My question about how close one could go to the lock frequency produced
a range of responses; from "what problem" to the total drop out
we experienced. Many made the point that for the line width involved,
lock is not really needed anyway.
At present, we are running off lock. The use of the filters refered to
(lock reject/bandpass and lowpass) is routine in this lab.
The replies are below, with headers removed.
========================================================================
I've been doing O-17 like a madman of late, using our GE Omega-500,
and I get wonderful (if not "total") isolation between my observe
channel tuned for O-17, and the Lock Channel. On other instruments,
those on which I've had problems...like my 200, I have observed problems
with the actual probe-circuits coupling with each other (i.e. when I
tune the observe probe-channel @ O17 freq, the 2H tuning dip degrades).
If you haven't already done so, double-check the tuning of the lock-channel
of your probe, after you tune the observe channel for O17. I _have_ done
sucessful O17 on my 200, but have had to iteratively tune the lock and observe
sides of the probe, and the lock still jumped when pulsing oxygen.
FYI, the lock signal is TOTALLY unaffected by pulsing O17 on my 500...even
when I'm pulsing a hundred time/second!
===========================================================================
Andy - We have done a lot of 17O here at both 300 proton and 500 proton, and
we always detune the lock, turn the lock trap off, and turn off the lock
controls. We run the spectra unlocked. If we want to shim on the lock, we
set up for 1H first, lock, shim, and then set up for 17O. You do not have
to detune the lock unless you are seeing interference patterns, glitches,
etc. We actually never considered running 17O while locked, so I guess I
can't really answer your original question, since we never have tried to run
in that configuration.
=============================================================================
We have seen this during 17-O operations. It is largely due to the lack of
geometric isolation inherent in a single coil probe. Instead isolation is
achieved by time-sharing the observation and lock systems.
Ten Mhz is not much of a separation due to the extreme amplitude difference
between the NMR signal of interest (in the microvolt range) and the unwanted
white noise on the transmitter output ( about 0.1V). Even with isolation
diodes, the broadband transmitter noise irradiates the lock receiver
sufficiently to cause partial saturation of the lock pre-amp. Any switching
transients on the lock transmitter only exacerbate the problem. BTW, the
lock transmitter gating frequency will often show up the observation channel
of the spectrometer by a similar reciprocal mechanism.
The solution is to use very high Q bandpass filters or a pair of sharp
cut-off high and low pass filters (on the pre-amp inputs) to reduce this
effect. These filters are not easy to fabricate, so they must be made by a
specialty supplier. The filters should be specified for about 60 dB of
rejection at the frequency of interest. On some spectrometers, it is
possible to use a diode blocking circuit on the transmitter outputs that is
tuned to the specific carrier frequency with an additional parallel LC circuit.
==============================================================================
Dear Andy-
I don't know if this will help answer your question or not,
but when I was in grad school I ran a lot of 17O spectra in 20mm
tubes using a Nicolet NT-200, and we used a concentric 5mm tube
filled with D2O for lock. At the time I wasn't very well versed in
electronics, so I don't know how much crosstalk etc. there was bet-
ween the lock and 17O observe, but I know that I didn't have a lot
of trouble locking the samples even with the small effective fill-
ing factor for the D2O in the 5mm tube, and the old NIC's didn't
have very good lock sensitivity in general. So, obviously at 200
17) and 2H are a lot closer than at 500, and I didn't really have
problems with it.
==============================================================================
Hi Andy, a couple of years ago I ran 17-O NMR on a GN-500 spectrometer
(General Electric) without any problems with 2-H lock. I don't remember if I
had to use 2H-Pass Filter in the lock circuit.
=============================================================================
Have you tried retuning your lock channel? Tuning the BB channel close
to the 2H frequency may be giving your problems.
Recall that many probes come from the vendor with their lock channel
slightly detuned anyways so they won't be too sensitive (I don't
exactly understand the reason for that myself - I always retune them on
d2o at 25 C). If your lock channel is detuned to the low side of 77
MHz, maybe the BB tuning is interfering.
Another item which isn't clear to me is whether you lose the lock once
you start pulsing on the BB channel or just as soon as you attempt to
tune. If it is occuring when you are pulsing, then it probably is rf
bleeding through into the lock channel. Dig up a 2H pass filter or even
try putting two in tandem to double the cutoff slope for use in the lock
channel. I once had a problem where an ungrounded copper air line
connected a lab containing a 300 solids instrument running carbon (lots
of power at 75.47 MHz) and a 500 solutions instrument. Every so often
the 13C pulse would coincide with a lock channel receiver and the lock
level would drop about 20%. It took me quite a while to figure out what
it was (grounding the copper line took care of the problem). Anyways,
the point is that maybe the 17O pulses are getting through to your lock
receiver. Try cranking up the lock power and turning down the gain and
see if it still happens.
So to sum up: (1) retune the lock channel, (2) filter the lock channel
better, and (3) crank up the lock power and turn down the lock gain.
==========================================================================
There is a student in our lab who is observing O-17 on an ARX500. The
lock is fine. She is using a 10mm broadband probe. It has been a while
since I looked at the tuning myself but I am the one who got her set up in
the first place. If you have some specific questions, I could go play with
it myself again to see how ours compares.
============================================================================
Why lock in the first place? O-17 is sufficiently broad that a bit of
inhomogeneity or drift won't make any difference. If your magnet drifts
enough to worry about, there is circuitry to correct it. We never lock
when running metal nuclei spectra to save the expense of deuterated solvents.
Good luck.
=============================================================================
Can you view your probe with a swept rf source and a bridge looking
at both channels simultaneously? The interaction between the channels
will be quite significant and you will have to tune them
interactively. You may also require some high Q bandpass filters for
each frequency to improve the isolation. In the past I have used two
sets of filters to separate H1 from F19--each filter package consited
of two filters..i.e. H1pass & F19 reject for one channel and F19pass &
H1 reject for the other channel. The filters were rf tubular tuned
cavity devices. For your application you can probably purchase small 1
inch diameter bandpass devices. They are commercially available with
various performance specs depending on the number of poles, bandpass
characteristics and insertion loss. You will have to check and specify
about 70db of isolation between O17 and D frequencies. Check with the
manufacturers for current suppliers.
==========================================================================
17-O was spec'ed out on our Omega-500 spectrometer when it was installed
in 1988. We have never run it since, however. I would highly recommend
that you check to tuning on the lock channel as well as on the X (17-O)
channel. I suppose it's possible that strangeness of coil/tuning design
of your particular probe could lead to problems. In general experience
I think the 10 MHz difference shouldn't be a problem.
Good luck,
===========================================================================
I never use a lock when I am recording oxygen-17 spectra. Just like
sample rotation the lock really is not necessary for 17O, since the lines
are very broad anyway (usually several hunderd Hertz).
I just turn off the sweep and shim on the shape of the FID. It is even
recommended to disconnect the lock cable completely from the probe to
avoid any interferences.
For other nuclei that are close to 2H try using a 2H-Pass-Filter in the
lock channel and/or a 2H-Stop-Filter in the Xnucleus channel.
===============================================================================
Hi Andy,
We frequently work within 5MHz of the 2H lock freq. ( at varying
fields 250 - 500).
Some points to note are:
Shim the sample with a different X tuning (especially with low lock
signals) thus avoiding "cross talk" .
Check if there are any 2H filters in the pre-amp circuitry and by-pass.
If you need to keep the lock on (I dont understand why though) then
turn down the power to the lowest possible.
Better yet run unlocked if possible disconnecting the lock cable from probe
and terminating with 50 Ohm. You may be surprised by how much better the
s/n appears depending on your instrument.
The 10 MHz restriction you mention derived, I believe, from the Bruker 2H
stop fitted to some pre-amps which if in place would filter 2H plus or
minus 10MHz.
==========================================================================
I think what you need is a lock bandpass filter in your lock line. The
lock bandpass is 77MHz +- 3MHz. If there is one already you should
check the filter cutoff of the bandpass filter. I haven't measured 17O
for a very long time but I think that the filter is good enough that
you are getting a proper lock signal. In your observe line you should
use in addition to the lock reject filter a standard lowpass filter.
===========================================================================
You may have second resonance in your tuning or trap circuits that is
starting to overlap
with D2. Also it could be just due to the broad Q associated with the lock
circuitry itself.
I have run simultaneous F19 / H1 and side band techniques as close as 5
MHZ.If there
is bleed thru swamping the lock you may want to consider a cavity resonator
/ trap and filter
in the lock signal path between the t/r switch and the probe. Cavity
resonators have very high
Q's and have about 800 KHz 3dB bandwitdths at these frequencies
.( Wacom Inc ,Waco Texas)
=================================================================
regards, Andy
============================================================
Dr. A.J.Rous Phone +353-1-7062442
Chemistry Department, 7062133
University College, Fax +353-1-2837873
Belfield,
Dublin 4, E-mail andy@fiachra.ucd.ie
Irish Republic URL http://fiachra.ucd.ie/~andy/
============================================================