kleing@ucsd.edu (Klein Gilhousen) (02/27/91)
In article <telecom11.121.2@eecs.nwu.edu>, motcid!ellis@uunet.uu.net (John T Ellis) writes: > 2. TDMA = Time Division Multiple Access > CDMA = Code Division Multiple Access > 3. The capacity increase of 3:1 is not on paper; we have shown the TDMA > capability to Pactel with a working demo system. The 20:1 increase > that is proposed for CDMA (Code Division Multiple Access, a form of > spread sprectrum transmission) is only on paper; nobody has > built a working example. > 4. The biggest problems with TDMA technology is the degradation of > speech quality with channel errors, and the fact that a POCKET-SIZED > digital mobile is several years away. Also the great difficulty in > expanding the speech compression to the final 6:1; that is the > theoretical limit for this application. The QUALCOMM, Inc. CDMA digital cellular system was demonstrated to the industry in November, 1989. This demonstration consisted of TWO cell-sites and a mobile unit all running CDMA. The system demonstrated all important features of the proposed CDMA system, including the "Soft Handoff" in which a call is processed simultaneously by two cells when the mobile moving from one cell to the other. In Feb, 1990, the demo system was taken to Manhattan to demonstrate operation in the high multipath and clutter environment of the mid-town skyscrapers. The system performed exactly as predicted (very, very well.) The referenced TDMA demo system did not (I believe) include a handoff capability, was a single cell-site and single mobile. This summer, we will be fielding our validation system here in San Diego. This will consist of several cell-sites and about 70 mobile terminals. The mobiles will be form, fit and functional prototypes of CDMA car phones. This system will prove the capacity improvement claims. I personally feel that the 20:1 claim we have made is quite conservative. We have had a hard enough time getting people to believe the 20:1 without sounding like we are completely crazy. As we have continued to improve our design, we have just held constant with the same claim, putting the increase in our hip pocket. CDMA addresses the issues of channel quality head-on by providing several additional forms of diversity. The result is a channel remarkably free of outages due to multipath, fading, interference, blockage, etc. There is no such "theoretical limit" on vocoder compression. The lower would be determined by the information transmission rate which is quite slow. The problem is to reduce the transmission rate without causing an unacceptably low quality. Both the TDMA and the CDMA approach use vocoders compressing to about 8000 bits per second. Testing has produced "Mean Opinion Scores" of about 3.9 on a scale of 5 where a wired phone is about a 4.5. Not too bad, but not quite as good as a wired phone. Improved vocoder algorithms should result in somewhat lower bit rates at this quality level. Interestingly, Motorola, along with AT&T, Nynex, Ameritech, Pactel, Nokia and Clarion have all committed to support the CDMA development. Of course, many of them also continue to keep the their TDMA developments alive until the marketplace resolves which system will win out. Klein Gilhousen, WT6G Qualcomm, Inc. Amazing NON-Disclaimer: In this case, I _DO_ speak for my company.