In the Cube Drillthrough Options screen, on the Columns tab, you should see all of the columns contained in all of the tables used in your model. Assuming you have only 2 tables, a fact and a single dimension, the columns may be something like
COLUMN TABLE
fkCol1 tblFact
fact1 tblFact
fact2 tblFact
pkCol1 tblDim1
Desc tblDim1
Assuming that you have run the optimise schema, the model will now not have a join between tblDim1.pkCol1 and tblFact.fkCol1
What you need to do is on the Filter tab of the Cube Drillthrough Options screen, add a filter like
"tblDim1"."pkCol1" = "tblFact"."fkCol1"
So basically you're using the name of the tables to make the join/s in the drillthrough filter, you shouldn't be trying to use the (models') Dimension name, it won't work.
HTH,
Steve.
Steve.