Hi avalon
A better approach is to convert all significant elements into equivalent noise temperature, then add them up.
There are some excellent sources on the net to guide you through the process. See for example
http://www.satsig.net/noise.htmAlso agilent have an Application Note on the subject number 57-1 which can probably be downloaded from [url]hp.woodshot.com[/url]