We have been having some problems with our AMX500 that I would like
to share with the group to see if anyone is having similar problems
or can offer some advice.
We were noticing some severe baseline problems on X nuclei, so I decided
to have a close look at the beginning of the FID. Right at the beginning
there is what looked like intense acoustic ringing going on for 180 usec
or so. (29-Si at 99 Mhz, SW 20 kHz, simple zg pulse sequence)
I dug out some old files
and old data with the same parameters did not show this effect. I removed
the probe from the magnet and tried again - still there. That rules out
any sort of NMR effect or acoustic ringing. I then replaced the probe with
a 50 Ohm load and the effect went away. I tried several other probes
and saw the same effect on each of them, as long as they were tuned
to resonance. So, the effect is somehow related to the RF ringdown
of the probe, but NO probe has a high enough Q to ring for 180 usec!
especially not at these fairly high X frequencies. I then tried
a proton probe and this seemed OK, but ringdown at 500 MHz is generally
much faster, and the probe is excited with much less power.
I then started hooking the scope up to the various test BNC connectors
on the SE451: One channel on the receiver output, one on SPFX (pulse
gate) and a third on EP which is the receiver blanking pulse.
I triggered on the leading edge of the pulse gate. I started a GS and saw
the following:
- Pulse gate looked as expected, P1 usec in length.
- The receiver unblanked (opened) approx, 1.6 usec after the trailing edge
of the pulse.
- The receiver output went wild and clamped at 15 V for ca 3 sec and then
rang for another 5 usec or so.
The DE (time from pulse to first data point) was 37 usec, so the receiver
should have been settled by this time, but I was looking at the signal
AHEAD of the analogue filters. What was this huge receiver outout
pulse doing to the filters? I opened the filter width to 100 kHz
and the effect (post filter) went away - along with the signal to noise,
of course! So, the output pulse is causing the filters to ring!
But WHY am I getting this receiver output pulse? Why is the receiver
opening at 1.6 usec after the pulse when the first data point isn't for
37 usec.?
I then looked into the UXNMR manual under "DE" and it said that "DE"
is composed of two equal periods, DE1 and DE2, and the receiver is opened
after the first one (ie. half way through DE). In my case this would
be about 18 usec, NOT 1.6! No matter what I set DE to the delay between
the end of the pulse and receiver turn-on remained fixed at 1.6 usec.
>From the large frequency dependent phase shifts that occured at large DE
I confirmed that DE was setting the time between pulse and first data
point.
I then simulated a longer DE1, by putting a fixed 5 usec delay between
P1 and GO. As expected, the receiver opened 6.6 usec after the pulse,
the output pulse from the receiver was pretty much gone on my baselines
were flat again. I guess a temporary fix would be to edit all the
pulse sequences and add a fixed delay between any pulse and GO.
ARRRGHHHHHH!!!!!!!!!!!!!! This shouldn't be necessary, the system
should work properly!
I also experimented with commands like P3:e and the proper receiver
pulses were generated, so at least that part of the hardware seems
to be working.
Although I am not certain of this, the problem seemed to develop about the
time that we upgraded software from 930601.1 to 930601.3 (or perhaps
when we upgraded to 930601.1)
So, I have two hypotheses:
1. There is a hardware problem in whatever timer generates the DE
delays. I tend to doubt this as I don't THINK there is a dedicated
DE timer.
2. This is a software problem, and the DE1 (or whatever) value is not
being calculated or set properly. Is this possibly
related to the development of the DMX where analogue filters are not
used and presumably the receiver can open earlier without problems?
(I'm really guessing at that one!)
Anyone out there have any ideas?
Would anyone be willing to put a scope on their SE451 and look
at the time between the end SPFX pulse and start of the EP pulse
when GSing on an X nucleus? These signals are available on convenient
BNC jacks on the SE451.
__
_________| |____________ SPFX
________
________________| EP
----> <----- this time
Thanks in advance!
-- Kirk Marat NMR Facility Manager Dept of Chemistry / Prairie Regional NMR Centre University of Manitoba, Winnipeg, MB, CANADA, R3T 2N2.. Ciminal Lawyer (n): redundant term, see WEASEL