I have been attempting to simulate
a SATRAS spectrum, but for some reason
it seems that the intensity of the central
transition is much weaker than it should be.
Can someone tell me what's wrong?

ivan


spinsys {
  channels     13C 27Al
  nuclei       27Al
  quadrupole   1    2           0.800e6    0.200    0 0 0
  shift        1    -117.229p   -53.33p    0.075    0 0 0
}

par {
  proton_frequency  399.745605e6
  spin_rate         6000
  np                8192
  sw                804000
  crystal_file      zcw4180
  gamma_angles      sw/spin_rate
  method            direct
  start_operator    I1z
  detect_operator   I1p
  variable rf       5000
}

proc pulseq {} {
  global par

  maxdt 1.0

  set tdwell [expr 1.0e6/$par(sw)]
  pulse  [expr 1.0e6/(4*$par(rf))] 0 x $par(rf) x
  acq
  for {set i 1} {$i < $par(np)} {incr i} {
      delay $tdwell
      acq
  }
}

proc main {} {
  global par

  set f [fsimpson]
  fsave $f $par(name).fid
  funload $f
}




To unsubscribe from this group, send an email to:
simpson-simmol-unsubscribe@yahoogroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service .

-- *********************************************************
Thomas Vosegaard
Laboratory for Biomolecular NMR Spectroscopy
Department of Molecular and Structural Biology
University of Aarhus
Denmark
E-mail: tv@... http://nmr.imsb.au.dk/tv
Tel: +45 8942 3873 Fax: +45 8619 6199
*********************************************************

the 13C channel stuff was just something else
I was trying out...but, I have tried omitting it
and it comes out the same...

ivan


--- In simpson-simmol@y..., "jhbalto" <jay_baltisberger@b...> wrote:
> When I look at your pulse sequence, I am wondering why you
> have the 13C channel but don't use it?  I dont think it should
> matter, but perhaps it is causing something odd?  I wouldnt think
> it would.  Also, your RF power seems very low? Is this true?  And
> did the
> pulse  [expr 1.0e6/(4*$par(rf))] 0 x $par(rf) x
> line come out correct?  I guess the 0 x pulse is on the 13C?
> Beyond that, it looks like what you have should work?  The only
> explanation I could se for wrong intensity would be the low RF
> power?
> Jay
>
> --- In simpson-simmol@y..., ibinhung0 <no_reply@y...> wrote:
> > I have been attempting to simulate
> > a SATRAS spectrum, but for some reason
> > it seems that the intensity of the central
> > transition is much weaker than it should be.
> > Can someone tell me what's wrong?
> >
> > ivan
> >
> >
> > spinsys {
> >   channels     13C 27Al
> >   nuclei       27Al
> >   quadrupole   1    2           0.800e6    0.200    0 0 0
> >   shift        1    -117.229p   -53.33p    0.075    0 0 0
> > }
> >
> > par {
> >   proton_frequency  399.745605e6
> >   spin_rate         6000
> >   np                8192
> >   sw                804000
> >   crystal_file      zcw4180
> >   gamma_angles      sw/spin_rate
> >   method            direct
> >   start_operator    I1z
> >   detect_operator   I1p
> >   variable rf       5000
> > }
> >
> > proc pulseq {} {
> >   global par
> >
> >   maxdt 1.0
> >
> >   set tdwell [expr 1.0e6/$par(sw)]
> >   pulse  [expr 1.0e6/(4*$par(rf))] 0 x $par(rf) x
> >   acq
> >   for {set i 1} {$i < $par(np)} {incr i} {
> >       delay $tdwell
> >       acq
> >   }
> > }
> >
> > proc main {} {
> >   global par
> >
> >   set f [fsimpson]
> >   fsave $f $par(name).fid
> >   funload $f
> > }

When I look at your pulse sequence, I am wondering why you
have the 13C channel but don't use it?  I dont think it should
matter, but perhaps it is causing something odd?  I wouldnt think
it would.  Also, your RF power seems very low? Is this true?  And
did the
pulse  [expr 1.0e6/(4*$par(rf))] 0 x $par(rf) x
line come out correct?  I guess the 0 x pulse is on the 13C?
Beyond that, it looks like what you have should work?  The only
explanation I could se for wrong intensity would be the low RF
power?
Jay

--- In simpson-simmol@y..., ibinhung0 <no_reply@y...> wrote:
> I have been attempting to simulate
> a SATRAS spectrum, but for some reason
> it seems that the intensity of the central
> transition is much weaker than it should be.
> Can someone tell me what's wrong?
>
> ivan
>
>
> spinsys {
>   channels     13C 27Al
>   nuclei       27Al
>   quadrupole   1    2           0.800e6    0.200    0 0 0
>   shift        1    -117.229p   -53.33p    0.075    0 0 0
> }
>
> par {
>   proton_frequency  399.745605e6
>   spin_rate         6000
>   np                8192
>   sw                804000
>   crystal_file      zcw4180
>   gamma_angles      sw/spin_rate
>   method            direct
>   start_operator    I1z
>   detect_operator   I1p
>   variable rf       5000
> }
>
> proc pulseq {} {
>   global par
>
>   maxdt 1.0
>
>   set tdwell [expr 1.0e6/$par(sw)]
>   pulse  [expr 1.0e6/(4*$par(rf))] 0 x $par(rf) x
>   acq
>   for {set i 1} {$i < $par(np)} {incr i} {
>       delay $tdwell
>       acq
>   }
> }
>
> proc main {} {
>   global par
>
>   set f [fsimpson]
>   fsave $f $par(name).fid
>   funload $f
> }

I just tried rerunning that input file on my machines and I am not
sure why it isnt working for you.  If you have that file as something
like
      rbse.in
it should make an output file
      rbse.spe
I will repost what I am running below and you can try it again.

As per what is happening in the main section, I will add some
comments to the code as well, so see the code below.
# self explanatory spinsys => 87Rb with 15.0 MHz Cq, 0.42 eta,
20 ppm iso cs, -50ppm anisotropy, 0.6 cs eta

spinsys {
   nuclei 87Rb
   channels 87Rb
   shift 1 20 -50p 0.60 90 30 90
   quadrupole 1 2 15e6 0.42 0 0 0
}

# here we will start with Iz as the initial density matrix and detect
I-
# the spin rate is 0 (static sample), large crystal file for good
pattern
# rf is rf power level in Hz, giving the t90 as the central transition
selective
# 90 pulse.  off_max represents the maximum offset for the RF
pulse frequencies
# which are stepped through by num_off steps.  zero_fill will zero
fill by
# 2^3 times and gb_add will add a gaussian line broadening to
the echo
# before fft.  Note echo will appear at 300us as the echo_shift
suggests

par {
   start_operator   I1z
   detect_operator  I1m
   spin_rate        0
   gamma_angles     1
   sw               1800000
#  crystal_file     rep320
#  crystal_file     rep2000
#   crystal_file     zcw986
   crystal_file     zcw4180
   np               1024
   proton_frequency 300e6
   verbose          1001

   variable rf            50000.0
   variable echo_shift    300
   variable gb_add        8000
   variable zero_fill     3
   variable off_max       775000
   variable offs          -off_max
   variable num_off       30
   variable tsw           1.0e6/sw
   variable t90           0.125e6/rf/2.0
   variable t180          0.25e6/rf/2.0
   variable acq_time      tsw*np
}

proc pulseq {} {
   global par

   matrix set 51 coherence {1}
   matrix set 52 coherence {-1}

# here we apply a pulse with offset offs

   offset $par(offs)
   pulse $par(t90) $par(rf) x
   filter 52

# we use same offset for the inversion pulse
   delay $par(echo_shift)
   pulse $par(t180) $par(rf) y
   filter 51
#
#   this is acquisition of np points
#   note that offset is reset back to 0 for detection
#   in a spectrometer, this may not be what you actually do

   offset 0
   acq
   delay $par(tsw)
   store 1
   for {set n 1} {$n < $par(np)} {incr n} {
     acq
#  we may as well use a stored propogator since this is a static
sample.....
     prop 1
   }
}

proc main {} {
   global par

# the n_off2 is actually double the num_off parameter since this
will be going from
# positive to negative offsets...

   set n_off2 [expr 2*$par(num_off)]
   for {set offcount 0} {$offcount <= $n_off2} {incr offcount} {
# here we set the offset for each of the 2xnum_off experiments
     set par(offs) [expr $par(off_max)*($offcount -
$par(num_off))/$par(num_off)]
     puts "Offset is $par(offs)"
     if [expr $offcount>0] {
# if this is 2nd or more experimeent, run expt as $f and add it to
$fidsum
       set f [fsimpson]
# the gmftmc command does a gaussian multiplication,
followed by ft and then
# magnitude calculation.  By doing this approach we get purely
absorptive spectra
# since FT of an echo has no imaginary component once it is
phased properly with
# 1st and 0th order phase.  The advantage of doing this is
normally the sqrt(2) s/n
# you gain, but in simpson it guarantees we dont need to worry if
the pulse offsets
# create any phase errors in the spectra that would have to be
corrected individually
#
       gmftmc $f
       fsave $f $par(name).$offcount.spe
       fadd $fidsum $f
       funload $f
     } else {
# if this is 1st offset, just create $fidsum data set
       set fidsum [fsimpson]
       gmftmc $fidsum
       fsave $fidsum $par(name).$offcount.spe
     }
   }
   fsave $fidsum $par(name).spe
   funload $fidsum
   exit
}

proc gmftmc {fidsum} {
   global par

# in this part we add a gaussian multiplation to the echo top if it
is echo_shift'd
   if [expr $par(echo_shift)>0] {
     set e_top $par(echo_shift)
     for {set n 1} {$n <= $par(np)} {incr n} {
       set c  [findex $fidsum [expr $n]]
       set em_time [expr $n*$par(tsw)-$e_top]
       set em_exp [expr $em_time*$par(gb_add)/400000.0]
       set em_val [expr exp(-[power $em_exp 2])]
       set re [expr [lindex $c 0]*$em_val]
       set im [expr [lindex $c 1]*$em_val]
       fsetindex $fidsum $n $re $im
     }
   } else {
     faddlb $fidsum $par(gb_add) 1
   }

# here is the zerofilling to 2^zero_fill
   set np_fill [expr $par(np)*[power 2 $par(zero_fill)]]
   fzerofill $fidsum $np_fill

   fft $fidsum

# here we do the magnitude calculation on the spectrum
   if [expr $par(echo_shift)>0] {
     for {set n 1} {$n <= $np_fill} {incr n} {
       set c  [findex $fidsum [expr $n]]
       set re [lindex $c 0]
       set im [lindex $c 1]
       fsetindex $fidsum $n [expr sqrt(($re*$re)+($im*$im))] 0.0
     }
   }
}

proc power {base p} {
  set result 1
   while {$p > 0} {
     set result [expr $result * $base]
     set p [expr $p - 1]
    }
    return $result
}

I have been attempting to simulate
a SATRAS spectrum, but for some reason
it seems that the intensity of the central
transition is much weaker than it should be.
Can someone tell me what's wrong?

ivan


spinsys {
   channels     13C 27Al
   nuclei       27Al
   quadrupole   1    2           0.800e6    0.200    0 0 0
   shift        1    -117.229p   -53.33p    0.075    0 0 0
}

par {
   proton_frequency  399.745605e6
   spin_rate         6000
   np                8192
   sw                804000
   crystal_file      zcw4180
   gamma_angles      sw/spin_rate
   method            direct
   start_operator    I1z
   detect_operator   I1p
   variable rf       5000
}

proc pulseq {} {
   global par

   maxdt 1.0

   set tdwell [expr 1.0e6/$par(sw)]
   pulse  [expr 1.0e6/(4*$par(rf))] 0 x $par(rf) x
   acq
   for {set i 1} {$i < $par(np)} {incr i} {
       delay $tdwell
       acq
   }
}

proc main {} {
   global par

   set f [fsimpson]
   fsave $f $par(name).fid
   funload $f
}

do you have something simpler...I don't
quite seem to follow what you are doing
beginning in the 'main' section...& for
some reason when I ran your input file,
I didn't get an output file...

ivan


--- In simpson-simmol@y..., "jhbalto" <jay_baltisberger@b...> wrote:
> OK, see if this is what you want?
>
> source "/home/jhbaltis/simpson/global/power.in"
>
> spinsys {
>   nuclei 87Rb
>   channels 87Rb
>   shift 1 20 -50p 0.60 90 30 90
>   quadrupole 1 2 15e6 0.42 0 0 0
> }
>
> par {
>   start_operator   I1z
>   detect_operator  I1m
>   spin_rate        0
>   gamma_angles     1
>   sw               1800000
> #  crystal_file     rep320
> #  crystal_file     rep2000
> #   crystal_file     zcw986
>   crystal_file     zcw4180
>   np               1024
>   proton_frequency 300e6
>   verbose          1001
>
>   variable rf            20000.0
>   variable echo_shift    300
>   variable gb_add        8000
>   variable zero_fill     3
>   variable off_max       275000
>   variable offs          -off_max
>   variable num_off       30
>   variable tsw           1.0e6/sw
>   variable t90           0.125e6/rf/2.0
>   variable t180          0.25e6/rf/2.0
>   variable acq_time      tsw*np
> }
>
> proc pulseq {} {
>   global par
>
>   matrix set 51 coherence {1}
>   matrix set 52 coherence {-1}
>
>   offset $par(offs)
>   pulse $par(t90) $par(rf) x
>   filter 52
>
>   delay $par(echo_shift)
>   pulse $par(t180) $par(rf) y
>   filter 51
> #
> #   this is acquisition of np points
> #
>   offset 0
>   acq
>   delay $par(tsw)
>   store 1
>   for {set n 1} {$n < $par(np)} {incr n} {
>     acq
>     prop 1
>   }
> }
>
> proc main {} {
>   global par
>
>   set n_off2 [expr 2*$par(num_off)]
>   for {set offcount 0} {$offcount <= $n_off2} {incr offcount} {
>     set par(offs) [expr $par(off_max)*($offcount -
> $par(num_off))/$par(num_off)]
>     puts "Offset is $par(offs)"
>     if [expr $offcount>0] {
>       set f [fsimpson]
>       gmftmc $f
>       fadd $fidsum $f
>     } else {
>       set fidsum [fsimpson]
>       gmftmc $fidsum
>     }
>   }
>   fsave $fidsum $par(name).spe
>   exit
> }
>
> proc gmftmc {fidsum} {
>   global par
>
>   if [expr $par(echo_shift)>0] {
>     set e_top $par(echo_shift)
>     for {set n 1} {$n <= $par(np)} {incr n} {
>       set c  [findex $fidsum [expr $n]]
>       set em_time [expr $n*$par(tsw)-$e_top]
>       set em_exp [expr $em_time*$par(gb_add)/400000.0]
>       set em_val [expr exp(-[power $em_exp 2])]
>       set re [expr [lindex $c 0]*$em_val]
>       set im [expr [lindex $c 1]*$em_val]
>       fsetindex $fidsum $n $re $im
>     }
>   } else {
>     faddlb $fidsum $par(gb_add) 1
>   }
>
>   set np_fill [expr $par(np)*[power 2 $par(zero_fill)]]
>   fzerofill $fidsum $np_fill
>
>   fft $fidsum
>
>   if [expr $par(echo_shift)>0] {
>     for {set n 1} {$n <= $np_fill} {incr n} {
>       set c  [findex $fidsum [expr $n]]
>       set re [lindex $c 0]
>       set im [lindex $c 1]
>       fsetindex $fidsum $n [expr sqrt(($re*$re)+($im*$im))] 0.0
>     }
>   }
> }
>
> proc power {base p} {
>   set result 1
>   while {$p > 0} {
>     set result [expr $result * $base]
>     set p [expr $p - 1]
>   }
>   return $result
> }

And by analogy, if you set only the first order on, does that imply
that there will be no second, even if the coupling constant was
large enough that you should have a second order coupling?
Jay



To unsubscribe from this group, send an email to:
simpson-simmol-unsubscribe@yahoogroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service .

-- *********************************************************
Thomas Vosegaard
Laboratory for Biomolecular NMR Spectroscopy
Department of Molecular and Structural Biology
University of Aarhus
Denmark
E-mail: tv@... http://nmr.imsb.au.dk/tv
Tel: +45 8942 3873 Fax: +45 8619 6199
*********************************************************

following up my question, are these two filters equivalent?

#  matrix set 2 elements {{4 1}}
  matrix set 2 coherence {-3}
#  matrix set detect elements {{2 3}}
  matrix set detect coherence {1}

Jay



To unsubscribe from this group, send an email to:
simpson-simmol-unsubscribe@yahoogroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service .

-- *********************************************************
Thomas Vosegaard
Laboratory for Biomolecular NMR Spectroscopy
Department of Molecular and Structural Biology
University of Aarhus
Denmark
E-mail: tv@... http://nmr.imsb.au.dk/tv
Tel: +45 8942 3873 Fax: +45 8619 6199
*********************************************************

I have been looking at some input files other people have given
me that use both reset and filter in ways that I don't think I quite
understand.
What is the use of the added time delay in the reset command? 
I am confused as to what "time" it is even changing?  Is this so
you can have the rotor in a spinning experiment change position
but not change the propogator/density matrices?

Filter command confuses me when people use the matrix set 1
elements or matrix set 1 totalcoherence.  I think I under stand the
matrix set 1 coherence notation, but the other two are a bit odd? 
In one of the paper examples, the elements { 1 4 } and { 4 1 } are
distinguished for a 3q to 1q and -3q to 1q MQMAS expt?  Could
anyone give a more complete description of what these different
notations are doing in the matrix set and filter commands?
Jay



To unsubscribe from this group, send an email to:
simpson-simmol-unsubscribe@yahoogroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service .

-- *********************************************************
Thomas Vosegaard
Laboratory for Biomolecular NMR Spectroscopy
Department of Molecular and Structural Biology
University of Aarhus
Denmark
E-mail: tv@... http://nmr.imsb.au.dk/tv
Tel: +45 8942 3873 Fax: +45 8619 6199
*********************************************************

And by analogy, if you set only the first order on, does that imply
that there will be no second, even if the coupling constant was
large enough that you should have a second order coupling?
Jay

following up my question, are these two filters equivalent?

#  matrix set 2 elements {{4 1}}
   matrix set 2 coherence {-3}
#  matrix set detect elements {{2 3}}
   matrix set detect coherence {1}

Jay

I was looking at documentation in the JMR paper and wondering
about the "quadrupole" interaction.  You can have the 2nd
argument represent 1 or 2 for first or second order respectively. 
If you set the value to 2, this will give the second order
interaction.  Does this also imply that the first order interaction is
present as well, or do you need to turn it on separately?  I know it
would seem strange (impossible) to not have the first order
present if the second is present, but I suppose if you are ONLY
working on central transitions of half-odd integer spins, this
might be valid for some cases.
Jay


To unsubscribe from this group, send an email to:
simpson-simmol-unsubscribe@yahoogroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service .

-- *********************************************************
Thomas Vosegaard
Laboratory for Biomolecular NMR Spectroscopy
Department of Molecular and Structural Biology
University of Aarhus
Denmark
E-mail: tv@... http://nmr.imsb.au.dk/tv
Tel: +45 8942 3873 Fax: +45 8619 6199
*********************************************************

I have been looking at some input files other people have given
me that use both reset and filter in ways that I don't think I quite
understand.
What is the use of the added time delay in the reset command?
I am confused as to what "time" it is even changing?  Is this so
you can have the rotor in a spinning experiment change position
but not change the propogator/density matrices?

Filter command confuses me when people use the matrix set 1
elements or matrix set 1 totalcoherence.  I think I under stand the
matrix set 1 coherence notation, but the other two are a bit odd?
In one of the paper examples, the elements { 1 4 } and { 4 1 } are
distinguished for a 3q to 1q and -3q to 1q MQMAS expt?  Could
anyone give a more complete description of what these different
notations are doing in the matrix set and filter commands?
Jay

I was looking at documentation in the JMR paper and wondering
about the "quadrupole" interaction.  You can have the 2nd
argument represent 1 or 2 for first or second order respectively.
If you set the value to 2, this will give the second order
interaction.  Does this also imply that the first order interaction is
present as well, or do you need to turn it on separately?  I know it
would seem strange (impossible) to not have the first order
present if the second is present, but I suppose if you are ONLY
working on central transitions of half-odd integer spins, this
might be valid for some cases.
Jay

I was looking at documentation in the JMR paper and wondering
about the "quadrupole" interaction.  You can have the 2nd
argument represent 1 or 2 for first or second order respectively.
If you set the value to 2, this will give the second order
interaction.  Does this also imply that the first order interaction is
present as well, or do you need to turn it on separately?  I know it
would seem strange (impossible) to not have the first order
present if the second is present, but I suppose if you are ONLY
working on central transitions of half-odd integer spins, this
might be valid for some cases.
Jay

OK, more of my attempts to learn/use Simpson.
I am trying to simulate DAS experiment....

The thing is, my DAS echo shifts to the right in time in this
experiment when it should stay put in t2.  It is as if the t1
evolution is not working right?  Maybe the rotor angle cannot be
changed this way?  Any ideas?

spinsys {
  nuclei 87Rb
  channels 87Rb
# shift 1 10 -50p 0.60 90 30 90
  quadrupole 1 2 4e6 0.42 0 0 0
}

par {
     start_operator   I1z
     detect_operator  I1m
     spin_rate        27000
     variable tr      1.0e6/spin_rate
     variable nprop   4
     gamma_angles     4
     sw               nprop*spin_rate
     variable tsw     1.0e6/sw
     sw1              60000
     variable tsw1    1.0e6/sw1
     crystal_file     rep320
#    crystal_file     rep2000
#    crystal_file     zcw986
#    crystal_file     zcw4180
     np               256
     ni               256
     proton_frequency 300e6
     verbose          1101
     use_cluster      1
     cluster_port     3265

   variable rf            70000.0
   variable t90           0.125e6/rf
   variable t180          0.25e6/rf
   variable echo_shift    10
   variable zero_fill     3
   variable acq_time      tsw*np
   variable ta_min        tsw1*ni
   variable echo_time     tr*echo_shift
}

proc pulseq {} {
   global par

   maxdt 1
   matrix set 50 coherence {0}
   matrix set 51 coherence {1}
   matrix set 52 coherence {-1}

   set angle1 79.19
   set angle2 37.38

   for {set t1 1} {$t1 <= $par(ni)} {incr t1} {
     reset
     set tsw2 [expr $par(tsw1)*$t1/2.0]
     set par(rotor_angle) $angle2
     pulse $par(t90) $par(rf) x
     filter 52

     delay $tsw2

     pulse $par(t90) $par(rf) x
     filter 50

     set par(rotor_angle) $angle2

     pulse $par(t90) $par(rf) x

     if [expr $par(echo_shift) > 0 ] {
       filter 52
       delay $tsw2
       delay $par(echo_time)
       pulse $par(t180) $par(rf) -y
     }

     filter 51

#
#   this is acquisition of np points
#
     for {set n 1} {$n <= $par(np)} {incr n} {
       acq
       if [expr $n > $par(nprop)] {
         prop [expr (($n-1) % $par(nprop))+1]
       } else {
         delay $par(tsw)
         store $n
       }
     }
   }
}

proc main {} {
     global par

     set f [fsimpson]
     fsave $f $par(name).rmn -rmn
     exit
}

I am trying to simulate the QPASS experiment of Massiot.  I am
wondering why in this experiment below, the echo seems to shift
to higher t2 values as t1 is incremented?  Has anyone tried a
similar simulation for PASS?

source "/home/jhb/simpson/cluster.in"
source "/home/jhb/simpson/global/power.in"

spinsys {
  nuclei 87Rb
  channels 87Rb
# shift 1 10 -50p 0.60 90 30 90
  quadrupole 1 2 4e6 0.42 0 0 0
}

par {
     start_operator   I1z
     detect_operator  I1m
     spin_rate        8000
     variable tr      1.0e6/spin_rate
     variable nprop   15
     gamma_angles     4
     sw               nprop*spin_rate
     variable tsw     1.0e6/sw
     sw1              sw
     variable tsw1    1.0e6/sw1/10.0
     crystal_file     rep320
#    crystal_file     rep2000
#    crystal_file     zcw986
#    crystal_file     zcw4180
     np               16
     ni               512
     proton_frequency 300e6
     verbose          1101
     use_cluster      1
     cluster_port     3265

   variable rf            70000.0
   variable t90           0.125e6/rf
   variable t180          0.25e6/rf
   variable echo_shift    15
   variable zero_fill     3
   variable acq_time      tsw*ni
   variable echo_time     tr*echo_shift
}

proc pulseq {} {
   global par

   maxdt 1

set delays { \
   1 0.1000000   0.2000000   0.3000000   0.4000000   0.5000000
0.6000000   0.7000000   0.8000000   0.9000000 \
   2 0.1291799   0.2325585   0.3253607   0.4309482   0.5242304
0.6299537   0.7231375   0.8283638   0.9199156 \
   3 0.1635639   0.2632728   0.3471366   0.4568109   0.5426028
0.6529745   0.7385292   0.8472323   0.9284580 \
   4 0.1943716   0.2769544   0.3513224   0.4646120   0.5431853
0.6583618   0.7369162   0.8501135   0.9242462 \
   5 0.1961527   0.2422587   0.3285490   0.4513264   0.5250075
0.6477848   0.7212225   0.8423374   0.9127756 \
   6 0.1669983   0.2006810   0.3359459   0.4518658   0.5178214
0.6452848   0.7121974   0.8390408   0.9039095 \
   7 0.1517745   0.1901459   0.3559386   0.4426394   0.5039939
0.6386728   0.6999142   0.8339029   0.8937398 \
   8 0.1400747   0.1843472   0.3465835   0.3988912   0.4857753
0.6298344   0.6857642   0.8278361   0.8827112 \
   9 0.1270763   0.1760999   0.3239001   0.3729237   0.5000000
0.6270763   0.6760999   0.8239001   0.8729237 \
  10 0.1172888   0.1721639   0.3142358   0.3701656   0.5142247
0.6011088   0.6534165   0.8156528   0.8599253 \
  11 0.1062602   0.1660971   0.3000858   0.3613272   0.4960061
0.5573606   0.6440614   0.8098541   0.8482255 \
  12 0.0960905   0.1609592   0.2878026   0.3547152   0.4821786
0.5481342   0.6640541   0.7993190   0.8330017 \
  13 0.0872244   0.1576626   0.2787775   0.3522152   0.4749925
0.5486736   0.6714510   0.7577413   0.8038473 \
  14 0.0757538   0.1498865   0.2630838   0.3416382   0.4568147
0.5353880   0.6486776   0.7230456   0.8056284 \
  15 0.0715420   0.1527677   0.2614708   0.3470255   0.4573972
0.5431891   0.6528634   0.7367272   0.8364361 \
  16 0.0800844   0.1716362   0.2768625   0.3700463   0.4757696
0.5690518   0.6746393   0.7674415   0.8708201 \
}
foreach {del t1 t2 t3 t4 t5 t6 t7 t8 t9} $delays {
   set pipul($del,1) $t1
   set pipul($del,2) $t2
   set pipul($del,3) $t3
   set pipul($del,4) $t4
   set pipul($del,5) $t5
   set pipul($del,6) $t6
   set pipul($del,7) $t7
   set pipul($del,8) $t8
   set pipul($del,9) $t9
}

   matrix set 51 coherence {1}
   matrix set 52 coherence {-1}

   for {set t1 1} {$t1 <= $par(np)} {incr t1} {
     reset
     pulse $par(t90) $par(rf) x
     filter 52

     if [expr $par(echo_shift) > 0 ] {
       delay $par(echo_time)
     }

     delay [expr ( $pipul($t1,1)*$par(tr) ) - $par(t90)]

     for {set n 1} {$n <= 4} {incr n} {
       pulse $par(t180) $par(rf) y
       filter 51

       delay [expr ( ($pipul($t1,[expr 2*$n])-$pipul($t1,[expr
(2*$n)-1]))*$par(tr) ) - $par(t180)]

       pulse $par(t180) $par(rf) -y
       filter 52

       delay [expr ( ($pipul($t1,[expr (2*$n)+1])-$pipul($t1,[expr
2*$n]))*$par(tr) ) - $par(t180)]
     }

     pulse $par(t180) $par(rf) y
     filter 51

     delay [expr ((1.0-$pipul($t1,9))*$par(tr))-$par(t90)]

#
#   this is acquisition of ni points
#
     for {set n 1} {$n <= $par(ni)} {incr n} {
       acq
       if [expr $n > $par(nprop)] {
         prop [expr (($n-1) % $par(nprop))+1]
       } else {
         delay $par(tsw)
         store $n
       }
     }
   }
}

proc main {} {
   global par

   set f [fsimpson]
   fsave $f $par(name).rmn -rmn
   exit
}

I have been using the -rmn format of the new version of
simpson.  I think there is an error in this format as RMN seems
to read it in with the wrong number of points in the two
dimensions if you have ni not the same as np.  Also, I think the
dwell times in the two dimensions get switched as you read
them into RMN.  Finally, it seems the number of data points
causes a problem with the current version of RMN. I am curious
if the -rmn save mode is for RMN vesion 1.2.4?  If not, perhaps
we can get the format from Phil and be sure it is correct in the
simpson program?
Jay

OK, see if this is what you want?

source "/home/jhbaltis/simpson/global/power.in"

spinsys {
   nuclei 87Rb
   channels 87Rb
   shift 1 20 -50p 0.60 90 30 90
   quadrupole 1 2 15e6 0.42 0 0 0
}

par {
   start_operator   I1z
   detect_operator  I1m
   spin_rate        0
   gamma_angles     1
   sw               1800000
#  crystal_file     rep320
#  crystal_file     rep2000
#   crystal_file     zcw986
   crystal_file     zcw4180
   np               1024
   proton_frequency 300e6
   verbose          1001

   variable rf            20000.0
   variable echo_shift    300
   variable gb_add        8000
   variable zero_fill     3
   variable off_max       275000
   variable offs          -off_max
   variable num_off       30
   variable tsw           1.0e6/sw
   variable t90           0.125e6/rf/2.0
   variable t180          0.25e6/rf/2.0
   variable acq_time      tsw*np
}

proc pulseq {} {
   global par

   matrix set 51 coherence {1}
   matrix set 52 coherence {-1}

   offset $par(offs)
   pulse $par(t90) $par(rf) x
   filter 52

   delay $par(echo_shift)
   pulse $par(t180) $par(rf) y
   filter 51
#
#   this is acquisition of np points
#
   offset 0
   acq
   delay $par(tsw)
   store 1
   for {set n 1} {$n < $par(np)} {incr n} {
     acq
     prop 1
   }
}

proc main {} {
   global par

   set n_off2 [expr 2*$par(num_off)]
   for {set offcount 0} {$offcount <= $n_off2} {incr offcount} {
     set par(offs) [expr $par(off_max)*($offcount -
$par(num_off))/$par(num_off)]
     puts "Offset is $par(offs)"
     if [expr $offcount>0] {
       set f [fsimpson]
       gmftmc $f
       fadd $fidsum $f
     } else {
       set fidsum [fsimpson]
       gmftmc $fidsum
     }
   }
   fsave $fidsum $par(name).spe
   exit
}

proc gmftmc {fidsum} {
   global par

   if [expr $par(echo_shift)>0] {
     set e_top $par(echo_shift)
     for {set n 1} {$n <= $par(np)} {incr n} {
       set c  [findex $fidsum [expr $n]]
       set em_time [expr $n*$par(tsw)-$e_top]
       set em_exp [expr $em_time*$par(gb_add)/400000.0]
       set em_val [expr exp(-[power $em_exp 2])]
       set re [expr [lindex $c 0]*$em_val]
       set im [expr [lindex $c 1]*$em_val]
       fsetindex $fidsum $n $re $im
     }
   } else {
     faddlb $fidsum $par(gb_add) 1
   }

   set np_fill [expr $par(np)*[power 2 $par(zero_fill)]]
   fzerofill $fidsum $np_fill

   fft $fidsum

   if [expr $par(echo_shift)>0] {
     for {set n 1} {$n <= $np_fill} {incr n} {
       set c  [findex $fidsum [expr $n]]
       set re [lindex $c 0]
       set im [lindex $c 1]
       fsetindex $fidsum $n [expr sqrt(($re*$re)+($im*$im))] 0.0
     }
   }
}

proc power {base p} {
   set result 1
   while {$p > 0} {
     set result [expr $result * $base]
     set p [expr $p - 1]
   }
   return $result
}

I was thinking about this.  Is the simulation you want one where
you say do a
90 - tau - 180 - acq
to get the whole echo and then vary the offset and sum up a
series of these?  I think the OFFSET command will do this (vary
the offset of the pulses but not the reciever).  I haven't tried this
ever, but once I get a chance, I will see if I can toss together a
quick example and put it here.
Jay

I tested the 90 time just as you would on the spectrometer.
Pulse and acquire. As expected, the RF power level is the actual,
not the effective RF and thus you scale 90 time by (I + 1/2) if you
want central transition pulses.  I sort of figured this would be the
case, but its nice to have a simulation work and know that theory
matches experiment!
Jay

Hi! I have been trying to simulate a very
wide static powder pattern in parts and then
sum the parts up to get the full pattern, but
I cannot seem to be able to just excite pieces
of the whole pattern.  Can anyone help me out?

Ivan

-- *********************************************************
Thomas Vosegaard
Laboratory for Biomolecular NMR Spectroscopy
Department of Molecular and Structural Biology
University of Aarhus
Denmark
E-mail: tv@... http://nmr.imsb.au.dk/tv
Tel: +45 8942 3873 Fax: +45 8619 6199
*********************************************************

If you use the filter command, can you still store the propogator?
Jay

New version that seems to be working:

source "/home/jhb/simpson/cluster.in"

spinsys {
nuclei 87Rb
channels 87Rb
shift 1 0 -50p 0.60 90 30 90
quadrupole 1 2 5e6 0.42 0 0 0
}

par {
    start_operator   I1m
    detect_operator  I1c
    spin_rate        12000
    sw               250000
#    crystal_file     rep320
    crystal_file     zcw4180
    np               512
    proton_frequency 300e6
    verbose          1001
    use_cluster      1
    cluster_port     3265

  variable rf        64102.6
  variable zero_fill 3
  variable tsw       1.0e6/sw/6.0
  variable tr        1.0e6/spin_rate
  variable t90       0.25e6/rf/2.0
  variable t180      0.5e6/rf/2.0
  variable ta        (tr/3.0)-tsw-2.0*t180
}

proc pulseq {} {
  global par

  maxdt 1

#       net effect here is to have 6 periods of tsw with coherence +1
#       along with 3 periods of ta with coherence -1 and 3 periods
of
#       ta with coherence of +1 to cancel out one another giving a
#       net evoluation where you have t1 evolving at 1/3 rotor
positions

  reset
#  pulse $par(t90) $par(rf) x
#  acq

  delay $par(ta)
  pulse $par(t180) $par(rf) -y
  delay $par(tr)
  delay $par(tsw)
  pulse $par(t180) $par(rf) y
  delay $par(ta)
  pulse $par(t180) $par(rf) -y
  delay $par(tsw)
  pulse $par(t180) $par(rf) y
  delay $par(ta)
  pulse $par(t180) $par(rf) -y
  delay $par(tsw)
  pulse $par(t180) $par(rf) y
  store 1

  acq [expr $par(np)] 1
}

proc main {} {
    global par

    set f [fsimpson]
    fsave $f $par(name).fid
    fsave $f $par(name).rmn -rmn

    faddlb $f 300 1

    set np_fill [expr $par(np)*[power 2 $par(zero_fill)]]
    fzerofill $f $np_fill

    fft $f
    fextract $f -30000 30000
    fsave $f $par(name).spe
    exit
}

The only real problem now is that I get a large 0 frequencey
peak that hides (I think) the proper result?  Is phase cycling
possible in SIMPSON?
Jay (jhbalto@...)



To unsubscribe from this group, send an email to:
simpson-simmol-unsubscribe@yahoogroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service .

-- *********************************************************
Thomas Vosegaard
Laboratory for Biomolecular NMR Spectroscopy
Department of Molecular and Structural Biology
University of Aarhus
Denmark
E-mail: tv@... http://nmr.imsb.au.dk/tv
Tel: +45 8942 3873 Fax: +45 8619 6199
*********************************************************

Hi,

There is a few things in your attached code which explain the error
message from the socket.

First of all, the problem is not related to the socket or clustering. It
is a timing problem in your pulse sequence.

Under MAS the propagator may only be re-used at equal times in a rotor
period. Consequently, you must ensure that the the length of your
propagator 1 is a multiple of the rotor period. This is not the case in
the attached file.

In this context it is also worth to notice that pulseid - although it
takes the time as argument - does not change the internal time of the
pulse sequence, i.e., its duration is 0.

Thomas

jhbalto wrote:

> I was trying to learn to use Simpson better.  I have been trying to
> build a basic MAT sequence using pi pulses.  The code I have
> looks like this:
>
> source "/home/jhb/simpson/cluster.in"
>
> spinsys {
> nuclei 87Rb
> channels 87Rb
> shift 1 0 -50p 0.60 90 30 90
> quadrupole 1 2 5e6 0.42 0 0 0
> }
>
> par {
> #    rotor_angle      63.15
>     start_operator   I1p
>     detect_operator  I1c
>     spin_rate        12000
>     gamma_angles     10
>     sw               gamma_angles*spin_rate
>     crystal_file     rep320
> #    crystal_file     zcw4180
>     np               1024
>     proton_frequency 300e6
>     verbose          1001
>     use_cluster      $cluster
>     cluster_port     3265
>
>   variable rf        64102.6
>   variable zero_fill 3
>   variable tsw       1.0e6/sw/6.0
>   variable tr        1.0e6/spin_rate
>   variable t180      0.5e6/rf/2.0
>   variable t1        (tr/3.0)-tsw-t180-t180
> }
>
> proc pulseq {} {
>   global par
>
>   maxdt 0.4
>
>   delay $par(tr)
>   delay $par(tsw)
>   pulseid $par(t180) $par(rf) y
>   delay $par(t1)
>   pulseid $par(t180) $par(rf) y
>   delay $par(tsw)
>   pulseid $par(t180) $par(rf) y
>   delay $par(t1)
>   pulseid $par(t180) $par(rf) y
>   delay $par(tsw)
>   pulseid $par(t180) $par(rf) y
>   store 1
>
>   for {set n 1} {$n <= $par(np)} {incr n} {
>     acq
>     prop 1
>   }
> }
>
> proc main {} {
>     global par
>
>     set np_fill [expr $par(np)*[power 2 $par(zero_fill)]]
>
>     set f [fsimpson]
>     fsave $f $par(name).fid
>     fsave $f $par(name).rmn -rmn
>     fzerofill $f $np_fill
>     fft $f
>     fextract $f -30000 30000
>     fsave $f $par(name).spe
>     exit
> }
>
> Is there something fundamentally wrong with this idea?  I
> wanted to calculate the propogator for a two rotor period MAT
> sequence and then propogate this multiple times to create a
> large number of t1 points.  I get standard output until it connects
> to my cluster and then it stops with:
> Success
> error: received garbage from socket: "error: a propagator was
> calculated at time 0usec
> relative to the"
> error: cannot read from socket
>
> Thanks in advance for any help. BTW, my email is not working
> quite right and if you need to send me anything, you can at
> jhbalto@...
>

• Combination of SIMPSON with the MINUIT optimization tools from CERN.
• Import/export from/to NMRPipe and RMN file formats for one- and multi-dimensional data processing and visualization.

I am trying to use FEXTRACT to pull out a small portion of a
spectrum in the form:
   fextract $f reduced_sw_lf reduced_sw_hf
and it gives me error :
   fextract: argument 2 must be double <frq-from> 
Is there way to put variables into the fextract?
Jay



To unsubscribe from this group, send an email to:
simpson-simmol-unsubscribe@yahoogroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service .

-- *********************************************************
Thomas Vosegaard
Laboratory for Biomolecular NMR Spectroscopy
Department of Molecular and Structural Biology
University of Aarhus
Denmark
E-mail: tv@... http://nmr.imsb.au.dk/tv
Tel: +45 8942 3873 Fax: +45 8619 6199
*********************************************************

cluster {

  computer1.domain

  computer2.domain

}

and then, in your simpson input files use the command
 

source "/path/to/cluster.in"

where you otherwise specified the cluster info.

Thomas
 

jhbalto wrote:

 OK, I just thought I'd chime in and say hello.
The current question I am wondering about was if there are any
global parameter type files where one could set the cluster
settings for a given machine so it runs them for all input files?
Jay

The SIMPSON - SIMMOL Discussion Forum is the place where SIMPSON and
SIMMOL users may exchange ideas and experiences.

If you use SIMPSON or SIMMOL on a regular basis you may subscribe to
this forum and receive email information regarding updates etc.

Enjoy

Niels Chr. Nielsen and Thomas Vosegaard