Unluckly, Self Tracking entities creates problems in the User interface(not only in the grid) for several reasons:
1) Due to lazy-loading if they are connected to other entities this other entities are NOT loaded before you put your object in the View. However, when the View try to access some property the original context might be already destroyed , thus causing null
2) Self-Tracking entities automatically imports database constraints imposed on the properties, so if for instance you don't save the key in a hidden field when the model binder try to read back your model you might experience a validation error, because
a required field has not been provided.
In your case probably the grid is trying to read some property containing other entitie of a 1 to N relation and it fails because of point 1).
The point is that the grid do changes-tracking to allow user undo changes in case of errors, and to make possible to understand what kind of operation need to be passed to the DB (insert, update, or delete) so it tries to save ALL ORIGINAL VALUES OF ALL
PROPERTIES into an hidden filed. This way it is able to rebuild the Old state of each row.
In the next version 1.0 I give the possibility to define exiplicitely the felds whose OLd value need to be saved. In case you use the themed datagrid it will assume they are the same of the columns you provided, so you don't need to provide any further information.
This should solve probably your problem, because in this case the Grid will not attempt to read connection properties but only the property provided in your column definition. However you might continue experiencing problems because of problem 2). In particular
you will be forced to provide in the View a value for all properties defined as required in the database(not nullable).
This problem is not specific of the grid...but is the reason why most of people prefer poco entities...;)
The new version 1.0 should be available in a couple of Days.