully checked out as well as is possible without real NMR signals. We will be using AMT amplifiers and a probe from Nalorac. While constructing this spectrometer we realized that we were going to have some difficulty constructing additional spectrometers in the future. The most serious problem was with the Ethernet controllers we had been using. For at least a decade we have been using Ethernet as the medium for connecting our spectrometers to their controlling computers. At the spectrometer end we have used a Multibus Ethernet controller manufactured by Interlan. Interlan has now stopped production of these controllers and there is no similar controller available from another manufacturer. Rather than making enormous changes in our spectrometer control software to support another board and risk having that board also go out of production, we decided to design our own board to emulate the current one. This actually turned out to be quite easy and we now have as many Ethernet boards as we will ever require. Our boards have comparable performanceand are significantly less expensive. Another major problem was the lock channel. This was originally designed nearly ten years ago, and several of the required parts are no longer available. The 750 MHz console used the last of our spares. We have completed a very nice design for a new lock channel which includes a T.I. DSP chip. Although our current design has always featured a computer controlled digital feedback loop, there had also been a fair amount of analog signal conditioning. The DSP chip will eliminate most of these analog components. The result should be an even more sensitive and stable lock than our current very good one. The pc boards for thenew lock have been layed out and sent out to be manufactured. The final problem was not quite as serious. We have from the beginning been using a Multibus Z80 CPU board to control each major subsystem in the spectrometer. Although the Z80 is an exceedingly old and quite slow microprocessor, its performance had always been found to be sufficient. Recently however, we have found ourselves running up against its limited maximum memory addressing capability. As we were also using up the last of our available stock, we decided that the time had come to design a new CPU board. The new design uses a highly integrated Intel 80C188 16-bit microprocessor and has more than four times the memory of the current design. Although the 80C188 is still not by any means state of the art, it will more than satisfy our needs for the forseeable future. The new boards have been fully designed and are expected in-house shortly. A fair amount of work will be required to port the current Z80 code, but because of the strong resemblance between the two computers, it should not be overwhelming.
Showing the most recent 10 out of 12 publications
