发信人: bard (HBT), 信区: EE
标 题: Re: RF transceiver已经没有前途了
发信站: The unknown SPACE (Fri May 10 12:37:43 2002) WWW-POST
Why do you only mention CMOS?
Actually it may not be the right direction.
Although after many year's study, the performance of CMOS RF has improved
significantly. And some lab or company have claimed that their CMOS RF has
exhibit comparable performance as those of GaAs. But it has been found out
that their CMOS RF is not cheap at all, which has deviated from its original
motivations. In order to achieve such performane, their CMOS circuit is quite
complicated, which usually requires much more masks and processing steps.
There were quite a lot arguments about this subject in the IEDM last year.
Novel Si-based devices for RF are good but not cheap at all. The cost of
substrate now is only small portion of the whole cost of RFIC.of course, for
processor or memory, the situation is different.
When we look at high-end products, at least at present, CMOS RF can not touch
this field at all. They are all GaAs, InP stuffs.
【 在 avante (avante) 的大作中提到: 】
: 现在热的是5GHz wireless LAN
: 因为channel bandwidth is only 200KHz
: 对VCO/frequency synthesizer的phase noise
: 要求很高,对mixer IP3的要求也类似
: 而COMS再厉害还是利害不过SiGe or BiCMOS
: 再看Wireless LAN
: 但是一是channel 少了
: channle bandwidth 宽了
: 二是Q of L is higher at higher frequency,
: so 很多指标比GSM反而更容易达到
: 就算可能CMOS在GPS,or pager,甚至wireless lan可能有作为
: 当然bluetooth 2.4G a paper published
: in ISSCC2002 by David Su, but still
: a quesion: ten years of RF CMOS, but
: how many products?
※ 来源:．The unknown SPACE bbs.mit.edu．[FROM: 188.8.131.52]