Dimension table on many side. Problematic?

I am really struggling to model my data into a star schema.

I've read tons of documentation and Kimball stuff and I feel like I understand a lot of it

However, when I look at my own data I suddenly have no idea what to do anymore.

I am wondering, is my data so different? Can the rules in my case be different? Or am I being dumb?

Few things to know about my data and my desired result:

• All my dimensions are basically fast changing, at the same rate as my facts.
• I have facts at different granularities.
• My dimensions will be used a lot in visuals therefore they must filter each other. For example I will make gantt charts and show dimensions over time. In regular star schema design dimensions do not filter each other. This is unacceptable for me. I considered creating a link table with all the keys. I tried implementing it and the result looked really stupid like this: dim tables(1)---(*)link table with keys(*)---(1)fact tables. Then I realized it would probably be a bad idea because:
  • It would probably just further complicate DAX.
  • The key table would quickly grow to hundreds of millions of rows because I had to merge all these keys together.
  • I cannot create hierarchies in Power BI because my dimensions are in different tables.

Let's say I have hierarchical dimensions like below. A to G 'could' all be different tables with each having multipe fields.

A can have one B

B can have one C

C can have one or more D

D can have one or more E

E can have one or more F

F can have one or more G

And I have several fact tables at different granularities like below.

FactTable1 has a 1 to 1 relationship with D

FactTable2 has a 1 to 1 relationship with E

FactTable3 has a 1 to 1 relationship with F

Because I decided that a link table with keys was a bad idea, and because I want dimensions to filter each other, and because I want to be able to create hierarchies, and because I want to avoid bidirectional relationships, I decided to merge A to G into 1 big dimension table. I read at several sources, that the fewer tables, the better. I would simply merge dimensions and create folders inside Power BI GUI if I want to bring structure to it for the users. And here a new problem arises: 

Because E expands on D, and F expands on E, and G expands on F, extra rows are being created all the time, resulting in the relationships below.

FactTable1 now has a 1 to many relationship with D

FactTable2 now has a 1 to many relationship with E

FactTable3 now has a 1 to many relationship with F

This is the opposite of what I've always learned. Fact tables should be on the many side, dimension tables on the 1 side.

However, when I think about it, I can't figure out what could go wrong:

If I filter A, B, C and D, FactTable1 results 1 row.

If I filter A, B, C, D and E, FactTable2 results 1 row.

If I filter A, B, C, D, E, and F, FactTable3 results 1 row.

This always allows me to navigate down to the granularity I want. It's almost like 'hidden' 1 to 1 relationships inside the model.

Is this a bad model? What could go wrong? I can find nothing on the internet about this situation. Or am I creating this situation for no good reason?

Is anyone able to help me in the right direction?