Broadcom's QAM decoder chips have been capable of handling 256 QAM at least since 1995 -- that's when we bought their test system module that has switchable QAM constellations and variable equalization string length. We ran it at 256, and it certainly works, albeit with the higher C/I you would expect from theory, which was difficult to achieve in the MDS environment we were examining at the time. Most of the use we had for the BRCM test system module was for live testing of BER and block error rate in digital signals carried over MDS links under controlled and ambient measured interference conditions, as well as variable equalization length (neutralizing varied depth of multipath interference.) For that, it is a fine tool -- tho a pricey one (~$15,000) for what is in it. I guess you always pay for calibration. The price of the BRCM chip set at the time was about $24 in quantity.
As I remember, USWEST tried 256 QAM down some new HFC cable they built in Omaha [?] back before they spun off their cable interests, also, I think, back in 1995, maybe 1994. My recollection is that they ran into some performance issues with the medium, tho they made it work. There may have also been an issue of being out in front of standards. However, both Thompson and General Instrument, at least, would support 256 QAM at that time if the service provider wanted it. Thompson was using their own QAM decoder chips, as I remember (so very long ago!). |
