The spin magnetic tunnel junction cant get converged

[qiuminz]

Dear  Quantumwise staffs:
  I  calculated  a magnetic tunnel junction  with exchange correlation SGGA-PBE ,but  I cant get  converged.  the best  result  is  dE and dH can reduce to  10-3.
I  have  tried all  methods  in  the convergence  guide but failed ,   and  all  pseudo potential  are tested  ,but it  does not work , could  you  please  help me ?  thank  you 
  my log file is over 400k  I cant post it

[Ulrik G. Vej-Hansen]

Please see also the new document about convergence of this kind of system: https://docs.quantumatk.com/technicalnotes/spin_polarized_negf_convergence_guide/spin_polarized_negf_convergence_guide.html

[qiuminz]

I have tried  the method in this guide,but the results are worse than before .I dont know why the bad concergence happen.

[Petr Khomyakov]

Did you also look at these guidelines https://docs.quantumatk.com/technicalnotes/negf_convergence_guide/negf_convergence_guide.html#negf-convergence-guide?

[qiuminz]

yes  i did  ,

[Petr Khomyakov]

Try to increase the central region size of your device, see the script enclosed.

[qiuminz]

the script doesnt work  ,the result is similar to before . thank  you anywhere

[Anders Blom]

I would try to set the scaled spins closer to their expected values based on a bulk calculations

[qiuminz]

Thank you for your reply