Measure from one to many relationship causing exponential table growth

I have had this a few times, and I think I am missing something regarding relationships or Measure behavior.

Let's start with this relationship.  The 'vdr Info' table has only one record per "VDR" which gives me a Table that shows only one VDR per row.  Good up to this point.

Now I add a "responsibility" table that also joins one-to many with the id_division field.

This table has something like this...

So what I want is to display the "resp_value" field in my Table.  It has to show the item that matches the id_division AND a specific bi_category.  So what I did is create a measure like this...

respVCodeErr = CALCULATE(
	FIRSTNONBLANK('IC eng_responsibility_list'[resp_value], TRUE())
    ,FILTER('IC eng_responsibility_list',[bi_category] = "Critical vCodes")
)

When I add this measure to the table, my number of records goes up exponentially to show me all permutations of column values.  Instead of 1 row per VDR, each VDR now has lots of rows, all with incorrect information...

Removing the Measure from the table sets everything right again.  Can anyone shed any insight?  Thanks!