The three page brochure lacks detail. It appears to be a phased array, but with what they claim is a superior concept and improved implementation of the feed network behind the radiators.

The bandwidth shown in the graphs (37 to 40 GHz) as a percentage bandwidth is good, but not amazing. At these frequencies, the loss is typically so high that VSWR itself is rarely a direct problem. But loss is so high that one wouldn't want to exacerbate it with high VSWR.

Being extremely high gain, they're obviously intended for fixed point-to-point links. Such applications are often installed in locations (e.g. part way up a tower) such that a simple dish installation is perfectly acceptable. Obviously the mm-wave radio itself needs to installed right there, or preferably integrated, to mimimize loss.

If it's really "50% less expensive" than the alternative technologies, then there's no need to avoid them.

"Invented in 1991" is 21 years ago. Not many were working in mm-wave radio (e.g. 40 GHz) in 1991. Some, researchers, but not many.

Their earliest patents must be getting ripe by now. Of course, they'll have layered on additional patents in the meantime. If you can find their patents at USPTO.gov, then you could follow the links (prior art, both backwards and forwards) to find others working in the same niche.

As a system designer, you could pretend to ignore the technology inside their antenna subsystem and simply focus on the performance (air interface, physical characteristics, etc.)
 

US Patent 5266961 reportedly expired at the end of Aug 2011.


Here's another one: http://www.google.com/patents/US5771567

Notice the 'Citations' and 'Referenced By' sections. The 'Referenced By' list leads you forward in time to the others in the same niche.
 

Thank you for your advice.I'll follow the patent to sourcing more supplier. I'll pay more attention during the negotiation.
attached is the IEEE file  

• http://files.engineering.com/getfile.aspx?folder=92f245f0-fa3f-4c9c-8f6f-dc

This array is Linear polarization, perpindicular to mounting surface. For similarity, picture a yagi antenna V pole oriented, atop a table that's simply rotated with Az over El rotator mechanics.

At zenith, its polarization is dependent on the azimuth pointing direction, so it can be rotating linear at zenith.
 
It needs a polarizer for circular polarization since it's inherently linear polarization.

It can do endfire V pole efficiently, but to get H pole or good circular, it'd have to sit up high above the surface that it's mounting on.  

Looks like a good option for higher frequencies due to the lower loss in the distribuiton across the aperture with it's slow wave structure.