--
Kind regards Nick Papior On 11 May 2016 11:44, "Seyed Mohammad Tabatabaei" <[email protected]> wrote: > > I have constrained the atoms corresponding to the electrodes and buffers in a siesta calculation I performed my transiesta calculation (to obtain the .EIG file and also to relax the scattering region) and just relaxed the scattering region without relaxing electrode and buffer atoms. I expected that would fix their positions. Here is the output file for the transiesta calculation in which I have fixed atoms from 1 to 128 and -1 to -128 which correspond to left and right electrodes and buffers. I am wondering if this has changed some constrained positions by 0.0001 Angstroms or so because I just copied the output relaxed position from the attached file for my transiesta calculation. Did you remove all files before re-running. > > I have also modified the .F95 file you mentioned and recompiled and its now giving me this error: > > "Error in GFfile: scat.TSGFL. Please move or delete" I think there is no ambiguity in this error message. > > On Wed, May 11, 2016 at 12:15 PM, Nick Papior <[email protected]> wrote: >> >> -- >> >> Kind regards Nick Papior >> On 11 May 2016 08:24, "Seyed Mohammad Tabatabaei" <[email protected]> wrote: >> > >> > Thank you very much for your help. >> > >> > Do I have change this and recompile transiesta? >> Yes. >> > >> > Also I have used 0 for MD.NumCGsteps and the positions were not updated by any means in the program. Is that a numerical issue in siesta? >> No, that is correct, would you expect otherwise? >> >> > >> > Bests, >> > >> > On Wed, May 11, 2016 at 10:11 AM, Nick Papior <[email protected]> wrote: >> >> >> >> You may change this line >> >> real(dp), parameter :: EPS = 1.0d-4 >> >> In m_ts_electrode.F90 to a lower value, say, 1e-3 or so. >> >> As you can see from your output there are some differences in the coordinates, although small. >> >> >> >> -- >> >> >> >> Kind regards Nick Papior >> >> >> >> On 11 May 2016 07:13, "Seyed Mohammad Tabatabaei" <[email protected]> wrote: >> >>> >> >>> This is also the POSITIONS.fdf input file for my left electrode. >> >>> >> >>> On Wed, May 11, 2016 at 9:25 AM, Seyed Mohammad Tabatabaei < [email protected]> wrote: >> >>>> >> >>>> Thank you for your time. >> >>>> >> >>>> The older version is "siesta-3.2-pl-5". >> >>>> >> >>>> I have checked my coordinates several times and I think they are currently "equivalent with respect to a common translational vector." so I would be really grateful if you explain to me the cause of this error. >> >>>> >> >>>> The inputs to both versions are identical so I only attach the output files. I have also attached the output file for left electrode calculation. >> >>>> Bests, >> >>>> Mohammad, >> >>>> >> >>>> On Tue, May 10, 2016 at 11:26 PM, Nick Papior <[email protected]> wrote: >> >>>>> >> >>>>> Dear Mohammad, >> >>>>> >> >>>>> 2016-05-10 17:46 GMT+02:00 Seyed Mohammad Tabatabaei < [email protected]>: >> >>>>>> >> >>>>>> Dear all, >> >>>>>> >> >>>>>> I have the two latest versions of transiesta installed on my computer. >> >>>>>> The older version does not print any errors for my electrode positions >> >>>>>> but the newer version (siesta-4.0b-485) prints this error for a 100 >> >>>>>> percent identical calculation: "The electrodes are not situated in the >> >>>>>> same coordinates. Please correct." and stops. >> >>>>> >> >>>>> Which version is the old version? >> >>>>>> >> >>>>>> >> >>>>>> I have read a similar thread about this issue and the suggestions >> >>>>>> therein but I think the newer version of transiesta falls short of >> >>>>>> understanding the fact that the positions of the electrodes may not be >> >>>>>> identical (same/equal) but they can be equivalent. Particularly, this >> >>>>>> sentence can be found in the manual of both versions: >> >>>>>> ".... Here, equivalent means that they can be made equal by a simple >> >>>>>> translation in space." >> >>>>> >> >>>>> Yes, this sentence is true. The electrode coordinates need not be EXACTLY the same. They only need to be equivalent with respect to a common translational vector. >> >>>>>> >> >>>>>> >> >>>>>> >> >>>>>> >> >>>>>> I would be really thankful to anyone who helps me understand the issue. >> >>>>> >> >>>>> We need more information, input/output of both the old and new version. >> >>>>> >> >>>>>> >> >>>>>> Bests, >> >>>>>> Mohammad, >> >>>>> >> >>>>> >> >>>>> >> >>>>> >> >>>>> -- >> >>>>> Kind regards Nick >> >>>> >> >>>> >> >>> >> > > >
