Prop2 FPGA files!!! - Updated 11 November 2017 - Version 27

195969798100

Comments

  • jmg wrote: »
    cgracey wrote: »
    I still need to get this BeMicro-A9 problem solved, somehow. I could just make two different images, but that seems ridiculous.


    Did you check to confirm the DIP sw is actually wired as expected ? - can you activate some other pin, based on the DIP setting to confirm - even using a similar equation syntax, in case Altera gets confused there ?

    The thing is, all the critical signals never change places, no matter the DIP_SW1 state. So, it shouldn't matter. Why it doesn't work is baffling.
  • YanomaniYanomani Posts: 386
    edited November 14 Vote Up0Vote Down
    cgracey wrote: »

    The thing is, all the critical signals never change places, no matter the DIP_SW1 state. So, it shouldn't matter. Why it doesn't work is baffling.

    Hi Chip

    When you say that they (critical signals) never change places, does it means that they haven't a single glitch when DIP_SW1 has its position changed?

    In other words, could any other statement in the logic equations revert the results in the next clock edge, immediately following the one that sensed (and was meant to react to) DIP_SW1 changing its state?

    Is it possible to single step the equations in the simulator, clock cycle by clock cycle?

    Only a stupid guess

    Henrique

    P.S. If the above guess was true, the resulting signals would toggle at the clock frequency.

    Please, forget it. Only a stupid guess! :blush:
  • Henrique, this top-level file is written in AHDL, so I don't know how you'd go about simulating it. It just seems too simple to not work.

    I've made a new BeMicro-A9 image with the SD card signal set permanently wired in. I'll put it up tomorrow for testing.
  • cgracey wrote: »
    Henrique, this top-level file is written in AHDL, so I don't know how you'd go about simulating it. It just seems too simple to not work.

    I've made a new BeMicro-A9 image with the SD card signal set permanently wired in. I'll put it up tomorrow for testing.

    Yes, that was going to be my next suggestion :) I will test it as soon as it's available.

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • cgracey wrote: »
    Henrique, this top-level file is written in AHDL, so I don't know how you'd go about simulating it. It just seems too simple to not work.

    I've made a new BeMicro-A9 image with the SD card signal set permanently wired in. I'll put it up tomorrow for testing.

    Yes, that was going to be my next suggestion :) I will test it as soon as it's available.

    Here it is:

    https://drive.google.com/file/d/1omGhklqFgAEEoR0jrSNupT_UUxO7Cye8/view?usp=sharing
  • cgracey wrote: »

    Thanks, I've been having a hard time with the DE2-115 and the SD card so I hookup one up externally to debug it but it's a hit and miss with some of the I/O pins. Some seem to be slow to respond so that they seem to work ok when you toggle slowly but never quite reach a logic low or high when being toggled as when the SD card is accessed. I've tried different pins and some are ok but others not so. This is using the add-on board from Parallax. But no matter, I will try out the CVA9 image.
    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • Peter, maybe I have some issues with the voltage/current settings of those pins.

    Okay, I just checked them for the DE2-115 and they're all set for 2.5V operation, with the Px pins being set for 8mA. Are those pins really 2.5V, or are they 3.3V? Do you see a constant slope when they transition?
  • cgracey wrote: »
    Peter, maybe I have some issues with the voltage/current settings of those pins.

    Okay, I just checked them for the DE2-115 and they're all set for 2.5V operation, with the Px pins being set for 8mA. Are those pins really 2.5V, or are they 3.3V? Do you see a constant slope when they transition?

    There is a 2-step slope I noticed but some pins like the clock pin seems to be ok. The data line to the SD seems to have trouble going to 0V, more like 1V but works ok if I just toggle it at a few 100kHz etc. I've tried different pins but now I'm on to the CVA9 which seems to work ok with the new V27z file.

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • Thanks for the explanation, Peter. I need to investigate what's going on there with the DE2-115.
  • I have moved these external SD pins around a bit and found for instance that P12 = CLK, P13 = MOSI, P14 = MISO works fine but P15 = CS only drives high to around 1.5V. I'm sure if I move this pin elsewhere it will probably work fine :) I will check again when I get back to it.
    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • .... but P15 = CS only drives high to around 1.5V...

    Does that maybe have a LED connected to it ?
  • jmg wrote: »
    .... but P15 = CS only drives high to around 1.5V...

    Does that maybe have a LED connected to it ?
    It's the DE2-115 breakout board, there's nothing connected to that pin or the other pins I've tried, with or without an SD card. My routines do drive LEDs but not on signal pins.

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • I've been having glitches while testing the CVA9, sometimes I can't get it to load but mostly I find that it just locks up even when it is just sitting there. What I think I would like is V27 for the CVA9 without SD mapping perhaps, then I could see if that's stable or not. Besides my CVA9 is plugged into a motherboard which also has two more SD card sockets and while I can test probe these sockets because they also have pin headers, the V27z mapping has made these slots inaccessible.

    In the meantime I will go back to V26 and test that.
    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • Peter, I will make you a v27 with the original pin mapping.
  • cgracey wrote: »
    Here is a new one, Peter, with the original pin mapping:

    https://drive.google.com/file/d/1-u8pggCf38aN9vudDy14pQnSX6V8mWd1/view?usp=sharing

    Thanks Chip, I will use that to compare to the other V27 versions too. The V26 version has been stable on the CVA9 whereas the V27z was always locking up at odd times as if there was a internal timing glitch.

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • cgraceycgracey Posts: 8,281
    edited November 15 Vote Up0Vote Down
    Peter, are you seeing problems with v27 on different FPGA boards? It seems maybe we've started having strange problems with I/O pins.
  • cgracey wrote: »
    Peter, are you seeing problems with v27 on different FPGA boards? It seems maybe we've started having strange problems with I/O pins.
    I'm going to check this properly for reliability on the CVA9 and also come back to the DE2-115 and exercise the I/O pins on that looking for problems.

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • Peter JakackiPeter Jakacki Posts: 6,613
    edited November 16 Vote Up0Vote Down
    I have been having problems with V27 on the CVA9, even with the zz version. If I ran a background timing loop reporting the current runtime in milliseconds, it somehow always ended up crashing around 750secs. But even without EXTEND loaded and just the kernel sitting there waiting for user input it would still lock-up.

    However, for a lark, I "0 CLKSET" in Tachyon and since I was already running at 921600 baud I simply changed the terminal to 230400 baud and there was my prompt. For another lark I decided to load my EASYFILE FAT32 module which I haven't been able to get going with V27 or even V26. Lo and behold, it works!

    Ok Chip, maybe there is a race condition somewhere that is only showing up with the CVA9 FPGA compile, but maybe it exists on the 123-A9 too. As you know, I'm doing a bit more than running a simple blinky loop so maybe this has't been noticed on the other boards. I will continue to find out more in the meantime.

    edit: The CVA9 still locks up at 20MHZ after about 50 minutes or so. Now I'm testing it at 5MHZ ($0F CLKSET) to see where it goes.
    MODULES LOADED:
    3C3C: EASYFILE.fth        P2 EASYFILE FAT32 SD FILE SYSTEM - 171115.0000
    1E80: EXTEND.fth          TACHYON FORTH EXTENSIONS for the P2 - 171124-0000
    Tue, 00 Feb 1997 00:00:00
      Parallax Propeller II  .:.:--TAQOZ--:.:.  V10171112.0000 V27 BOOT
    ----------------------------------------------------------------
    TAQOZ# C:
    Mounted 7803.AA00-F84E.1690          NO NAME     FAT32   4,119MB (512/cluster)
     ok
    TAQOZ# D:
    Mounted 6043.BCD2-3036.3334          NO NAME     FAT32   7,944MB (512/cluster)
     ok
    TAQOZ#
    
    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • Peter, are you using GETRND for any purpose? I'm just trying to think of things that have a long period before certain values come up.
  • @Peter,@Chip
    I've been running my collection of code on both P123-A9 and BeMicro CV-A9 with V27.
    I'm not seeing any lockups at 80MHz with 8 cogs running and lots of cordic,hub etc activity.
    One program is 250k and loads first time every time.
    I'll keep throwing code at them and see what turns up.
    Melbourne, Australia
  • There's no GETRND being used, in fact TAQOZ can just be sitting there waiting for console input and it locks up. But if I go back go V26 everything works just fine.
    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • There's no GETRND being used, in fact TAQOZ can just be sitting there waiting for console input and it locks up. But if I go back go V26 everything works just fine.

    It seems to lock up at a particular clock offset, right?

    I don't know what to make of this. Could you see if you have any trouble with the DE2-115 image? I don't know what else to try, at the moment.
  • ozpropdev wrote: »
    @Peter,@Chip
    I've been running my collection of code on both P123-A9 and BeMicro CV-A9 with V27.
    I'm not seeing any lockups at 80MHz with 8 cogs running and lots of cordic,hub etc activity.
    One program is 250k and loads first time every time.
    I'll keep throwing code at them and see what turns up.

    Maybe this has something to do with the SD card pins. I don't know. I'm hoping we figure it out soon.
  • cgracey wrote: »
    ozpropdev wrote: »
    @Peter,@Chip
    I've been running my collection of code on both P123-A9 and BeMicro CV-A9 with V27.
    I'm not seeing any lockups at 80MHz with 8 cogs running and lots of cordic,hub etc activity.
    One program is 250k and loads first time every time.
    I'll keep throwing code at them and see what turns up.

    Maybe this has something to do with the SD card pins. I don't know. I'm hoping we figure it out soon.

    Yeah, I will just keep plodding along and maybe I will stumble across the problem :)
    BTW, the z and zz versions would eventually crash.


    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • Peter, is the time at which it crashes proportional to the clock frequency? Or is there some randomness to it?
  • cgracey wrote: »
    Peter, is the time at which it crashes proportional to the clock frequency? Or is there some randomness to it?

    It seems to be kinda proportional, about 8 mins at 80MHZ but closer to 50 mins at 20MHZ ! But either way, with any V27, it crashes :(

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • cgracey wrote: »
    Peter, is the time at which it crashes proportional to the clock frequency? Or is there some randomness to it?

    It seems to be kinda proportional, about 8 mins at 80MHZ but closer to 50 mins at 20MHZ ! But either way, with any V27, it crashes :(

    Peter, could you disconnect your USB cable and know if it's still working okay? Something with that long of a period might be hardware related. Just thinking.
  • cgracey wrote: »
    cgracey wrote: »
    Peter, is the time at which it crashes proportional to the clock frequency? Or is there some randomness to it?

    It seems to be kinda proportional, about 8 mins at 80MHZ but closer to 50 mins at 20MHZ ! But either way, with any V27, it crashes :(

    Peter, could you disconnect your USB cable and know if it's still working okay? Something with that long of a period might be hardware related. Just thinking.

    I've tried a few things like that but then there's V26, rock solid, no glitches.......... I'm not ruling out finger trouble though, but I just can't seem to put my finger on it.

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • Peter, are you using interrupts? How about skipping?
Sign In or Register to comment.