Dear AMMRL,
Thanks for the helpful responses about shimming to:
Thomas Dickinson, Joe Dumais, Charlie Fry, Jerry Hirschiger,
Josh Kurutz, Robert Peterson, David VanderVelde.
The consensus is that high order shims are not to be adjusted manually
because their main component is weak and is highly mixed with other shims.
Responder #6 mentioned an engineer from Resonance Research suggesting
specifically to never adjust X3 and Y3 manually.
* stick with manually shimming Z1-Z4,X,Y,XY,XZ,YZ,XZ2,YZ2,X2-Y2 (resp #6)
* for other shims use 3D gradient shimming - looks like this is the only
solution available on-site to most labs.
* check/adjust cryoshims (resp #3), if these are not set correctly, high
order shims may not be able compensate for imperfections.
* use magnet mapper instrument (resp #3) - reduces shimming effort from
weeks to one day, b/c mapper looks at larger volume than probe.
Thanks!
-Evgeny.
Below is the content of the entire thread.
0) --- My question-------------
Dear AMMRL,
What's the idea behind adjusing Z3X and other high order mixed XYZ shims?
Has anyone done this? I notice that some of those are set quite high on our
system.
There's a good description on how to shim up to Z2X and such in
shimming ain't magic
http://www.acornnmr.com/Sam/shimintro.htm but there is
no mention
of higher order shims.
I'm hoping to improve a lineshape and sensitivity a little (especially
lineshape near the baseline)
on our 800 MHz system.
Thanks. Will appreciate any advice and post the summary.
1) ---------------------------------
Have you tried Varian's 3D gradient shimming for systems with z-only pfg's?
It's done wonders for our 500. Christine Hofstetter's the expert on this.
2) --------------------------------
My experience with high order shims is they are typically adjusted once
during initial mapping & installation of a magnet, then left alone.
Once the magnet drifts over time, remapping can be done to once again
improve homogeneity.
I'm not sure which brand of NMR you're using but Varian has recently
developed a gradient 3D shimming package which can be used periodically to
reshim all the high order shims and help maintain great line shape and water
suppression. We've used it here and the results have been fantastic.
3) ---------------------------------
Before putting very much effort into adjusting the higher orders with the
room temperature (RT) shims, I suggest checking the cryoshims first. Do
that by setting the RT shims all to zero and checking the raw lineshape of
tapwater. If it is assymetrical, or has a very broad base, you might want
to adjust the cryoshims first. A good cryolineshape should be less than
~30/300/1000 and symmetrical but it is widely different for each
manufacturer, depending on the strength of the RT shims. Check with the
mfr. to see what their recommended cryolineshape is for that system. If it
is worse than the mfr. suggests, then your RT shims might not have enough
power to correct it, and the magnet may need the cryoshims reset. All shims
are not equal strength. The higher order gradients are the weakest shims,
so having the cryoshims set correctly is of great importance for the higher
orders.
If the cryolineshape is OK, or you want to try RT shimming even if it’s not
ideal, then I recommend using a field mapper to set the higher orders. That
is really the only way to determine just which shim needs adjusting and by
how much. Otherwise, RT shimming the higher orders is a guessing game
because the shims are so weak it is very difficult to see their effects. It
can be done without a field mapper, but it requires very diligent methodical
repetition of shimming sequences, looking for very small changes in lock
level or lineshape – a seriously difficult task. The best sequence to
adjust the shims is dependent on the inhomogeneity gradients present and the
strengths of every other RT shim, due to interaction. A field mapper will
figure it all out and tell you exactly what to move and how. Usually only 2
or 3 mapping and shimming iterations are necessary, so the mapper will do
the job in a matter of 4-8 hours, rather than the 4 – 8 days and more you
can easily spend trying to do it by hand. The reason a mapper works so
efficiently is because it maps a larger space than the sample occupies, so
it can see the inhomogeneity with better resolution in the outer areas most
affected by higher order gradients. Shimming by hand only gives you a
picture of the sample, where the resolution of high order inhomogeneity is
poor.
4) -----------------------
Those high order shims are important. But they don't change too much from
sample to sample. They should be set pretty well by the installation
engineer (or by gradient shimming). The numbers can be quite high.
If I don't have 3D gradshim, I always go through the X and Y shims to the
highest order. I do them in pairs (X, Y) (XZ, YZ) (XZ2, YZ2) and so on, to
the highest order, then back down to the lowest order. I just do it on the
lock level. Although the numbers don't change too much, I've seen the lock
increase quite a bit and the shimming improve a lot by doing this.
I almost never see any effect from the other off-axis shims (XY, X2-Y2, X3,
Y3, and all the combinations with Z).
5) -------------------
I am not sure if you meant to ask how to judge which high order shim to
adjust from the lineshape or why some shims have such a high current value?
At the risk of stating something you may know I will take stab at answering
this.
The higher the order the more the shim begins to only significantly impact
the fringes of your sample volume. Due to the way these high order affects
are created within the shim and the space limitations within the shim set
(which are severe) these shims are generally very weak and that is why the
numbers are so high. You have a few volts positive or negative to work with
on each shim supply and to get a significant impact from a high order shim
you need to use most of that voltage. Shims used to be complicated
windings dedicate to producing a single shim as best as possible but issues
like the leads entering into the windings and other imperfections in the
wire, space limitation (again) and the winding caused the result to not be
purely the intended one. One of the solutions to this that has been
employed successfully is based on using a set of shapes of conductive strips
that are created on a flexible sheet and then wrapped into the shim tube
instead of the old windings. The resulting affects on the magnetic field
from running current through this set of shapes can be decomposed through
linear algebra to produce a set of coefficients corresponding to the
intensity of the current applied to each of the available circuits to
produce more pure shims that one can with simple windings. As I understand
it, applying this type of shim required mapping to generate the exact
coefficients for that shims set. This was referred the as a matrix shim
set. It is my understanding that this idea was developed and applied by
Piotr Starewicz from Resonance Research.
At least one of our very high order shim is at its maximum on our 600 and
others are quite large. It is not bad to have one shim that I have always
simply shimmed these by the lock and generally the response is quite weak.
6) ------------------------------------
One of the Resonance Research people we had here (can't remember his name,
but was a good guy) told me that many high order shims should not be
"manually" adjusted. His view was that the shim space with modern shim sets
is too large for human, linear thinking to adequately work with. The
typical result of people adjusting high order shims is then a deterioration
in line shape quality. [Of course, at the time he was performing an
automated shimmap of a couple of our magnets (one with 28 shims, the other
40). So it was to RRI's benefit if I believed him.] But his comments were
compelling, especially considering his experience with manual and mapped
shimming. He specifically talked about not adjusting X^3 and Y^3; Z3X would
have been included I'm sure as one to not adjust.
So more shims allow automated shimming procedures, by his view, to get all
shims to have small to moderate sized current values, and to enable the
workable shims (about 12 or so) to fit better when adjusted in.
Now that does not necessarily assist you in trying to eek out that last bit
of line shape (unless you're willing to invest in a remap of the magnet).
But I've always stuck to the primary set of shims when manually adjusting
(Z1-Z4,X,Y,XY,XZ,YZ,XZ2,YZ2,X2-Y2). Occasionally I'll try something else,
but nearly always end up setting it back to the stored value. Will be
interested in comments back, if sufficient to summarize.
--
Evgeny Fadeev, Ph.D.
Director, BioMolecular Spectroscopy Facility
1212 Natural Sciences 1
University of California Irvine
Irvine, CA 92697
telephone: 949-824-5842
http://www.physics.uci.edu/~biomolenmr
http://nmrwiki.org - Share your magnetic science!
Received on Tue Oct 20 2009 - 08:27:23 MST