Page 1 of 1

Esfit - Warning: Rank Deficient

Posted: Wed Sep 04, 2019 12:13 pm
by smbetts
Hello,
After the EasySpin Academy, I have been working on trying to fit my spin-labeled protein spectra, and am having issues. I need to fit with 2 components for a nitroxide spin label and this is the code I have:

Code: Select all

clear,clc,clf

[B,spc] = eprload('M9_Only_S184R1a_pR_190806');
Spc = spc{1}; % to pick out only the derivative spectrum
B = B/10; %G to mT

Exp.mwFreq = 9.830268; %GHz
Exp.ModAmp = 2/10; %mT
Exp.nPoints = 2048;
Exp.Temperature = 298; %K
Exp.CenterSweep = [3511/10 150/10]; %mT
Exp.CrystalOrientation = [];

Sys.S = 1/2;
Sys.Nucs = '14N';

%mobile component
Sys1 = Sys;
Az_m = mt2mhz((-349.2+353.02)/2); %MHz
Sys1.A = [Az_m/20 Az_m/5 Az_m+60]; %MHz
Sys1.logtcorr = -8.7; 
Sys1.Potential = [2 0 0 0.2];
Sys1.lwpp = 0.6;
[B2,spc2] = chili(Sys1,Exp);

%immobile component
Sys2 = Sys;
Az_i = mt2mhz((-348.27+354.96)/2); %mT -> MHz
Sys2.A = [Az_i/5 Az_i/5 Az_i+15]; %mT -> MHz
Sys2.logtcorr = -8.2218; %6 ns
Sys2.Potential = [2 0 0 2.5];
Sys2.lwpp = 0.5;
[B3,spc3] = chili(Sys2,Exp);

Sys1.weight = 0.5
Sys2.weight = 1-Sys1.weight

Var.g = [0.005 0.005 0.005];
Var.Potential = [0 0 0 2];
Var.weight = 0.5
Var.logtcorr = 0.5
Var1 = Var;
Var2 = Var;
Var1.A = [Az_m/5 Az_m/5 Az_m/5];
Var2.A = [Az_i/5 Az_i/5 Az_i/5];

esfit('chili',Spc,{Sys1,Sys2},{Var1,Var2},Exp);
Upon running this and choosing any of the fitting algorithms, the following appears repeatedly in the command line, while no fitting occurs (no matter how few parameters I tell it to fit):
Warning: Rank deficient, rank = 1, tol = 2.057952e-11.
> In rescale
In esfit
In esfit_genetic
In esfit

Are there any issues with my code that I am unaware of that is causing this to happen? I attached the DSC file so that this code can be ran if necessary.
Thanks ahead of time!
-Sam

Re: Esfit - Warning: Rank Deficient

Posted: Thu Sep 05, 2019 2:17 pm
by Stefan Stoll
To reproduce the problem, both the DSC and the DTA file are needed,

Try to select maxabs scaling in esfit.

Re: Esfit - Warning: Rank Deficient

Posted: Fri Sep 06, 2019 6:21 am
by smbetts
The maxabs scaling worked. The only issue now is that the fit spectrum seems like it is going to be shifted in field from my experimental.
esfit_S184R1a_pR_test.png
esfit_S184R1a_pR_test.png (247.18 KiB) Viewed 2878 times

Will this effect the parameters I obtain from the fit? If not, I'm all set, but I attach both data files in the case there is still an issue.
Thank you very much!

Re: Esfit - Warning: Rank Deficient

Posted: Tue Oct 08, 2019 4:16 pm
by Stefan Stoll
Field shifts occur if (1) you don't know the field axis exactly (which is often the case), (2) you don't know the microwave frequency exactly (which is sometimes the case), or/and (3) you don't know the g values exactly and you don't fit it.