Test Equipment
Robert Santini (santini@chem.purdue.edu)
Wed, 05 Jan 1994 12:22:10 -0500
>Date: Wed, 5 Jan 1994 12:12:11 -0500 (EST)
>From: SANTINI@CV3.CHEM.PURDUE.EDU
>Subject: Test Equipment
>To: SANTINI@CHEM.PURDUE.EDU
>Cc: SANTINI
>
>Here are my comments on the value of various kinds of electronic support
>equipment in the NMR lab. This is in response to a posting to the
>discussion group by Bill Stevens. As Bill suggested, I am responding to
>the whole group in the hope that further discussion will be stimulated.
>I have and know how to use everything that was mentioned by him. I will
>comment on the subjects that he discussed below. Then, I will add some
>suggestions of my own.
>
>Stevens suggests the following equipment:
>
>(1) Nice multimeter . . .
>
>This is an essential item. The cost of a suitable unit is so low that I
>marvel how anyone can buy an NMR spectrometer and pass on this gadget.
>There are several reputable vendors here, but there are a lot of low
>cost offerings that are not really suitable. Read accuracy
>specifications carefully. These will tell you which units to consider.
>While Bill does not mention the display resolution, I greatly prefer to
>have mulitimeter with a total display count of 19999 (this is called a 4
>- 1/2 digit display) to the more common 3 - 1/2 digit displays. The
>cost of the better version has come way down recently. A hand-held unit
>is probably best for use around a spectrometer. Here, I prefer a unit
>that has a "touch and hold" feature. I have spent too much time trying
>to probe a circuit while simultaneously trying to read the digital
>display.
>
>(2) Oscilloscope . . .
>
>An oscilloscope is an essential item. This should be planned as part
>the basic cost of any NMR installation. For general use, I am not yet
>completely sold on the type that digitally capture a waveform. These
>are subject to Nyquist type sampling limitations, although these units
>are a lot more user friendly now. However, a traditional analog scope
>does not do a great job on low repetition rate waveforms. I use both
>types, often simultaneously. Either type is useful only if the operator
>fully understands its capabilities. As an example, it is essential to
>know how to use the delayed sweep mode to diagnose most NMR console
>logic circuits. Another important subject is the scope probe itself.
>It is not possible to attach a "standard" probe to a high frequency
>device without loading that circuit to some extent. The proper use an
>active probe with a built-in preamplifier and-or 50 ohm terminator is
>essential.
>
>The overriding selection parameter for an oscilloscope is its bandwidth
>specification. One can diagnose a sizable fraction of NMR console
>problems with a scope bandwidth of just 100 MHz. I regard this a the
>minimally acceptable value in a service oscilloscope. A 300 MHz
>bandwidth is suitable for all but the transmitter and decoupler circuits
>in a high field (say 500 MHz) system. Probably, 90% of all signal
>frequencies in the NMR system, i.e. fundamental plus at least a couple
>of harmonics, fall within the 300 MHz frequency range. One must be
>circumspect in the selection of a unit that can operate over a bandwidth
>of greater than 500 MHz. Expense aside, many of the digital units
>utilize a repetitive sampling principle above some nominal time per
>division setting. They are not effective in this mode unless the
>observed wave form is both completely recurrent and has relatively
>little noise associated with it.
>
>(3) Wattmeter . . .
>
>A directional wattmeter will indicate correct tuning of a high power NMR
>probe circuit better than any other low cost equipment. They are
>inexpensive enough that I leave one in-line on the decoupler channel of
>all of our systems. They are not particularly frequency selective.
>This limitation must be remembered if one attempts to optimize a high
>power transmitter with a lot of spurious frequency content on its
>output. That will be a problem only when the transmitter is very far
>out of tune or very sick. Again, once the characteristics of the
>measurement device are understood, it becomes immensely useful.
>
>(4) Frequency counter . . .
>
>A frequency counter can be used to isolate errors in a frequency
>synthesis circuit quickly. The important requirement is that the
>counter can be referenced to the master clock of the NMR system by way
>of an external input. (Some counter manufacturers call this a ratio
>mode.)
>
>The counter can give the frequency, or frequency ratio, to high
>precision (usually 8 or more significant figures). This is far beyond
>what can be resolved with an oscilloscope. This high precision will be
>required anytime that the NMR frequency seems to be slightly in error or
>if some offset frequencies are "missing" in an otherwise functional
>spectrometer. The cost of a suitable counter is small enough that its
>purchase should be planned as part of an installation.
>
>(5) Spectrum Analyzer . . .
>
>I have used a spectrum analyzer for many years. This is an example of a
>device of great subtlety and equally great cost. The learning curve is
>steep before it can be used effectively, thus I do not recommend its
>routine purchase in most NMR labs. It becomes useful if new circuit
>development or instrument modification is carried out. Then, the
>frequency domain data is invaluable. There are two reasons to consider
>its use. First, the spectrum analyzer has a larger dynamic range than
>almost any other measurement system (at least 60 dB). Second, it can
>record data without ambiguity at frequencies above 300 MHz. I use one
>above 300 MHz in place of an oscilloscope. My comments above concerning
>active probes and 50 ohm attenuators apply to this device.
>
>(6) What else to have ?
>
>I could generate a long list under this heading. The one item that I
>would recommend, that was not mentioned by Stevens, is a radio frequency
>sweep generator with a VSWR bridge. This is far more effective than
>anything I have seen to date for optimally tuning all probe circuits and
>RF filter boxes associated with the modern spectrometers. This is a
>must-have item for anyone doing inverse detection or solids work,
>especially if samples with high dielectric constants are encountered.
>The equipment cost can be justified. It will be used on a routine basis
>by all users. (If a spectrum analyzer with a tracking generator is
>available it will function well in this mode.)
>
>(7) Service contract ?
>
>I have not been particularly happy with service contracts that involve
>more that parts and telephone consultation. We have always been able to
>achieve better results by teaching someone who knows the NMR issues
>thoroughly the fundamentals of electronics. I believe that this works
>out since the NMR spectrometer is often its own best test instrument. A
>careful reading of the artifacts in typical test spectra can isolate 90%
>or more of all problems. After that, the choice is between swapping
>modules or trouble-shooting to the component level. I invariably
>isolate problems to the component level on most analog and RF boards.
>The computer board problems are best isolated with software diagnostics
>and swapped out. Digital boards that reside in the interface between
>the two extremes are usually field repairable.
>
>(8) Diagnostic Software ?
>
>Most manufacturers have some sort of test software that their own people
>use to set up the spectrometer during its final test sequence. If this
>is available it can be the best "test instruments" one can have. One
>real advantage is that it mostly stands alone in operation. The only
>way that I have ever been able to get this stuff is at the time of
>purchase and at additional cost. Good Luck!
>
>(9) Extender Boards and Cables ?
>
>It does not help to have nice measurement capabilities if one cannot
>actually set up the individual spectrometer modules in a physically
>accessible configuration that will run in the console. There seem to be
>two philosophies here. One vendor charges a sizable cost for these
>fixtures. The other tends to include them with the system. They have
>been a tie-breaker for us in the final stages of the purchase courtship
>ritual.
>
>(10) Electronics Training ?
>
>It is my sense that the practice of NMR is now so complex that at least
>three experts are needed at least part of the time. They are the spin
>mechanician, the UNIX guru, and the electronics zen-master. I have not
>seen many people who successfully manage all three disciplines for long.
>
>One way to get an electronics novice started is to send a willing person
>to a short course such as the one offered by NMR Concepts. The rest is
>a matter of personal motivation and experience. In a University
>environment, there are usually very good instrumentation courses in the
>Physics Department. Be advised that working with circuits is fun! It
>can be addictive over time.
>
>
>
Robert E. Santini
Department of Chemistry
Purdue University
W. Lafayette, IN 47907-1393
santini@chem.purdue.edu
Voice (317) 494-5227
Facsimile (317) 494-0239