your thoughts on FPGA?

[ahack3r]

I'm researching FPGA boards at the moment for my autonomous motion tracking project(s)...what are your guys thoughts on it? I'm slightly confused as to how they work...can you communicate to it with a computer via serial?

[waltr]

An FPGA is just firmware configurable logic. One writes a description of what the logic is to do using VHDL, Verlog or by a schematic. This 'code' is built and fitted to the logic cells of the FPGA then loaded into a PROM for the FPGA to load on power-up, hence:field-programmable gate array (FPGA)

Code can be written for an FPGA to do asynchronous serial or almost any other protocol. You would still need an external logic level to RS232 level shifter.

Try a google search on FPGA for more detailed info.

[Soeren]

Hi,

An FPGA (Field Programmable Logic Array) is more or less a GAL/PAL/PEEL on steroids. You can build a µcontroller in an FPGA (if you're sharp), but if you're confused about them, I think you'd be better off using a ready made controller.

What are your reasons for wanting an FPGA over a µcontroller?

[madsci1016]

I thought learning and playing with FPGA using VHDL was fun when I did it, but I agree with Soeren, it's a different animal. The hardest thing to adjust to was what makes sense in usual programming structure doesn't work in VHDL. If your project doesn't have the time for you to learn FPGA, stick with uC.

[ahack3r]

Hi,

An FPGA (Field Programmable Logic Array) is more or less a GAL/PAL/PEEL on steroids. You can build a µcontroller in an FPGA (if you're sharp), but if you're confused about them, I think you'd be better off using a ready made controller.

What are your reasons for wanting an FPGA over a µcontroller?

my reasons for looking into FPGA's are because of these articles: http://www.societyofrobots.com/member_tutorials/node/369 http://www.optoiq.com/index/machine-vision-imaging-processing/display/vsd-article-display/4862174297/articles/vision-systems-design/volume-15/Issue-8/Departments/Technology_Trends/MACHINE-VISION-SOFTWARE-Graphical-software-tools-target-embedded-applications.html

I heavily suggest you read both of those links, I think there may be more to FPGA's than what was originally thought by the Robot communities. Also, I'm an Electrical Engineer major, so I'm looking for the real deal...not just a quick and dirty Arduino project with some stolen roborealm code.

[madsci1016]

I heavily suggest you read both of those links, I think there may be more to FPGA's than what was originally thought by the Robot communities. Also, I'm an Electrical Engineer major, so I'm looking for the real deal...not just a quick and dirty Arduino project with some stolen roborealm code.

No one here put down FPGAs as a possible solution, we are just trying to make the point that if the goal of your project is to make a robot, there are easier ways. If you want your project to be about using FPGA to make a robot, have at it. But you are going to be spend more time on the FPGA part and less on the robot.

I'm a degree'd Electrical Engineering (under)graduate, Soeren & waltr are seasoned EE experts.

I took grad level classes in VHDL and computer architecture and designed a single cycle core architecture in FPGAs as a term project; it's no picnic.

[ahack3r]

I heavily suggest you read both of those links, I think there may be more to FPGA's than what was originally thought by the Robot communities. Also, I'm an Electrical Engineer major, so I'm looking for the real deal...not just a quick and dirty Arduino project with some stolen roborealm code.

No one here put down FPGAs as a possible solution, we are just trying to make the point that if the goal of your project is to make a robot, there are easier ways. If you want your project to be about using FPGA to make a robot, have at it. But you are going to be spend more time on the FPGA part and less on the robot.

I'm a degree'd Electrical Engineering (under)graduate, Soeren & waltr are seasoned EE experts.

I took grad level classes in VHDL and computer architecture and designed a single cycle core architecture in FPGAs as a term project; it's no picnic.

really? thats very impressive! thanks for the information, I realized that going FPGA would be harder but my train of thought was "if its what they'll be using in College, I mine as well start now". But if you say its too hard to start with, I'll take your word.

[waltr]

Since you are in school for EE this could be a good opportunity to use a PLD (any type of programmable logic array).
They are getting more and more powerful and are slowly replacing processors in many types of embedded systems. They are not going away.

[ahack3r]

Since you are in school for EE this could be a good opportunity to use a PLD (any type of programmable logic array).
They are getting more and more powerful and are slowly replacing processors in many types of embedded systems. They are not going away.

my thoughts exactly, I want the upper hand in my upcoming EE classes and want a skill that can be directly helpful with my degree, and after graduation.

[cyberfish]

The biggest advantage of FPGA is probably speed. You can get orders of magnitude higher speed/price and speed/area using an FPGA vs a microcontroller, if the application is right. An FPGA can be optimized to be extremely fast at the very specific thing you want to do, while a microcontroller has to be able to do a lot of things that you don't need.

That said, unless you really require that much computing power, an FPGA is just free head-ache.

Think of it like a massive breadboard, that you can "wire up" programmatically.

Whereas a microcontroller is a fixed circuit. You give it some input, it gives you some output, like a black box.

It's perfectly possible to wire up a microcontroller in an FPGA. But if that's all you do, you are wasting all the FPGA's power of parallel processing, and might as well just use a microcontroller, which would be cheaper and probably faster at being a microcontroller.