A query on "Errors in memory testing" (ACA 1234 + ACA 500 Plus)

Caution: Non registered users only see threads and messages in the currently selected language, which is determined by their browser. Please create an account and log in to see all content by default. This is a limitation of the forum software.


Also users that are not logged in can not create new threads. This is a, unfortunately needed, counter measure against spam. Please create an account and log in to start new threads.

Don't Panic. Please wash hands.
  • Just a quick addition: Even just launching "MuScan" instead of MuForce will cause the Freeze with MBRtest-2. So a direct conflict of MuForce with MBRtest-2 is excluded.

  • That sounds like you're still in the cloaking device menu. You missed the "ESC"-press that takes you back to the main menu.

    I'm now pressing escape and still have an issue with 3.1 too in 68000 mode, but this time the screen does not even go white.. If I boot 3.1 without enabling 68000 mode it boots fine, the same goes when removing the ACA1234 attached. It only seems to have issues booting once I enter 68000 mode. I've even purchased a fresh CF card and installed 3.1 on it and still the issue persists, so I seem to be able to rule out any workbench configuration issues.

  • Could it be that you've selected overclocking beyond a compatible frequency for 68000 operation?

  • If the Amiga Test Kit AND the MBRtest-2 software are both 'thrown out' as containing 'suspect code'...
    I could fine with that... as long as there is an alternative software with similar functionality to permit testing.


    If however, there is only one RAM test software and a new hardware product comes onto the market and it doesn't work with that RAM test...

    It's not acceptable to say all RAM test software is bad and leave it at that.


    We can't just 'eliminate' everything that is able to show a problem.

    That would be like gouging your eyes out so you can't see the issues, problem solved. :D


    So, what I'd like to know is what Amiga test software is "guaranteed" to work with ACA1234 ?


    Thank you,

  • If the Amiga Test Kit AND the MBRtest-2 software are both 'thrown out' as containing 'suspect code'...
    I could fine with that... as long as there is an alternative software with similar functionality to permit testing.

    AmiTestKit is not "out" per se. It does not show any errors after the CPLD update. And the only reason why I wanted to use a different program in addition to AmiTestKit in the first place was that it was buggy before, attempting to test memory with an ACA500plus that was never added to the freemem list. Luckily, the authors are there, and they fixed it. This time, the test was actually right and it displayed an error that was really there. It was just


    If however, there is only one RAM test software and a new hardware product comes onto the market and it doesn't work with that RAM test...

    It's not acceptable to say all RAM test software is bad and leave it at that.

    I agree - and I don't even exclude MBRtest-2 completely. I only observe that the "freeze" is also happening on the B1230-IV if the MMU has been activated with MuTools, and I also see on the logic analyzer that the CPU just hangs and doesn't even attempt to use the bus (AS line stays high). This is a clear indicator of a software problem.


    Everybody seems to agree that the B1230-IV is the big role model in terms of compatibility. I accept any comparison with that card, as long as the software is identical. Still, as you (and others) have pointed out: Testing is an art of its own. And it's always a good idea to question the test method as well, not just run after results without looking at how they were obtained. Feel free to use any memory test software - if it works on other hardware, it shall also work on the ACA1234 and other iComp products.

  • Everything is set at 7MHz.

    Please double-check the global settings: If the max. frequency selected there is higher than the actual frequency, it will be chosen during Kickstart unpacking to reduce boot time.


    If that doesn't help, I'll need a more detailed description of what goes wrong. So far, I only got "issues" and "does not work"-like descriptions, which doesn't really help with my imagination of what actualöly goes wrong.

  • Please double-check the global settings: If the max. frequency selected there is higher than the actual frequency, it will be chosen during Kickstart unpacking to reduce boot time.


    If that doesn't help, I'll need a more detailed description of what goes wrong. So far, I only got "issues" and "does not work"-like descriptions, which doesn't really help with my imagination of what actualöly goes wrong.

    That was it, I've now successfully updated the CPLD on the ACA1234. It's probably worth adding a note in the wiki to take the CPU speed in global settings down to 14MHz

  • It's probably worth adding a note in the wiki to take the CPU speed in global settings down to 14MHz

    I really can't repeat all kinds of setting-recommendations in every possible spot. The information about max. overclocking frequency is in the ACA500plus description, and it should only be set to a higher frequency for testing and only left at a higher speed if it's deemed to be successful.


    The CPLD update is pretty easy to do, but having too many unnecessary instructions may put off other people. So far, I'm very pleased with the outcome: Lots of reports of successful updates, even when the card appeared bricked at some point. This has made me confident enough to continue shipping the ACA1234 today, including the JTAG cable by default, and roll out the next CPLD update when our internal tests are completed.

  • We're almost ready with the current update. Prepare for a surprise: A so-far undocumented feature. Sorry it took so long, but this time we wanted to be sure that everything works in all eight modes of operation (four speeds on two target computers), and with all CPLD versions.

  • Was tinkering in the lab today and noticed my ACA1234 was still bricked from the last round of testing...
    And noticed the new release V3 CPLD code and 1.4 update for the ACA500+
    Happened to have my new A2000 board in my hands... so decided to use it to do a remote un-brick...
    Went smooth...I won't post the pics...would be heart stopping... :D
    A2000 motherboard sitting on anti-static foam block on a pizza box, perched over my Checkmate with cables dangling and a picoATX PSU bodged into the A2000... Just not a pretty sight...but my ACA1234 is working great now and apparently my A2K parallel port is working too :D


    The new features in the ACA500+ menu, really nice..

    And I've not gone back and run all of the software I was having trouble with...
    But I did run a few of the games that stood out in my mind, and I have to say huge improvement...
    "Agony" now running 'perfectly' to my eyes..


    I'm guessing your new memory controller has solved many of the issues I was seeing before.
    I've been running for hours without a crash...just playing tunes while I solder away, so far all seems quite stable...

    At first glance, much improved...

  • I'm guessing your new memory controller has solved many of the issues I was seeing before.

    The new memory controller was a necessity to shrink it's size - timing-wise, it's pretty much identical to the old one, but it makes room in the CPLD for other fixes. The main thing that CPLD V3 fixes is 33MHz operation on the ACA500plus, which *sometimes* did not read the correct data. So the main change is in the host computer interface, not the memory controller :-)


    Note that the ACA1234 is multiple accelerators in one, as each speed and host computer have their own optimizations, while the memory controller only makes a difference between "50MHz or not". Those who have run "bustest" might even have noticed that the V3 CPLD is a bit faster on chip ram at 33MHz on the ACA500plus than it was before. Just a tiny bit, but measurable.

  • Yay, so I finally received my JTAG cable and did the CPLD update. After that I did some tests with the games that used to crash at 33 MHz and with Amiga Test Kit. Looks like everything works fine, at least for now. I'll be doing more tests soon. Thanks Jens for releasing the fixes!

  • Hello!,


    Thanks Jens for the updates and the advice to make them.


    I comment that I have been able to update the CPLD successfully.


    The paralell to JTAG cable has not yet reached me, due to the poor service that the Argentine post has. I'm still hopeful that one day it will come, but I couldn't wait any longer. I have built the cable following the instructions on the wiki.


    http://wiki.icomp.de/wiki/Budd…your_own_programmer_cable


    First I performed the CPLD version 1 update, with the ACA 1234 mounted on the ACA500 plus, after updating the ACA500 plus firmware to version 0.14. This update was successful. Obviously the memory error problems disappeared in the test of the Amiga Test Kit v1.18


    The other day I saw that the version 3 update for the CPLD was already available and I proceeded to update. Then I installed the new firmware and tested the ACATOOL 2.9 application. (Here I had a hard time finding the successful connection of the JTAG, but after some tests it connected well and I stayed like a statue for a while).


    The ACA1234 accelerator works very well with my ACA500 plus, so I'm thinking of buying another one to leave connected to it.


    Now I have permanently connected the ACA1234 to my A1200 (which has just arrived after a small repair).


    I have installed the Workbench 3.2.1 on a Compact flash card put into the CF adapter of the ACA1234, and it really is very fast. I have mapped Kickstart 3.2.1 with ACATOOL 2.9's maprom option and it works great.


    These days I am going to continue testing programs and games, but so far everything is perfect.


    Thank you for everything, I take this opportunity to wish you a good end of the year and an excellent new year

  • The last reply was more than 365 days ago, this thread is most likely obsolete. It is recommended to create a new thread instead.