Dear Sean Xie,
Perhaps I know this problem. When we upgraded ARX's to xwinnmr 2.0,
we found that signal averaging over a relatively short period led to an
increase in S/N. Clearly there were enough scans to be assured that the
conditions did not lead to major saturation of the signal. Then if the
experiment was allowed to continue for several hours, often (but not
always) there would be no signal left, not even the solvent peak. We lived
with this problem for a long time and mostly avoided it by acquiring many
blocks of no more than 256 scans then adding the FID's using the au program
"fidadd". Finally I worked out that the problem was caused by going to the
acqu window while the acquisition was going on. Apparently the phase
cycling information could get lost and the scans started to subtract.
Whether or not it happened probably depended on where you were in the 8
scans of the phase cycling of zg when you did it. Once I worked out that
this was the source of the problem, a fix was put into xwinnmr2.6. I am
quite sure that the source of the problem that I observed was gone.
However, I am now told by a reliable user of our equipment that it still
occasionally happens on the ARX. They ran exactly the same sample 3 times.
Both times when starting the acquisition with zg using a large ns, the
signal to noise ratio was quite poor. The one time it was run with a
relatively small number of scans and started with multizg to obtain many
files that were added after the fact, the signal to noise was quite
acceptable for the same total number of scans. So, although the source of
the problem that I found is now fixed (I have tried it many times and it is
fixed!), there may be something else that also can cause it. I have
encountered at least one other person who has had a problem of this type on
an older instrument that was upgraded. With xwinnmr 2.6 the problem does
not occur on a brand new Avance that was delivered in March. We ran
exactly the same sample as mentioned above on the new Avance. I know that
people at Bruker have been trying to track this down, but it is a pretty
subtle problem and may not be noticed by users who just think they have a
difficult sample. If other people have noticed this symptom, it would
probably be helpful to let Bruker know along with a description of your
hardware including EC levels. In the meantime, you will probably get
perfectly good data if you block average.
Jane Strouse
At 08:19 AM 6/29/00 -0400, you wrote:
>Dear All:
>
> I would appreciate it if someone could give me some help in
>upgrading xwinmr2.6. We have an AVANCE DMX500 NMR system (5 years
>old). Since April, we have been struggling with upgrading from
>xwinnmr1.5 to xwinnmr2.6 (patchlevel 5) with IRIX 6.5.5, and ran into
>many problems. Most of the problems have been solved with the
>support from Bruker. Now the last problem is: "the experiment will
>failed because of lossing FID after a while. It seems OK at the
>beginning and then we could see the FID after 1 or 2 hours".
>
> We have trying many ways including upgrading TCU and RCU, but
>the problem still exists. Actually, the new RCU (ECL06) did not work
>on my system (I do not know why? we tried two RCU boards). I have to
>put back my old RCU (ECL03) since the old one at least let users to
>collect data.
>
> IF you have a similar system (DMX500 with three
>amplifiers:BlarH100, Blax300 and Blax300; four channels), could you
>email me back.
>
> Thank you in advance for your help.
>
>Best regards,
>
>
>Sean Xie
>
Dr. Jane Strouse
Dept. of Chemistry and Biochemistry
UCLA
Los Angeles, CA 90095-1569
(310)-825-9841 - voice
(310)-825-0393 - FAX
strousej@chem.ucla.edu
This archive was generated by hypermail 2b29 : Sun Dec 31 2000 - 01:15:02 PST