VASP Forum

Dear All,

During the relaxation of a charged cluster in a cubic box using ALGO = Normal, the calculation converges, but I obtain the following message:
BRMIX: very serious problems – the old and the new charge density differ
old charge density: 44.00000, new: 42.00000, difference: 0.524E+00

When I use ALGO = All, the calculation also converges, and this message no longer appears. However, I instead get:
BOPT < 0, no correction made 

The BOPT < 0, no correction made message also appears for some neutral clusters when using ALGO = All.

Should these messages be ignored, or do they indicate a problem with the calculations?

I have attached the input files and log outputs.

Many thanks in advance,
Ibrahim

At least in my case runing your input files I did not see this BOPT < 0 error.
Can you try restarting from a clean directory or remove WAVECAR?
But I would not worry about it.

As for the BRMIX error, unfortunately that always occurs when starting from scratch for charged systems when using the https://www.vasp.at/wiki/Category:Elect ... ency_cycle
This is because the initial guess for the charge density is always for the neutral system.
In the first five electronic minimization steps there is no update of the change.
As soon as there is the first update of the charge you see the BRMIX error.
In this context this is not a problem: the charge will become the one you specify in the INCAR file and the calculation should converge.

If you are having issues with this calculation you could perhaps try to do it in two steps:

1. neutral system (without NELECT, LDIPOL and IDIPOL)
2. charged system restarting from the WFs of the first step.

See Dipole correction section here:
https://www.vasp.at/wiki/Troubleshootin ... Correction

Hope this helps!

Ah, another feature that we added recently in VASP is the kernel truncation:
https://www.vasp.at/wiki/KERNEL_TRUNCATION/LTRUNCATE

This might also help!