Shop OBEX P1 Docs P2 Docs Learn Events
Problem with SX-Key on *one* computer — Parallax Forums

Problem with SX-Key on *one* computer

peterjpeterj Posts: 24
edited 2004-10-06 22:54 in General Discussion
I recently purchased the SX-Tech Toolkit and serial transfers are *very* unstable on my primary workstation, an A-Bit NF7S running Windows XP. The serial port in question works fine with my modem, and the SX-Key seems to work fine on other computers. When I go to program the chip, the transfer will lock up midway through, or sometimes the software will fail to recognize the SX-Key entirely. I was able to program the chip once on this machine, but only after about a dozen attempts.

Fortunately I have a spare computer on a KVM at the moment that I can use to ptogram the SX-Key, but I'd really like to find a solution for my primary workstation. Has anyone else had any similar problems? Any potential solutions or work-arrounds?

-p.

Comments

  • PJMontyPJMonty Posts: 983
    edited 2004-09-15 22:56
    Peter,

    1 - On the other machine(s) that work ok, what operating system (Win98, XP, etc) are you using?

    2 - XP seems to be problematic for the SXKey. One common solution is to lower the FIFO buffers from the typical 14/16 to something like 4/4. Try lowering them substantially and see if that improves your situation.

    I have attached a screenshot showing my settings. Don't take this as gospel, but merely as a starting point. You can also try unchecking the "Use FIFO buffers", and remove them from the equation completely.
      Thanks, PeterM
    Serial FIFO settings.jpg
    636 x 285 - 26K
  • Josiah WoodJosiah Wood Posts: 16
    edited 2004-09-16 21:50
    I've been having the same problems, and I tried modifying my FIFO settings as PJMonty suggested, but nothing improved.
  • peterjpeterj Posts: 24
    edited 2004-09-17 01:11
    Of the machines that do work, one is running Windows 98 and the other is running XP. It does not seem to be *specifically* an issue with Windows XP. Rather, it seems to be a combination of XP and certain hardware.

    I tried reducing the number of FIFO buffers, as well as disabling them entirely, but this did not help in any way. [noparse]:([/noparse]

    Any other hints? Is there any official word on this from Parallax?

    -p.
  • richlrichl Posts: 3
    edited 2004-09-21 18:53
    My Windows 2000 machine (2 GHz) does not work reliably, but the XP (200 mhz) one does work. When doing serial I/O, most of the cycles of the machine seem to be used by the downloader. This gives the impression that they are using a loop to test the serial I/O instead of using events and threads. The serial I/O may thus be sensitive to the speed of the computer. Since the assembler is very fast, using an old dog of a computer to run SX-KEY is not much of a penalty.


    - rich
  • peterjpeterj Posts: 24
    edited 2004-09-21 20:49
    Interesting. My failing machine is also rather fast (athlon 2400) and the working machine a bit slower (1 Ghz Intel)

    -p.
  • Josiah WoodJosiah Wood Posts: 16
    edited 2004-09-21 23:10
    Would one of the USB adapters solve this problem? Such as the "Parallax USB2SER Development Tool" or "FTDI US232B USB to Serial Adapter" from Parallax?
  • peterjpeterj Posts: 24
    edited 2004-09-22 16:34
    It's possible that a usb->serial dongle would get arround the problem, but that is somewhat of a hack. I'm hoping that Parallax will provide an updated version of the IDE that fixes this problem on high-speed machines before my spare machine has to go away. Failing that, I'll probably just scrounge an old machine. It's pretty easy to find a free PC in the sub-300 Mhz range these days (I gave away a bunch of machines like that last time I moved) and I do have a free port on my KVM so I don't need to bother with the extra keyboard/mouse/monitor.

    -p.
  • Jon WilliamsJon Williams Posts: 6,491
    edited 2004-09-22 16:50
    We have found an IDE issue with very high-speed machines -- I don't know that it will address your specific problem (I run on a 2.2 gig P4 without any issues), and yet it will be corrected shortly.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Jon Williams
    Applications Engineer, Parallax
    Dallas Office
  • Michael ChadwickMichael Chadwick Posts: 80
    edited 2004-09-30 17:12
    Any ETA on the revised IDE with the bug fix for fast machines?
  • Jon WilliamsJon Williams Posts: 6,491
    edited 2004-09-30 22:23
    Not yet -- it also has to do with the size of the program in the Javelin (virgin EEPROMs have bigger problems). It is being worked on now and will be posted as soon as it has been tested and deemed to be reliable.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Jon Williams
    Applications Engineer, Parallax
    Dallas Office
  • peterjpeterj Posts: 24
    edited 2004-10-01 00:21
    Can you elaborate on this? Are you referring to the SX code that implements the Java byte-code interpreter? Is the problem in the assembler (i.e. the assembler not generating proper machine code) or the driver that programs programs the chip from the HEX file (i.e. will any large hex file have difficulties?) Or am I misunderstanding entirely?

    -p.
  • Jon WilliamsJon Williams Posts: 6,491
    edited 2004-10-01 00:27
    The problem is in the JIDE, having to do with timing the ID process. It will be corrected shortly.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Jon Williams
    Applications Engineer, Parallax
    Dallas Office
  • TreatTreat Posts: 1
    edited 2004-10-06 21:04
    News on the correction would be welcome. I'm trying to run SX-Key on a relatively slow (966MHz) processor with XP Service Pack 2, and I've encountered every issue identified above. Ability to program, let alone debug, is completely compromised. Thanks,
  • Jon WilliamsJon Williams Posts: 6,491
    edited 2004-10-06 22:54
    Forgive me ... I posting information regarding the Javelin IDE (JIDE) here -- too many long nights.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Jon Williams
    Applications Engineer, Parallax
    Dallas Office
Sign In or Register to comment.