Dear All
Happy New Year!
Thank you to everyone who replied to my request for information which
was posted just before the Christmas vacation.
A quick reminder of the problem.
Irrespective of sample/temperature/experiment/User, we noted that over a
period of time the lock level would drop (say from 80 to 10 or 0%, the
time interval for this was variable but of the order of hours). This
would happen even if nothing was running and a sample was left in locked
over night at ambient temperature. None of the experiments performed
during this time incorportated gradients. On checking the spectrum
before reshimming or resetting anything revealed that homogeneity had
been lost. Reading back the shim file made immediately before the
experiment started would result in the lock level be regained and
homogeneity restored.
We did all the usual things; rebooting software, rebooting console,
checking connections at various points (in particular RT shims). To no
effect.
We then switched off our gradient amplifier (a Highland L700) and have
been monitoring over the last few days. Lock level and shimming have
been steady for several days now, both with and without experiments
running (we have had the same sample in for ca 5 days doing various
things). I am not sure what is going on and clearly as we do not want to
have to switch the gradient facility off (to standby) some further
investigation is required - but at least we can do some work now! This
it seems is exactly what Josh Kurutz has seen (see note 4)so we will
follow this up with Varian - I would be interested to hear how Josh has
progressed with his testing.
A summary of some of the advice forwarded to me is provided below;
1. We have an INOVA 500 that behaved similarly. The issue was a
connector to the backplane in the console. The problem was actually
intermittent, so that it was difficult to diagnose until it began to be
difficult to recover the shims after they went bad. The problem
connector contains a bunch of wires (about 9) that all go into one
plastic connector that plugs into the backplane. One of the pins
actually burned and shorted.
2.double-check the connection to the RT-Shim stack under the magnet.
I've had people bump the connector when changing probes, and if the
connection is not good, the supply will max-out on voltage and error
out. Then, try halting EXPPROC, power-cycle the shim supply, then
re-boot the whole console.
3.It turned out to be a cable that connects the DC power to the
backplane where the shim boards are mounted. We replaced this power
distribution cable, and the problem never came back.
4.We traced the probable cause of our problems to, of all things, our
gradient amplifier. It seems the amplifier constantly puts out a
small amount of current to the z gradient in its resting state. If
this current is unstable, the current in the Z gradient is unstable,
which gives rise to fluctuations in Z1 field homogeneity that are
reflected in the lock level. Note that this particular homogeneity
issue has nothing to do with the shims. When I replaced our
theoretically problematic gradient amplifier with one from a
different Inova, we regained stability. Our solution, was to purchase a
"deadband circuit" that dams out these gradient fluctuations. It costs
about $660. We are in the midst of testing this.
Thanks again to everyone for the help with this.
Julie
Received on Thu Jan 11 2007 - 19:52:55 MST