The convergence criteria for a SOL101 Linear analysis including contact is defined in the Connection Properties of the contact elements of the model.
With my version of Femap with NX Nastran two criterion are available :
Using the first criteria would lead to an output .f06 file with the following text :
1 LINA+ APRIL 8, 2016 NX NASTRAN 5/ 5/15 PAGE 642 NUMBER OF CONTACT STATUS CHANGES: 4 (NCHG: 5) NUMBER OF INACTIVE CONTACTS: 2784 NUMBER OF STICKING CONTACTS: 1593 NUMBER OF SLIDING CONTACTS: 375 CONTACT ITERATION CONVERGED
As you can see convergence is reached for this analysis because the number of contact changes between two contact loops is less than 5 (convergence criteria). This is good. However the number of sliding contacts is still quite high.
My point here is to discuss with you the reliability of the criteria Number of changes. Shouldn't the number of sliding contacts be taken into account when a static contact analysis is performed ?
The criteria is not about how many are sticking or sliding, it is about how many change status (between inactive, sticking and sliding). The table in the .f06 is telling you that you have 4752 contact elements and that only 4 of them changed status between the previous iteration and the current iteration.
Recall that a status of sliding does NOT mean that the contact element is physically continuing to move. It means that the tangential traction is equal to the normal force times the coefficient of friction (and sticking means that the tangential traction is less than the normal force times the coefficient of friction). This is small displacement linear contact, so the contact pairing is not updated as it would need to be if there was large sliding involved.