CCL: Boron NT fails with "Diagonalization in DiagDN via DSPEV failed"

From: "Mike Edmund Miller" <really_mike]~[yahoo.com>
Subject: CCL: Boron NT fails with "Diagonalization in DiagDN via DSPEV failed"
Date: Wed, 2 Mar 2011 11:30:16 -0500

 Sent to CCL by: "Mike Edmund Miller" [really_mike{:}yahoo.com]
 Dear CCL members,
  A boron nanotube which I am trying to optimize fails as indicated in subject
 header. The route line for that job is:
 #T Opt Freq=raman b3lyp/6-31g(d) scf=(NoVarAcc,maxcyc=9999)
 I included the "NoVarAcc" keyword in an attempt to solve the original
 error of "Inaccurate quadrature in CalDSu". Other unsuccessful
 attempts to deal with the inaccurate quadrature error included 1) another WF
 guess, i.e.,  "Guess=INDO" and 2) an ultrafine grid together with a
 opt=tight.
 I have also tried tweaking the input coordinates manually.
 Thank you for your ideas / guidance you could provide in this matter.
 Mike