Bruce Williams, an electronics engineer with 18 years of experience
in NMR, has the following opinion:
Electronic Equipment
Regarding your question about the spectrum analyzer: I can honestly say
that in the eighteen years that I have been maintaining as well as designing
NMR spectrometers that I have never had to avail myself of what a
spectrum analyzer has to offer in order to diagnose a problem. Ninety five
percent of all problems require no more than a good oscilloscope, along
with a good set of schematics and a knowledgeable technician. A
multimeter is useful for setting or reading voltages exactly but a 'scope
will tell you more about the nature of the voltage being observed. In
addition, a 'scope can make use of a current probe to observe current
waveforms. The spectrometer itself is all the spectrum analyzer that you
will need. By observing the FID or transform of a reference sample, you
can determine if the problem is in the data system or the analog system
(images and such). In the few instances where a spectrum analyzer might
be used, other methods obviate the need for an expensive spectrum
analyzer.
If you have money to spend, a far more useful instrument to buy,
and costing less than a spectrum analyzer, is a synthesized signal generator
with calibrated output such as the Fluke 6060B. It is useful when you
suspect that the gain through a system is not correct. It is difficult to
impossible to measure the gain of a preamp using an NMR signal but is
easy with a signal generator.
We have found other instrumentation to be useful at times. These
include: 1) a power supply to power printed circuit boards that have been
removed for testing or preamps that have been removed for gain checks,
etc. 2) a capacitance meter, 3) a digital multimeter, 4) a sweep generator
and rf bridge for tuning probes or checking suspected bad probes, and 5) a
30db, 30 watt attenuator used in conjunction with the 'scope for testing
the outputs of transmitters.
If I had to list these in order of importance, it would be as follows:
1) oscilloscope
2) digital multimeter
3) 30 db attenuator
4) sweep generator/rf bridge
5) synthesized signal generator
6) power supply
7) capacitance meter
I consider the first three "must haves", the fourth almost so, the fifth very
nice to have but not absolutely necessary, and the rest nice to have but I
could get by without.
Expertise
A typical yearly "maintenance agreement" can run from $30,000 to
$40,000 per year. These include parts and labor. However, a typical
scenario is this: something in your spectrometer breaks so you call the
company. They say "sure, we will be glad to come out and fix it but we're
real busy right now and the soonest we can get there is two weeks from
tomorrow". However, he has trouble at the previous stop and is delayed a
couple of days. He finally calls and says he'll be there tomorrow.
However it is January and on the day the engineer is supposed to arrive,
all flights are canceled so he doesn't get there 'till Friday afternoon. He
looks at the spectrometer and determines that it needs a part which he
orders. However, it is the weekend and everything stops while he goes
home for the weekend. Because he lives far away, he doesn't arrive again
until Monday afternoon. The part hasn't arrived and on checking, he finds
that they had to backorder it. So he goes home and says he come back
when the part arrives. Well, you get the idea. All of the above has
happened to us when we had to call in an engineer from the company
when a spectrometer was in warranty. The point is that for the same
money, you can hire a good technician and have this person available at all
times. This is especially important in a lab that has a lot of spectrometers
such as ours - we have seven. Seven "maintenance agreements" would be
seven times more expensive than one technician. Enough said.