There is not any backbone sampling so I generally find RMSD to your crystal framework being < one.0. Unwind really will do specific bb sampling So offers a decreased Electrical power structure than my protocol but might also introduce the variations that you noticed. I'm pasting my typical possibilities file https://eduardoeqzjr.activoblog.com/26895243/https-fastrelx-com-things-to-know-before-you-buy