Hi Stevie
Instead of using computer power on optimizing parameters for the D2O peak, I would either only fit the region of the spectrum not influenced by this peak. You can do this by specifying optional regions to the frms command (syntax would be something like frms $f $g {0 10000} {12000 14000}).
If you have important features under the D2O peak so the above idea doesn't work, then I would suggest you only optimize the linewidth of the D2O peak (if you cannot estimate a fixed value for it) and then handle the relative scaling of the two spectra using the approach I sketched in a mail to the simpson-simmol user group in June 2002 (Subject: Re: [simpson-simmol] overlapping lineshapes using simpson). This procedure avoids using minuit parameters to fit it - so it's less prone to falling into wrong minima.
The maxdt parameter defines the maximum time over which the Hamiltonian is considered constant in the case of MAS. This matters for the speed of the calculation if you have pulses. With maxdt 1 you will split the pulse into 1us steps and diagonalize for each step.
When you have a pulse sequence which only consists of "delay 9999", it is probably using the method gcompute and a start operator of I1x. So it starts with x coherence on the nucleus you want to detect, so no need for pulses. gcompute is a fast method to propagate when you have a rotor-synchronized pulse sequence - and a free-induction decay sampled with a spectral width of n*spin_rate is rotor synchronized. In this case the pulse sequence shouldn't contain any information on the sampling - this is specified by the spectral width and spin rate (and gamma_angles, which needs to fulfill the requirement sw = spin_rate*gamma_angles). The pulse sequence needs only contain the information on the first rotor period.
Thomas
On Sep 9, 2008, at 22:43 , Stephen Greenwald wrote:
Hi guys,
I'm a relatively new SIMPSON user. I've managed to figure quite a bit
out on my own, with much struggling, but there is something I cannot
seem to tackle. what is the best way to handle broadening of peaks
caused by something other than the four spinsys interactions?
For example, I have fit a static powder pattern of fully deuterated
oxalic acid-dihydrate. The spectrum is a clean pake doublet (the
oxalic acid deuterons) with a broad D2O peak in the middle. I could
not, for the life of me, figure out how to simulate the D2O peak and
so ended up zeroing out that portion of the spectrum and fitting only
the pake portion.
I tried to use faddpeak, but the program would not compile with
variables inside the double curly bracket faddpeak takes asw an
argument (ie "faddpeaks $f 1e-5 {{0 $mn(int) $mn(lb) 1}}" was not
proper syntax).
So what is the bet way to handle broadening? "faddpeak" might do the
job in some instances but it would fail for any spectra with
interacting spins. What if I want to generate a spectrum of two peaks
of different broadness that also had jcoupling or dipole interactions?
One other question: I have seen a few examples whose pulseseq section
contains only "maxdt 1" and "delay 9999". What does this signify?
How is a spectrum generated without any pulses?
Thanks
Stevie
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
Offline 
Send Email