Actually the issue is a different story.

It turns out that my secondary tables had multiple columns with FK constrains, some of then they were null values, from the business perspective there is no issue with that. the error SSAS was throwing had not a clue of what was the real issue, since the value "62" it does exist in every dimension, but in a lower level a secondary dimension with a FK which i was not using at that time had null values.

my conclusion is that SSAS is not able to process dimension tables with nullable FK columns. the error is thrown at the cube processing step, the dimension processed successfully every time, i guess it is a flaw in the way SSAS build the query for the cube.

does anyone knows if there is a configuration somewhere to set is processing behavior differently?