First, I'd like to thank all those that took time to respond to my last
outburst. The answers ranged from "It can't be done" to "It should be easy",
and everywhere in between (including the expected "Why don't you just use
an MSL?").
So I've spent several late nights and a weekend or two (i.e. not married)
trying some of your suggestions that were sent my way concerning the use
external address advance to obtain a CPMG T2 decay by sampling points on
the tops of echoes.
I should remind you that I'm working on an AC spectrometer, with ASPECT 3000 and
a process controller.
I'm sorry for making this a long post, but I think it's important that the
whole boring story be told...
The obvious first question... does EXTAD.AUR work? YES! And works well.
Consider...
;EXTAD.AUR
1 ZE
2 D1 ;relaxation delay
3 P1 PH1 ;90 degree pulse
4 D5 ;D5=DE/2
5 D7 ADC ;initialize ADC with D7=2 usec
6 D5 PH2 ;set detector phase, D5=DE/2
7 P8:X ;take a data point, P8 = 1 usec
8 D8 ;D8 is equivalent to dwell time
9 L1 TO 9 TIMES C ;loop for TD data points in the FID
10 RCYC=2 PH3 ;loop to line 2 for NS scans
11 EXIT
PH1=0 0 2 2
PH2=0 0 0 0 3 3 3 3
PH3=R0 R0 R2 R2 R1 R1 R3 R3
Now, this sequence works quite well, with D8 set as low as 8 usec. Comparing
FID's and spectra from this automation and the normal proton acquisition,
there is no noticable difference in the data. Some have suggested that
there is some overhead in the looping command in line #9, but I see no
evidence for it.
Now, moving from this automation to the CPMG automation _should_ be relatively
trivial. You've guessed by now that it isn't.
The closest to success that I have been is with the following:
;T2CPMGXA.AUR
1 ZE
2 D1 ;Relaxation delay
3 P1 PH1 ;90 degree pulse
4 D5 ;D5 = DE /2
5 D7 ADC ;Initialize ADC, D7=2 usec
6 D5 PH2 ;Set detector phase, D5=DE/2
7 P8:X ;Take the first data point on CPMG, after
; DE+D7 usec of dephasing
8 EOSC ;This MUST be included to get the 180 pulse train
; to fire
9 D0 ;Ideally, D0 = D2, the tau spacing, but it is not to be
10 P2:E:T PH3 ;Refocussing pulse
11 D2 ;Refocussing delay
12 D2 ;Allow magnetization to dephase
13 P2:E:T PH4 ;Second refocussing pulse
14 D3 ;D3=D2-(2*D5+D7), so that refocussing is still occurring
; when we are preparing for our next data point.
15 L1 TO 4 TIMES UPR ;Loop for next echo, next data point
16 D20 ;Short delay
17 RCYC=2 PH5 ;Loop for NS scans
18 EXIT
;A simple phase program for test purposes
PH1=0
PH2=2
PH3=1
PH4=3
PH5=R0
If line #8 EOSC is not included, the 180 pulses are not generated. BUT
the EOSC statement takes 2 msec to execute. So, if you have a long T2,
and wish to use this with echo spacings of 2 msec or longer, it works
fine as long as you compensate for this 2 msec by making D0=D2-2 msec.
Re-writing the CPMG sequence in a more explicit fashion (i.e. I have to
draw pictures to envisage what's happening...) might give something like:
;T2CPMGXB.AUR
1 ZE
2 D1 ;Relaxation delay
3 P1 PH1 ;90 degree pulse
4 D5 ;D5=DE/2
5 D7 ADC ;Initialize the ADC, D7=2 usec
6 D5 PH0 ;Set the detector phase, D5=DE/2
7 EOSC ;Necessary for 180 pulses to be generated
8 (D2) (D3 P2 PH2):T:E ;D2 is echo spacing, D3=D2-(D7+2*D5)
9 (D2) (D3 P8):X ;Take a point on the top of the echo
10 L1 TO 8 TIMES UPR ;Hardware loop for L1 echoes
11 RCYC=2 PH3 ;Loop for NS scans
12 EXIT
PH0=2
PH1=0
PH2=1
PH3=R0
Now, this may seem odd, but if line #7 EOSC is not included in the sequence, the
180 degree pulse train does not get generated (don't ask me why). If we leave it
in, the pulse train gets generated, with great timing down to D2=50 usec (which
is pretty good for an AC). But, of course, the fatal flaw is that no data
is collected. (Yes, the ADC is lit...)
Even if we try to include an initialization of ADC and phase setting within the
echo loops, by eliminating lines #5 and #6 and replacing lines #8 and #9 with:
.
.
8 (D2 ADC) (D3 P2 PH2):T:E
9 (D2 PH0) (D3 P8):X
.
.
... we retain the spacing of the pi pulses, but still no data is acquired.
I think I've gone on long enough. Maybe I'm optimistic in saying that I think
we're getting warm here (or maybe I'm just a sucker for punishment).
Any more ideas out there?
Again, thanks to those who took the time to reply to my last post.
Cheers,
Tim Bremner
===============================================================================
Z Y' : DR. TIM BREMNER, Inst. for Polymer Research |
| / : :: Department of Chemistry, U. of Waterloo |
|/ : : : :: Waterloo, Ontario, CANADA, N2L 3G1 |
-----o====>X' : : : : : ::::::::::::::::::::::::::::::::::::::::|
/| : : : : :: Ph: 519-885-1211 x2510 |
/ | : : :: Fax: 519-888-6179 |
/ | :: E-mail: tbremner@polymer.uwaterloo.ca |
===============================================================================