Does anyone know how to do this?  I have a bunch of ideas written down at the Radio Analysis page, but honestly it feels like stabs in the dark.  I don't have any true noise generators with adjustable bandwidth (donate! :)), so my simulated wideband interference signal consists of a bunch of peaks instead (FM'ing a carrier with a sine wave).  Any tips or help appreciated.

Also seeking 1W+ 20dB+ 6GHz+ attenuators, and splitters/combiners (with good isolation) that can work @ 6GHz+.  I'd like to simulate a full cell site (3 PtMP sectors + some PtP links) within a totally confined coaxial system, and inject arbitrary noise into it.  Given the cost of these microwave attenuators + splitters, it might be easier to set it up with real antennas instead.  But that has problems of environmental noise and multipath / echoes.

Linked from the Radio Analysis page is the RouterBoard Metal 5SHPn page.  It has the result of 1 performance test so far, dealing with amplifier clipping leading to increased sideband energy.  This is of great importance in designing sectorized systems.  When these unexpected sideband emissions leak into adjacent antennas, you can shoot yourself in the foot.  So I'm trying to characterize the tolerance of the radios to nearby or in-band interference.  Doing this in TDMA/OFDM space is not trivial.

--Bart