Pulling together the best technologies from merger partners ST and NXP, ST-Ericsson has announced two new mobile connectivity chips that together deliver a complete solution for handsets and smartphones. The CG2900 is a combo chip that includes Bluetooth 3.0, FM receive/transmit, and GPS, the three most popular technologies for feature phones. For smartphones and other devices that require Wi-Fi, the new CW1100 offers a single-chip 802.11b/g/n device that is designed to cooperate with the CG2900. Both products use leading-edge 45nm technology and are scheduled to sample in 4Q09.
The CG2900 combines proven FM and GPS technology from NXP with ST's Bluetooth core. The Bluetooth stack is upgraded to support Bluetooth 3.0 high-speed mode, which routes Bluetooth requests through the external Wi-Fi chip, if available, and Bluetooth Low Energy (LE) mode. The GPS core has been improved from the earlier GNS7560 (GloNav) device to deliver –165dBm tracking sensitivity, a figure matched only by Infineon's standalone Xposys chip.
The CW1100 is one of the few mobile Wi-Fi chips that supports 802.11n, although it operates at 2.4GHz only and not in the less crowded 5GHz band. It is the only standalone 802.11n chip that integrates a CMOS amplifier powerful enough to eliminate the need for an external PA; the CW1100 is rated at +21dBm of output power. (Broadcom integrates a similar PA in its BCM4329 Wi-Fi combo chip.) This integration keeps the total solution size to 45mm2, the smallest Wi-Fi design yet reported.
Although ST-Ericsson is the fourth vendor with a GPS combo chip, it is the first to announce a 45nm device. The CG2900 matches the feature set of existing devices while supporting the latest Bluetooth standards and offering better GPS performance, particularly when compared against CSR's BlueCore7. The CW1100 offers 802.11n support and requires little board area. If ST-Ericsson delivers these products as planned, they will set a new technology standard in mobile connectivity. --Linley
Linley Gwennap, president and principal analyst
Complete coverage of competing products appears in our report "A Guide to Mobile Connectivity Chips."
Posts
5 comments:
If the CG2900’s performance matches or exceeds the competition, the power and space savings will likely lead to an increased uptake of this part.
http://deviceconvergence.wordpress.com/2009/09/12/st-ericsson-continuing-to-innovate/
Hi Linley, that's a nice summary, thanks! I'm curious about two points you make though:
1) 802.11n PA: At MWC09, CSR insisted that BCM4325 does require an external PA in practice because of FCC EVM specifications, but they didn't have a BCM4329 to be certain there. Are you implying Broadcom has assured you this has changed with that chip, or that CSR wasn't being completely honest?
2) GPS comparison with Bluecore7. While I'm sure they beat it nicely, is that really the point? Shouldn't it be compared to BCM2075 or XPOSYS instead? It's not clear to me BC7830 has a single design win yet. And this ST-Ericsson chip is unlikely to beat SiRFStarIV in anything but raw sensitivity; see http://www.gpspassion.com/fr/articles.asp?id=175&page=9 for example.
Also CSR is likely to be sampling one or two new (pre-SiRF HW) GPS chips around 4Q09 too (BC7-GPS+ and BC7-MM). While the HW/RF will be based on CSR R&D, I suspect the firmware & software will also be heavily influenced by the SiRF acquisition. We'll see what happens. After that, I'm sure we'll see 40nm 802.11n+BT+FM+GPS chips sooner rather than later :)
Good questions, Newbie. (1) Broadcom claims that most of their customer's BCM4329 designs do not include an external PA, although a few still do. Given the specs on the CG2900, it should be in the same boat.
(2) Comparing GPS chips is difficult, particularly on the basis of incomplete information. The sensitivity number that ST-Ericsson released is extremely good, even compared with BCM2075 or Xposys. Whether the overall performance lines up as well remains to be seen. But even if some standalone GPS products beat the CG2900, the chip is really competing against integrated solutions; in that regard, it should beat out CSR and TI on the basis of GPS alone.
(3) Of course, other vendors will be coming out with 45nm combo chips as well. It will be interesting to see what is shown at MWC. I think the big news here is that ST-Ericsson has pulled together technology from the merger to field a very competitive combo chip, whereas neither predecessor company had a strong position in mobile connectivity. This validates the merger and puts the pressure on Broadcom, CSR, and TI to respond.
Thanks for taking the time to reply Linley - that's very informative :) I definitely agree that these are very solid products and they validate the ST-NXP merger (and indirectly also the Ericsson one via the bundling opportunities). I'd argue it also puts tremendous pressure on Atheros and Marvell as they risk being marginalized further.
Just one quick point: my understanding is that CSR will stick to 90nm for their next GPS combo chips in the short-term. However some non-CSR insiders are speculating they might have a few tricks up their sleeve (APE2 DSP as GPS baseband probably and they might be early adopters for Tela's AreaTrim which TSMC seemingly renamed Slim Standard Cells on 130nm). If you have a call with them when they announce one of these chips, it might be interesting to see if they can be tricked into talking about those things ;) (plus their reasoning for sticking on 90nm in general)
Hi Linley, thanks for very informative summary.
A quick point on tracking sensitivity. At power levels lower than -160dBm quality of local clock (drift) and user motion plays significant role due to deep integration. Do you have insights if deep sensivity really pans out in end-user equipment (where clock source undergoes thermal variations) or it is ideal lab numbers?
Post a Comment