Questions about general ASIC design.

Hi.

I have a few questions about ASIC design. I tried some searches on the forum and found little general information. It was mostly specific on one thing.

My knowledge:
I am quite new to ASIC design and I have no experience in this field. I have been working with Altera FPGA systems and microcontrollers the last two years and have a fairly descent understanding of IP components and VHDL design. I have also been working with multiple soft processor systems and the use of real time kernels. However I still feel this is a bit new to me. The designs have all been on FPGA so it has all been digital design.

Current project:
I am currently working on a small project that uses a microcontroller based on the ARM M3 Cortex. The controller reads values from an ADC connected to a sensor and transmits the processed data over a Wi-Fi network or a Bluetooth connection. The system currently consists of pretty big modules on a prototyping board. It’s pretty basic.

Now to my questions:
Let’s say I wanted all of the components in my above mentioned current project in one chip.
-CPU
-WiFi
-GSM modem (maybe)
-Bluetooth
-ADC

1. Can one easily combine analog circuitry like in an ADC with digital logic in an ASIC?

2. Is there a lot of IP cores available for these kind of chips or do you have to do a lot yourself?

3. How small can a chip featuring the above mentioned components become?

4. What kind of software tools and/or hardware tools is best for ASIC design?

5. What is the main difference when working with ASIC vs FPGA?

6. Any good up to date books with design examples and/or tutorials?

7. Anything else worth mentioning?


Thanks for any help. I would really like to know more about ASIC design.

-Stian

Regarding point 7:
If you are new to VLSI, then I would not recommend you do that project. Its quite a big task to undertake. You would have to do a lot of research on IP cores, not to mention that you also wanna do a mixed-signal design! My advise would be that you do something very simple. Something which you can learn the CAD tools, design methodology, and asic design flow from. Having done that, then you MAY be able to tackle that mammoth.

Hi Stian,

This task is not for beginner but if you are so brave good luck.

1. It is possible to combine analog and digital. You need custom design software for it.
5. The main difference between ASIC and FPGA is backend. You should learn backend flow.

Its good to get a perspective on what kind of scale such a project has. Taking the above mentioned design to an ASIC design is not the current plan. I just want to know the possibilities and pros/cons of doing so.

After reading a bit more here on the forums i found that there is a lot of different software for the different stages of the design. What would be a good place to start when it comes to learning the software tools? I read a bit about Cadence and i have used software from Cadence before.

Now the questions i consider the most important of the above is 3, 4 and 6.

Especially how small it is possible to make a chip like that (estimate/educated guess)? What is possible/reasonable?

-Stian

1. Mixed-signal design is very specific, and it require some skill to avoide failure. High accuracy analog part require a lot of efforts in integration with noisy digital core,- it is a separate song!
2. There are a lot of IPs on the market, practically everything can be bought. It is a question of money, project budget, which is typicall limited. An experienced designer from company, having good relation with a chip manufacturers, can reduce IPs cost by proper choicing the manufacturer.
3. Strongly depends from technology node and relation between analog and digital parts,- analog isn't scaling such good as digital. In 0.18um node such chip can be about 10sq.mm.
4. For mixed signal with significant custom design portion Cadence is preferable. Verification is better from Mentor. Digital design with external analog IP usage - Synopsys.
5. Back end mainly. FPGA is based on ready blocks with a limited flexibility. In ASIC design you have more freedom and more risk in design.
6. Necessary to define preferable specialization direction. General descriptions are practically usefullness.
7. A lot of else, but not in brief description

1. ASIC boards are available in the marked
2. Already these IPs are proven and available...IP developers(companies) surely have this IP in their portfolio....
4. Only the way you handles it...
5. If we found error in FPGA it is easy to fix and if it is in ASIC it is too costly to fix it.