0.13um RFCMOS and BB SOC or SiGe RF +0.13cmos BB?

[GDF]

I would like to know in WLAN application, which one is cost effective?
Why SOC is the trend dispite the obstacles? By the way, who knows
the Average Selling Price of 0.13um and SiGe0.35um?

Thanks

 

[rfsystem]

That depend a little bit on the partioning.

If you can integrate the ADC/DACs on the 0.35u SiGeBiCMOS it tend to be more cost effective to do 2-chips. The scaling of analog tend to stop below about 0.25u. So active and passive devices could profit from scaling. So a cap for a filter has less area because the oxide thickness is also reduced but the supply voltage scale down in about the same manner. To get the same specs the cap value have to be higher. That is one example but nearly all analog functions show similar scaling stops.

So a 90nmBB+350nmRFAD could be a winner.

 

[GDF]

The ADC often tends to be implemented with digital circuit due to the limitation of
pin counts, so any good way to integrate it with RF? Is there any part available in
the market?

how about 90nm RF? TI performs aggressive approach in this topic.
For example, they integrate 90nm digital RF and its baseband, is it cost effective
than 90nm basebaand+BiCMOS RF?

 

[rfsystem]

Pin limitation is not an issue. Analog IQ uses 8+ pins. A serial interface like Jedec-61 could be used. The emission of the serial interface into the LNA is the issue. I think it is simply an IP/KnowHow problem. If someone have found a way to do ADC on RF it will happen because it is clearly cheaper. I have heard about multimillion $ contracts to integrate ADC for WLAN on a single project for only one process generation.