WLAN is now in CMOS, and a commodity -- everyone has them. Bluetooth same thing, UWB not selling, Cellular Transceivers all in CMOS and everyone will have them soon.
What is left in RFIC Design ? WiMAX ? Is there any reason to study or go into RFIC Design anymore especially in BiCMOS ? What is left ?
Also software defined radio -- will it kill RF Front Ends ? All digital architectures coming like at Texas Instruments ? Why study RFIC ?
Please give me your opinions and post them -- thanks.
-todays designs for commercial RFICs products are largely in BiCMOS. You're right for WLAN, but for Cellular or for high performance systems is still largely used. The reason is simple: less cost/performance.
-CMOS has really sense where SOC is the system strategy.
-being a good BiCMOS RFIC designer is the best way to become a good CMOS RFIC designer
-SiP allows the use of many different technologies in the same product and is delaying pure CMOS
-Software radio is still too young: ADC conversion is the bottleneck
-subsampling TI: do you know why there is not a product but a lot of papers, book and so on?
It seems that all these are statement, but is not like that, it's just my opinion.
the subsampling architecture still suffer from high noise , and still in the beginging , ald the subsampling , they still need LNA , and a mixer for sampling , this means they still need analog
even the high speed didigtal , and IO dsign , like serdes and so on considered analog
even in very high speeds like PCI express and SATA IC's u can find the bit rate 6 Gbps , so this speed need a very good analog even u can say RF designer
Puppet
can you give me some evidence on what you say about current products in BiCMOS?
as far as I know, for example, all Nokia platforms for their cellular phones are in 0.35/0.25 SiGe BiCMOS. The reason is the low cost vs high performance of these stable and mature technologies and also evolution of their products (it's fast to redesign some prodution IC for new generation adding or modifying something respect to redesign a complete new one).
i personally know no one is getting hired in north america in rfic design in cmos-bicmos that are young and have no experience.
is there anything left ... 3G has not taken off, yet now professors working on 4G... wimax looks like it will not do anything, WLAN is pretty dead ... all being done for very cheap in china and india and what else is left .. cellular now in cmos and getting low cost... USB is not taking off, bluetooth...
what is left in rfic design ... 60gb-s radios .. radar ... what is left to do.
why bother ...
people say there are jobs, yet no one is hiring young people. so what is the point of getting into it...
Commercial UWB is not selling because it's not fully defined and there are few viable products that offer true QoS - watch over the next year or 2 things may change.
I think it takes at least 5 yrs to have something big to change. That's why, for most of the classes you took, need to relearn the technologies to update yourself. I think there is more into RFIC design, could be a new type of technology will overcome many of the disablility CMOS or HBT has.
From the employer risk position it is getting more difficult to complete a design team team starting from the expert which modify the rough foundry designkits up to the MAC layer SW which understand which code have to be in HW and SW to fullfill critical timings.
Sometime no single company could provide all the knowledge and the cooperation is still increasing. So there is further need to cooperating in increasing manner between companies. For starters it seems that the final systems getting more far away in terms of knowledge. For people already in the job it is still difficult to follow the increasing rf system complexity. Sometimes the systems gain faster complexity overall than people could capture. That narrows the number of people which understand nearly everything. I does not think that knowing a little of all helps or that managing on report base works. That message told the engineering rf systems is getting more risky or simple more difficult to map to the right people in time and space to get the job done.
Single die CMOS for a complete radio system may be a small niche if other factors are more important. Because of the above mention complexity and risk a rare situation is to compare different technology approaches in terms of cost, area, power and development resources.
Take for instance a 90nm CMOS WiMAX SOC design versus a
65nm CMOS BB+MAC
130nm CMOS ADFE
250nm SiGeBiCMOS Radio
SiGe PA
GaAs Switches
So 5 technologies versus 1
Now compare the
Production Cost
Test Cost
Yield
Assembly Area
Power Consumption
Design Risk & Resources
To make money it is more important to know what could be done than simply give CMOS a try. That is research!
now there is a lot of work in RFIC design for mmwave, radars, imaging, and for gigabit wlan and so on.
when i talk about rfic i mean for lower frequency standards, under 10ghz or so.
also reason i say these things is because silicon labs just got bought by nxp, and silicon labs pumped cmos rfic for cellular for many years, then just quit the market all of the sudden. TI making low cost digital RFIC's, and also now WLAN only being done by few companies -- again all in cmos.
RFIC design is dead? Of course not but what's dead?
I think there are still some challenges in RFIC design.
a. CMOS v.s. BiCMOS: CMOS will overcome BiCMOS if the breakdown voltage of CMOS isn't reduced while the gate lengtg of MOS decreases. Is it possible for advanced CMOS process?
Ex. Designing a CMOS TV-tuner (for cable and DVB-T) is still knotty.
b. 0.13um, 90nm...CMOS with Cu interconnect and low-k dielectric will have very good device characteristics and UWB, 60GHz wireless standard, Ku/Ka band LNB, LMDS...etc provide the challenges to CMOS RFIC design. As the design frequency goes higher, the distributed concept (convenctional microwave design) not lumped-elements concept will be introduced as the development of IBM 60GHz transceiver. How interesting it is. Therefore, you can image how many funny things you can play. Are you sure you have the chance to play even for experienced designer?
c. What's dead? Everyone knows the answer: small capital design houses and their RFIC designers. No money, no chance to play the advanced technology. Oh! I'm talking myself. Help! Help!