Shop OBEX P1 Docs P2 Docs Learn Events
Beginner friendly FPGA boards $55 ? — Parallax Forums

Beginner friendly FPGA boards $55 ?

prof_brainoprof_braino Posts: 4,313
edited 2011-07-24 10:32 in General Discussion
http://www.kickstarter.com/projects/13588168/papilio-fpga-shield-for-arduino?ref=category

Is there such a thing as beginner friendly FPGA? What would one need to program the FPGA?

For $55 we have to add our own SRAM,
For $85 the board includes 4Mb SRAM

Would an ardinuo shield compatible prop be a good match for this?

Comments

  • LeonLeon Posts: 7,620
    edited 2011-07-24 07:29
    The free version of the Xilinx ISE Design Suite would be used for development.

    The configuration software would need to be ported to the Propeller, it would be much easier to use an Arduino. Or, a Xilinx programming cable could be used.

    The Cyclone II board I mentioned recently might be better value.
  • edited 2011-07-24 07:51
    http://www.kickstarter.com/projects/13588168/papilio-fpga-shield-for-arduino?ref=category

    Is there such a thing as beginner friendly FPGA? What would one need to program the FPGA?

    You should ask him what beginner friendly means. Does beginner mean that you put in the work or they will teach everything to you? How long did it take anyone to learn Microsoft Basic in the 80's? I learned it but it took months to perfect it. How long will it take to learn FPGA?

    Does anyone know how to make a bit stream in FPGA?

    Here is a stepped tutorial in doing FPGA design:

    http://www.xess.com/appnotes/ise-10.php

    Here is a VHDL tutorial in learning FPGA:
    http://www.seas.upenn.edu/~ese171/vhdl/vhdl_primer.html#_Toc526061343

    Here is another tutorial:
    http://www.fpga4fun.com/PongGame.html


    Now do most people understand these tutorials?

    Obviously, you need SMD soldering skills (don't scorch the board or you will void the warranty) if you want to add SRAM. Beginners in SMD shouldn't try to SMT solder a $55 dollar board in learning because I think it isn't as easy as everyone says it is.

    Too bad sites like Nerdkits don't have any tutorials on adding SRAM or I might have bought one of their kits. These people come from schools and know more than me but don't want to teach concepts like SRAM either.

    No. I don't think these things are easy anymore. Is he going to come to your house and teach it to you?

    You should also be wary of claims. The kickstarter part that says "pledge 1200 or more" and he says, "I can't guarantee they will work..." is a risky proposition.

    If there are bugs in this system, how are you going to know?
  • LeonLeon Posts: 7,620
    edited 2011-07-24 08:06
    The tools are much more complex than the software used for program development, but using FPGAs isn't particularly difficult. One can flash an LED in a few lines of VHDL or Verilog, for instance.
  • alex123alex123 Posts: 102
    edited 2011-07-24 09:45
    Why an FPGA? Would the CPLD work for your application instead? You program CPLDs in VHDL or Verilog the same way as FPGAs.

    Good part about the CPLDs is that you don't need the config PROM. The Flash memory is built in the CPLDs so you're looking at a single chip solution vs. 2 for FPGAs. You still can find the PLCC packaged CPLDs so it's easier to solder them than fine pitch QFP's.

    Xilinx is the most popular but the development tools are not so cheap. I'm using an Altera CPLD (MAX II). You can buy a Jtag programmer (USB-Blaster) for $15 on eBay.
  • prof_brainoprof_braino Posts: 4,313
    edited 2011-07-24 10:13
    Thanks for all the answers.
    alex123 wrote: »
    Why an FPGA.

    To experiment with FPGA. Is there a beginner friendly CPLD? That might also be an interesting experiment.

    Further search shows that a Papilio FPGA board that is NOT an arduino shield already exists for $75.

    Word is that FPGA is useful for a specific, complex task you want to do in hardware. The example was get a gene sequence with a couple billion pairs and look for patterns. The task is to randomly break the data into chunks of a hundred and do pattern matching. The task can be programmed into FPGA, and you can turn a bunch of them loose on the search space and its really fast. This is what you get for your $75 FPGA setup.

    The analysis was that for a GENERAL task, you can add an $8 prop each time you need more resources, and program custom peripheral in software. This is a more reasonable general solution. I already have a whole bunch of props.

    The finding is that the FPGA approach in general and this piece of hardware specifically will be lower on the "to do" list until a need is found that calls for FPGA, or FPGA becomes much cheaper.

    Thanks again!
Sign In or Register to comment.