ASIC project idea?

[Borkil]

I'm in an asic design class and we are supposed to come up with an idea for a final project; however, I don't really know what would be an interesting idea to design for. What are some ideas you guys have that would require an asic to do?

Here are some that I have thought of so far:
AES encryption (the prof has said that a lot of people have done encryption/decryption and I think several are thinking of doing AES this semester)
MP3 decoder
hdd controller (Not sure if an asic would be the best implementation)

 

[schenley101]

Does it have to be a digital design?

 

[WhoBeDaPlaya]

RF, analog, digital or mixed-signal?
Also, what level is this class? Introductory, intermedia, advanced, grad?

For my intro level classes, I've done an ADC, DAC, SERDES, 16-bit pipelined CPU and fully-diff op-amp.

 

[exdeath]

NES PPU with native RGB output.

Things like a HDD controller are better off running from software. It's almost impossible to work with parsing and navigating complex and dynamic data structures such as even basic FAT in a static hardware design in VHDL/Verilog; you are better off with an embedded CPU core.

 

[Borkil]

Does it have to be a digital design?

introductory
No, I don't believe so. So far the only requirements are
1. There needs to be some reasonably compelling reason why the selected idea would not or
could not be done in software on either a microcontroller
2. The idea needs to be useful as an ASIC (or an FPGA) for reasons due to requirements for
small size, low power, and/or high performance.
3. The proposed design needs to make use of moderately complex standard interface
formats such as MIDI, RS232, etc. or it needs to be designed to work with commercial
parts that make use of common interface formats that are de -fa cto standards such as the
serial or parallel interfaces used with LCD controllers.

 

[TuxDave]

Do you get to synthesize it to an FPGA afterwards? What about anisotropic filtering?

 

[Modelworks]

How about something useful beyond just a project for a class. A logic analyzer/digital scope that outputs to a graphical LCD is a project that almost requires ASIC type control to get accurate samples. AES, MP3, HDD are all things I can do with any micro, but doing a scope or analyzer with off the shelf micros has a lot of limitations that ASIC can overcome.
http://www.sump.org/projects/analyzer/

 

[William Gaatjes]

For those interested : Spartan 3 series from xilinx are also a great feature for the hobby guy / gal.

Xilinx ise VHDL editor is free. But you do no need to acquire a free license by use of your email. This license is coupled to the physical hardware of your pc.
But it has a simulator and everything else you need.

http://www.xilinx.com/products/design-tools/ise-design-suite/ise-webpack.htm


The fun part of these devices is that if you connect them to a SPI flash memory such as for example the AT45DB041D from Atmel(4Mbit dataflash), that these fpga chips can download the VHDL code from the SPI flash memory.
Thus by using a microcontroller (with proper program), one can download the VHDL bit file into the SPI flash memory from the pc and then the fpga will copy the bitfile from the SPI flash memory and configure it self.

Thus by using the xilinx ise editor and making use of this simple hardware, fpga logic is available for home use as well.
Of course it is also possible to bit bang from the pc the data into the SPI memory. Using for example an FT232H or FT2232 from FTDI and using the D2xx library. A microcontroller is not needed for downloading the bit file and everything can be from the pc side.

http://www.ftdichip.com/FTProducts.htm
http://www.ftdichip.com/Support/SoftwareExamples/MPSSE/LibMPSSE-SPI.htm

It may very well be that various projects already exist on the internet.

Example :

 

[William Gaatjes]

A document from xilinx how to do : Indirect Programming of SPI Serial Flash
PROMs with Spartan-3A FPGAs

http://www.datasheetdir.com/go/-xxm-jklm-bljl.pdf

 

[William Gaatjes]

For those interested :
Forgot to mention that the spartan 3A FPGA can also be programmed as a slave device without the use of an external program. My colleague made me aware of that option. This allows with a bit of handy programming and linker usage, to not use an external flash memory but store the bit file in the microcontroller memory.

Here is the user guide pdf for those interested :
http://www.xilinx.com/support/documentation/user_guides/ug332.pdf

See page 195, 196 for starters.

Do be aware that you might have to do some bit reversal.
If it does not work, please check that first.
The 50000 gates version, is available in a 144 pin casing that can be soldered by hand.

 

[CycloWizard]

FPGA is not the same thing as an ASIC. An ASIC is an application-specific integrated circuit, while an FPGA is a field-programmable gate array. The latter can be programmed to accommodate different applications, while the former is, by definition, application-specific: it's hard-wired. I doubt any professor teaching a course on ASIC design would be impressed if someone were to submit an FPGA-based project.

 

[William Gaatjes]

You are pulling my chain ...
I doubt that a professor at a college will spend a few 100.000 dollars per student to create masks (simplified :to chemically etch the silicon die) for a lot of asics that might not even work.

For prototyping and about how to use VHDL, fpga's are very useful.
ASIC's are also written in VHDL.


Here is a wiki :
http://en.wikipedia.org/wiki/Application-specific_integrated_circuit

An application-specific integrated circuit (ASIC) (pronounced /ˈeɪsɪk/) is an integrated circuit (IC) customized for a particular use, rather than intended for general-purpose use. For example, a chip designed to run in a digital voice recorder is an ASIC. Application-specific standard products (ASSPs) are intermediate between ASICs and industry standard integrated circuits like the 7400 or the 4000 series.

As feature sizes have shrunk and design tools improved over the years, the maximum complexity (and hence functionality) possible in an ASIC has grown from 5,000 gates to over 100 million. Modern ASICs often include entire 32-bit processors, memory blocks including ROM, RAM, EEPROM, Flash and other large building blocks. Such an ASIC is often termed a SoC (system-on-chip). Designers of digital ASICs use a hardware description language (HDL), such as Verilog or VHDL, to describe the functionality of ASICs.

Field-programmable gate arrays (FPGA) are the modern-day technology for building a breadboard or prototype from standard parts; programmable logic blocks and programmable interconnects allow the same FPGA to be used in many different applications. For smaller designs and/or lower production volumes, FPGAs may be more cost effective than an ASIC design even in production. The non-recurring engineering (NRE) cost of an ASIC can run into the millions of dollars.

 

[CycloWizard]

I'm in an asic design class and we are supposed to come up with an idea for a final project;

In case you missed it, William... It is very possible to produce ASIC masks for well under $100k. I can produce a very large, complicated one for less than $100k, let alone a simple one for a course project. But that's beside the point.

I agree that an FPGA might be helpful if he's called on to actually prototype the thing, but he's simply asking for ideas about what the ASIC should do - not how to do it. Your posts were therefore "out of left field" as we say and I wanted to clarify for any outsiders who might not know the difference.

 

[xboa721]

Along the lines of encryption you might find an algorithm for fractal compression. I believe there are some successful models available and implementing in hardware would make the process much more efficient over software/fpga implementation. e.g. a .fpg file format would be cool : fractal picture experts group or something like it......

Your new standard could be implemented and potentially marketed on video graphics cards for lighting fast standard compression PLUS some ground breaking fractal compression algorithms.

 

[William Gaatjes]

In case you missed it, William... It is very possible to produce ASIC masks for well under $100k. I can produce a very large, complicated one for less than $100k, let alone a simple one for a course project. But that's beside the point.

I agree that an FPGA might be helpful if he's called on to actually prototype the thing, but he's simply asking for ideas about what the ASIC should do - not how to do it. Your posts were therefore "out of left field" as we say and I wanted to clarify for any outsiders who might not know the difference.

That is cheap indeed. But is it worth the trouble for a course ?
Would it not be better to first use the fpga and when functionality has been proven create masks and the prototype asic ? Although i can understand that in conjunction with a company that desires an asic with specific functionality, a project group can be formed out of students as part of the course. That makes sense to me.

But i always post extra info for those interested. Some just might finally find what they where looking for to do...

 

[CycloWizard]

That is cheap indeed. But is it worth the trouble for a course ?
Would it not be better to first use the fpga and when functionality has been proven create masks and the prototype asic ? Although i can understand that in conjunction with a company that desires an asic with specific functionality, a project group can be formed out of students as part of the course. That makes sense to me.

But i always post extra info for those interested. Some just might finally find what they where looking for to do...

I agree that your extra information was helpful - I was just trying to provide the context to help those unfamiliar with the area. Without prior knowledge of the relationship between FPGA and ASIC, no one would know what you're talking about.

 

[William Gaatjes]

I agree that your extra information was helpful - I was just trying to provide the context to help those unfamiliar with the area. Without prior knowledge of the relationship between FPGA and ASIC, no one would know what you're talking about.

Indeed, you are very right. My positive and energetic enthusiasm does not work well with words it seems. I sometimes fear to create the opposite effect of what i intended. :| But, that is something that can be solved by emulation with proper preliminary research.

 

[Borkil]

Thanks for the info guys.

This project only needs to be simulated and will not built; however, the prof did say if the simulation is extremely long then we can map it to a fpga and use that for testing.

It's a group project and my group has decided to build a CAN to USB interface. Though I think its kinda a boring/stupid idea...

 

[Mark R]

Sounds like a classic project that is far better suited to a microprocessor/microcontroller rather than an FPGA/ASIC.

 

[William Gaatjes]

Sounds like a classic project that is far better suited to a microprocessor/microcontroller rather than an FPGA/ASIC.

I agree.
It is straight forward conversion of one communication protocol into another.

Most micro controller even have CAN and USB peripherals on board. Only glue code is needed.