I have a question concerning the time-step value for a molecular dynamics simulation (using Discover Module). I have been using the default value of 1.0 fs - it works for some cases but not in others (especially when I perform simulations at large stress values. The energy deviation between subsequent steps is too large and I obtain an error stating this, so I must reduce the value of the time-step). At the moment I am trying to use different time-step values (smaller than 1.0fs), however I would like to ask whether there is a way that this time-step can be automatically adjusted by the program itself if it 'feels' the need to do so.

Thanks for your help!

Christine