thank you very much for the fantastic and fast response I received on my
questions on "Room Preparations for NMR". This is really a great group.
It is great to know that one can get help that fast and knowledgeable.
Thanks to everybody who responded. Here is a summary of the more
detailed answers. I apologize for not being able to include everybody's
answer. Since I think the most useful way to present the answers is
sorted according to question number, I cannot include the names of each
contributors. I apologize for that as well.
At the end I will also include tips for the installation which were
given on top of my questions.
My original e-mail and questions were:
we are preparing an area for a 200 MHz NMR spectrometer for our
chemistry teaching laboratory. Since it has to go cheap, I would like to
ask your opinion on the necessity of the following items. In particular
I would like to know:
1) Concerning fluorescent lights, do they cause problems? Which
frequencies do they typically emit. (In the room I could detect 160 MHz,
125 MHz, 80 MHz, 64 MHz on the oscilloscope mostly related to a weather
station.) Are these frequencies clean enough to cause identifiable
spikes or do they vary? Is it worth the money to install a special
low-RF light ballast (on 15 lights in the room)?
2) Do you regard a separate earthing cable aside from the earth in the
normal three prong connectors a worthwhile investment?
3) Could you please comment on the effects to consider from conduits and
piping in the ceiling close to the magnet.
The majority of answers stated that no special measures were necessary
especially not for a 200 MHz teaching instrument.
Here are some more detailed answers.
1.)
I have seen some effects from fluorescent lights (which disappear when
the
lights are extinguished), but they seem to be minor and I think it isn't
worth replacing or modifying them.
If this is a new machine I wouldn't worry. Things should be adequately
shielded. If not, I would take a wait and see approach. If there is
interference, I have found that double shielded cable such as Belden
9913
between the T/R switch and the console helped alot.
I have a Gemini 200 running as an open-access NMR in one of my labs.
The
instrument is on the same power as everything else in the building that
has
emergency back-up, though the NMR has a line-conditioner. We have
standard
fluorescent lights (but on separate power from the NMR). The power has
been fairly carefully routed away from the magnet and RF console, and
the
re-inforcing bar in the concrete under the magnet is aluminum, and the
room
has anti-static carpet. We don't have noise problems detectable in any
standard experiments. I imagine you could get away without being as
careful. I think that most of the noise issues wouldn't make it into
experiments until you go to much higher field/sensitivity/much longer
experiments.
We have fluorescent lights in our NMR rooms: we have 300, 500, 600
instruments. There is no obvious problem. We tried "on/off" tests
during
the installation. Try to keep the lights as far as possible from the
magnet...10 feet, maybe. You might want to consider the newer fixtures.
We have incandescent track lighting in our NMR room (500 MHz magnet).
It
works very well, and it's convenient to have directional bright light.
We
leave it off when we don't need it, and the bulbs hardly ever burn out
(if
it's not kept on 24 hrs a day).
This is a perennial problem for me in several labs where FT processing
of
data is used, including NMR, FT-MS and Ion Trap MS. I can clearly
demonstrate this type of noise with an RF spectrum analyzer at many of
the
frequencies that you mention. This does not surprise me given that the
lamps are essentially gas discharge tubes. (BTW, I was not aware that
there were special low-RF ballasts available, but I would very much
appreciate a lead to obtain more information about them.) We have tried
all solid-state ballasts. They seem to help, but they are about 4X as
costly as conventional inductive ballasts. Another thing that seems to
help is to replace the lamps themselves every two years or so.
Apparently,
mercury vapor gradually attacks the filaments. This creates a high
frequency noise source over time.
I never saw influences from these lights to any spectrometer.
Furthermore I never saw any influence from stray frequencies
even from neighboring rooms. But to be somewhat more
precise: this is valid for a correct cabeling only. After
some years the cable connectors become dirty and this can produce
some odd effects. But a simple cleaning procedure of the connector
is sufficient to remove all frequency artefacts from the spectra.
We have normal flourescent lighting in all of our NMR labs and have
never
seen a problem from them. Flourescent lighting puts out mostly 120Hz
radiation. You might want to have the lights directly over the magnet
removed, so you don't hit them with accessories and tools, like the
cryogen
transfer lines. In some cases we removed the suspended ceiling tiles
for
the same reason.
no you cannot put any fluorescent lighting it is a night mare to
controll frequency problems.
2)
I would strongly recommend that the instrument be on its own circuit
(breaker), but the existing ground should be fine.
We have never had to to this, only the power line ground is used on all
of
our instruments. You would also have to worry about ground loops if you
provided multiple ground paths.
You should definitely have "dedicated" circuits for the NMR console and
host computer. We also had the electricians install "isolated ground"
circuits for the computers and computer-peripherals. Normally the
ground
wire in a recepticle box is connected to the third-pin of the recepticle
and is also jumpered to the metal box/conduit. In an isolated ground
circuit, the ground wire goes from the third (ground) pin of the
recepticle directly back to the electrical panel with no other
connections
Each outlet has its own ground wire. This helps to avoid RF noise from
ground loops.
This has been very effective for me in a building where the power is
distributed in a 4-wire "delta" configuration. There, it is well worth
the
cost ($1000 per spectrometer). Please note, you must route the lead to
a
separate earth ground inside a conduit. In our 5-wire "whye" configured
building, the power line ground seems to be adequate -- so far.
A good Earth ground is a good idea if it really goes to Earth, and not
back
to the same conduit ground you are concerned about. For instrument
grounding in our department, we have a direct copper connection to the
concrete reinforcing wire under the foundation. This connection is not
always necessary, but I recommend it if you run into noise problems.
You
should have it available if you can afford it to be installed properly.
PFG and V.T. cables can transmit noise into the probe, so they might
need
to be separately shielded and grounded.
yes you are best protected when you shield with indipendent earthing
using copper rods to nearest water table.spikes do not take kindly to
electronics espacially rf racks.also the shielding room should be
grounded
in the same manner.
3)
As far as conduit (and other magnetic materials in walls and ceilings)
are concerned, you should have no problems as long as you satisfy the
specified clearance in the manufacturer's siting instructions. This
assumes that these magnetic materials are stationary. Any residual
field distortions will be shimmed-out during installation, and you
shouldn't have any problems unless some of the magnetic structures
nearby are moved (this actually happened to us once...some magnetic
solvent containers were stacked on the opposite side of a wall near the
magnet, and our non-spinning shims needed major readjustment).
As long as they don't move, you should be able to shim out any effects.
If
they vibrate, moving metal in the fringe field could modulate your
field.
Avoid placing the magnet adjacent to conduits containing high current
conductors (like the building mains). Also avoid placing the magnet
near
big electric motors and transformers: i.e. be aware of what's on the
other side of walls/floors. Light use conduits don't seem to be much of
an issue.
This is not much of a problem if things don't move around. We had an
unfortunate situation recently where a cast-iron drain line under our
500
MHz magnet was rotated during maintenance on the ceiling of the lab on
the
floor below. The magnet required adjustment of its cryoshims afterward.
Power line conduits don't seem to cause much trouble provided they were
assembled with conductive pipe sealant at the threaded joints and they
are
well grounded. This precaution appears to suppress noise due to both
electrostatic and inductive pickup.
You need clearance for He fills.
Other than that, you'll need clearance for the guy with the heavy rusty
acetylene torch that'll want to work on the conduit control
conveniently located
right over your magnet. (In other words, any servicable parts near the
magnet?
What do you do when a pipe bursts and work needs to be done? - re route
or
consider the location). Advertise that the magnet cannot be turned off
(except at a cost of $20k or so :-)) and rope off the forbidden zone
with
plastic fencing (black and yellow plastic chain link helps too :-))
We have and electrical room (lots o' high power relays & goodies) just
out
of the 5 G line, no problem.
Conduits and pipes have been rerouted so that any maintence will
hopefully keep
the workers a somewhat acceptable distance away from our magnets.
The spectrometer vendor will provide you with specifications as to
circuit
requirements. If you don't have these already, ask the sales rep for a
copy of their siting requirements. Varian provides a very complete
booklet and I'm sure Bruker does as well. (They will send you this info
even if you haven't ordered the system yet). Also, they have power
monitors that can be sent out to your lab to check for voltage
fluctations, spikes etc. On higher field systems this is done routinely
as part of the installation package. You might have to make a special
request for the 200. If you have old, crudy circuits, this might be
justified.
if the piping is not shielded and is within the five gaouge line it will
certainly affect the signal.a spacial rf shielding is availeble for such
application.
4. Additional tips.
It is unlikely that any of these issues are of concern. If you are
relatively vibration-free, at 200 MHz you should be ok, especially if
structural elements are outside the 5G line.
We had a surprise when the liq He transfer tube on the first fill with
the removable extra long extension went beyond the ceiling. Luckily
fluorescent lights weren't there. We simply removed a ceiling tile
(permanently).
Be sure to allow extra circuits for auxilliary equipment:
oscilloscopes,
heat guns, soldering irons,.... This may be your best chance to get
things set up right: take advantage of the opportunity. It is harder
to
add/change things later.
Also, be sure to thoroughly clean up after the electricians are done:
they generate lots of iron filings.
As it turned out with our installation, the most obvious problem was
floor vibration causing some spikes near the peaks. The vibration is
caused from an air conditioner. The air cushions removed most of the
noise. I am trying to get the fans balanced but those things seem to
take time. We have ICP , Xray , mass spec, local radio stations , etc
and dont see any intefering signals. There was some power supply
sidebands on the spectra but those were removed by shielding the power
supply with a metal plate. Good luck.
Remember that the floor above the magnet will have a strong field, so
you will probably need warning signs in that room too - any unlucky
pacemaker wearer who went into the storage area above our magnet would
have a bad day indeed (we have signed posted
because of this).
The main source of problems in our lab has actually been temperature
fluctuation. This may not be relevant for a 200 MHz magnet, but we see
lots of parallel and anti-parallel diagonals, which can be traced back
to
cyclical fluctuations in temperature (over several minutes, since the
thermostat has a max and min difference of a few degrees). Forgive me
if
I'm telling you things that you already know, but if you haven't thought
about it, make sure there's lots of steady, constant temperature air
flow
through the area.. good luck!
I believe that it is important to suppress spurious pickup from the
environment. There are so many ways that it can show up from sources in
the spectrometer itself over time.
The newer multi-sync video monitors on many PCs are major noise sources
in
the range 15 kHz to several MHz. These can get into the system via
their
many harmonics.
The stray field from a 200 is not very large. Check the site
preparation
manual for your instrument to see if the 5 or 10 Gauss field will even
extend into the floor and ceiling. My guess is probably not very far.
Even if they do, conduit and piping should not affect you as long as
they
are not moving. Their effect will be shimmed away when the magnet is
energised. If work is planned to move, install, or remove these things
directly above (or below!) the magnet, it should be done before magnet
energisation. Also, check to be sure that there are no steel cabinet
doors
or refrigerator doors directly above or below that will be swinging
regularly.
Interesting question and difficult to answer. I had a probem here
and I will try to describe it. For our highfield machine we need
active cooling of the magnet (donw to 2K) and this is done by
a vacuum pump. This pump must work without any interrupt
and that's why we have a USV of course. There batteries are sufficient
for about 10 hours for the pump only but for only one hour
together with the spectrometer console. That's why the magnet
control unit stops the power supply for the console after 1 minute
without external energy to ensure the life of the magnet.
To avoid this unwanted effect we connected the whole system
to an internal emergency power supply system, which should
be unavaible for a maximum time of 15 seconds only. The
connection was done by a 100 meter cable with 4 mm2 copper.
For a currect of about 15A this gives a voltage difference of
about 5 Volt over the complete distance. At the location
of the spectrometer due to this effect we have the called
voltage differenc between internal ground (15A) and earth
(without current). An this difference gave odd effects.
We had to choose bigger cables (10 mm2) to ensure stable
spectrometer work.
Thank you very much again.
Ulrike
-- Ulrike Werner-Zwanziger Department of Chemistry Indiana University Bloomington, IN 47405Phone: (812) 856-4629 Fax: (812) 855-8300 WWW: http://nmr.chem.indiana.edu/~uwz