Dear J. Pesic, I understand your frustration, and Paolo's response is very cogent...the Fortran "ones and zeroes" at k(q) near zero need to be tuned "a priori or a posteriori" in either lambda.f90 and/or its output. I've dealt with this issue over the years attempting to calculate the Eliashberg-McMillan strong coupling superconductivity parameters in copper oxide perovskites. Please visit www.w2agz.com for relevant publications/presentations and references. I've found some success combining algorithms and code from PWSCF, GIBBS2, and ELK.
I visited and spoke at your institute in Belgade back in "the bad old days" of the mid-1970s, and recall fondly the hospitality of the faculty and the beauty of the contryside. Best, -Paul -----Original Message----- From: pw_forum-boun...@pwscf.org [mailto:pw_forum-boun...@pwscf.org] On Behalf Of yelena Sent: Thursday, April 14, 2016 6:49 AM To: pw_forum@pwscf.org Subject: [Pw_forum] electron-phonon calculation problem Dear all, I've been having problem with electron-phonon calculation. I keep having negative phonon frequency in gamma point, -0.00007, nowhere else, (so it is not "real" instability, but some numerical problem) but it keeps interfering with my lambda.x calculation and all i get is NAN for alpha2F and so on. So is there any trick to avoid it? This is clearly numerical issue, because other modes are positive and phonon dispersion looks as it should be and there is no any instability. Can I somewhere manually change that -, maybe? I've tried to increase tr2_ph (and decrease it, also) but it keeps happening just the same. And tried larger k and q grid. Still the same. Best Regards, PhD Student J.Pesic Center for Solid State Physics, Institute of Physics Belgrade, Serbia _______________________________________________ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum _______________________________________________ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum