[SQL Server 2000] - Selecting Nested Top N

Hi There,

We have problems getting the right records from a joined table.

First we have a table with all our Customers (eg. [Algemeen-B])
We join them with a table representing all Contacts per Customer (eg. [Algemeen-C])

An SP joins the tables because we have a few dynamic filters and parameters.

What we need is to link a maximum of @amount contacts per Customer.
Where each contact has a RANK (the TOP @amount ) joining the Customer Table.

We come thisfar:

-- The SP results this SQL select:

select *
from [Algemeen-B] left outer join
(
select * from [Algemeen-C] AS C
where exists
(
select * from [Algemeen-C]
group by bedrnummer
having bedrnummer=C.bedrnummer And min(rank) <= C.Rank and count(*) <= 2
) ) S
on [Algemeen-B].bedrnummer = S.bedrnummer
where KWP>=8

---
In this case I should return a maximum of 2 contacts per each Distinct Customer.
But I think the: 'And min(rank) <= C.Rank and count(*) <= 2' is wrong.

How can I select the TOP 2 Ranked Contacts for each Customer?
Because when there are no Contacts for the Customer it should return only the Columns for [Algemeen-B]
and NULL values for the Colums for [Algemeen-C].

If you need more info Tell me..
Hope someone can help me out..

Colin