scaling low resolution data

[Gregers Andersen]

I am processing data with max resolution 10 A. Processing images
with DENZO  works ok, but attempting to scale the files together
with ROTAVATA results in higly fluctuating B values, while the 
scale factors seems more appropriate. Part of the problem is possibly
that there are not that many reflections at all.
But can anyone think of reasons why rotavata shouldn't be suitable for
the job, or are there more general problems in scaling low resolution
data. At least Wilson statistics doesn't hold beyond 4.5 A.

Can anyone think of a better program to scale such data than rotavata?

Any help or comments are aprreciated
Gregers