Thanks to all of you for your replies, they were all extremely useful.
In particular, it seems I can now explain a strange behaviour (that was also one of the reasons why I opened the topic) that I was having with contact analysis results: when I set for example 10% it often happens that two contacting surfaces interpenetrate each other (thing that would be impossible)--> so this is because the gap is amplified? also even if the initial gap is zero? (conincident contacting surfaces)
thanks
bye