1995 was to be the breakthrough year for CDMA in the US. The FCC made the 1900 MHz band available for digital mobile telephony and launched the concept of PCS (Personal Communication Service) for the mobile technologies approved for the frequency band: CDMA, GSM and TDMA.
Åke Persson recalls: “It turned out that half the American operators went for CDMA. Our decision at Ericsson not to get involved in CDMA deals led to disappointment in our US marketing division. The pressure on us then went on increasing as the CDMA operators grew in strength.”
Outside the US, CDMA did not pose a short-term threat to GSM, but Qualcomm had made it clear it intended to develop CDMA2000 as a 3G system. This would offer better data-transfer possibilities than GSM.
Uddenfeldt recounts: “This was a genuine threat that we had to make up our minds about. Our strategy eventually became clear: we were going to develop our 3G solution on the basis of GSM’s core network. Then investment in GSM could be described as one step on the way to a broadband, 3G solution.”
One advance in the GSM process was called GPRS (General Packet Radio Services, often referred to as 2.5G), a platform for mobile data-network services in which the data can be transferred at rates of 30–100 kbps, for which there was a technical specification in 1997. Because information is transmitted in “packets” when needed, phones can be connected to the internet all the time.
AT THE CUTTING EDGE
Håkan Eriksson was at the cutting edge in the work on 3G. In a career review session in 1991, he told his manager that his personal objective was to run a research project based on wideband CDMA (Wideband Code-Division Multiple Access, or WCDMA).
Eriksson recalls: “Within Ericsson, we studied both TDMA and CDMA. For a while, we worked with Philips on a research project, while Nokia worked with Alcatel. But the Nokia project with Alcatel failed, so Nokia joined Philips and us instead. Then Philips decided to pull out of mobile telephony, and their research team moved over to Ericsson’s new R&D team in Nuremberg. From then on, Ericsson and Nokia worked on WCDMA as a 3G technology, the same way we had worked on GSM.”
The route to victory turned out to lie in Ericsson’s relationship to the government telecommunications monopoly in Japan, NTT and its mobile telephone arm, DoCoMo.
Beijer says: “The Japanese got no support for their JDC standard outside Japan. When they discovered that GSM was gaining ground, they tried to counter it by renaming their system PDC (Pacific Digital Cellular). But this was a marketing ploy that flopped.” He suggests the Japanese fiasco was partly due their failure to realize the importance of roaming.
With 2G a lost generation for Japan, it wanted to make up for lost ground in 3G. And in 1994 Ericsson and DoCoMo reached agreement on a joint development project focusing on WCDMA. A year later, Ericsson was able to demonstrate a pilot system with a bandwidth of 5 MHz for a few major customers. As a result, DoCoMo ordered its first trial system from Ericsson.
This determined Ericsson’s approach to 3G – and Håkan Eriksson’s next assignment was to move to Montreal, to the research unit that had been set up there in connection with ERA’s contract to supply operator Cantel.
His task now was to try to shift D-AMPS onto the GSM track and create from these two platforms a joint technology for higher data-transmission rates. He succeeded, and the solution was named EDGE (Enhanced Data Rates for GSM Evolution, sometimes known as 2.75G), with its technical specification release in 1998.
EDGE offers theoretical transmission rates, if all eight time windows are used, of 471 kbps in both directions. Because the specification of the handsets can be somewhat simpler if they do not receive and transmit at the same time, the usual approach is, however, to use four time windows downstream and two upstream, which offers 236.8 kbps downstream and 118.4 kbps upstream.
EDGE’s modulation technology can be considered a return to the adaptive modulation techniques proposed and demonstrated by Torleiv Maseng in the Paris tests in 1987. Beijer says: “Instead of always dimensioning for the worst possible radio channel, you go on transmitting with higher speeds if you happen to have a good connection. Maseng must have had the last laugh when EDGE was incorporated into GSM.”
Author: Svenolof Karlsson & Anders Lugn