Dear Colleagues,
This is my original problem description:
"We recently experienced the loss of lock during triple res Biopack
experiments, e.g. gNhsqc and other backbone assignment experiments. We start
with the lock as high as 95-100 and the gradients knock it down, fluctuating
between 50 and 70. At some point during the run the lock drops suddenly (not
gradually) and is lost.
This is an Inova 600 with a triple res HCN probe with xyz gradients, vintage
2003. The gradients connect to the probe though pins at the bottom of the
upper barrel.
The protein sample is 20 mM phosphate, 150 mM salt, 10% D2O, nothing unusual
about it.
The lock itself is stable, there is no problem with a locked sample over
time. Gradient shimming seems OK.
An overnight watergate run is fine, even with gradient strengths comparable
to the triple res ones.
I did not see any damaged/broken gradient connections in the upper barrel.
But I did not cut the heat shrinks for a better look.
I somehow think that the problem is related to a gradient problem, with
noise leaking out of the gradient amplifier affecting the lock. But I may be
wrong."
Many thanks to Mike Strain, Jerry Hirschinger, S. W. Simon Sham, Greg
Heffron, Josh Kurutz, Brian Breczinski, George Gray, Pierre Audet, Neil
Jacobsen and David VanderVelde for their prompt response and suggestions.
Also to Eugene Chernioglo and Ghirmai Meresi from Agilent's NMR service for
their help.
Some suspected an improper acquisition parameter or setting or suboptimal
locking. But we were not doing anything new and the problem was reproducible
time and again, with a variety of samples and triple res experiments.
Others pointed to a lock problem. During further troubleshooting there were
indications of lock misbehavior. The lock baseline was in the -30 to -40
range rather than ca. 0, suggesting some dc offset. Also the lock drop with
gradients on was about 30 units rather than the usual 10 or so. Voltage
measurements along the lock path did not reveal anything unusual.
A malfunction of the lock loop under the stress of the multiple strong
gradients in each scan seemed a plausible cause for the problem. The lock
loop filter and lock sense circuitry are part of the Z0/Z1 board in the shim
power supply. Indeed, the Z0/Z1 board failed repeatedly during the super
shim diagnostic program. But the replacement of this board did not resolve
the gradient-caused lock loss problem. Other lock control circuits are on
the lock transceiver board. The replacement of this board did resolve the
lock problem.
Thank you all!
Lazaros Kakalis
Rutgers University at Newark
Department of Chemistry
Olson Laboratories
73 Warren Street
Newark, NJ 07102-1811
Tel: (973) 353-5040
Fax: (973) 353-1264
kakalis_at_andromeda.rutgers.edu
Received on Thu Feb 07 2013 - 13:11:56 MST