• Please review our updated Terms and Rules here

Amiga 1000 Software Error

E

eight088

Experienced Member
Joined
Oct 31, 2018
Messages
158
Location
Perth, WA Australia
Hi all

I've been working on resurrecting an amiga 1000 for a friend, and after diagnosing some chip ram fault I have it to the point where it is mostly working. I've tried various games and while the odd few don't work (possibly bad disk or incompatible with the NTSC machine I have), most of them are working fine. However, when I boot into workbench 1.3, and try to open the disk directory I get greeted with a System Request Software Error - task held error. I've done a little research on this problem and can't really find anything to help with further trouble shooting, so thought I would ask the brains trust that form part of this forum!

The machine is a NTSC A1000, with an insider 1mb memory card which is plugged into the CPU socket, and also has a 256kb memory expansion cartridge which is plugged into the front of the machine. The memory expansion cartridge had some bad memory chips, which I have replaced.

I have run amiga test kit and it passes the memory test (1.5mb), but the CIA timers have a 1% fail so I was wondering if this could be a potential problem? (refer to attached pictures).

Any advice on how to further troubleshoot this will be greatly appreciated.

Cheers :)
 

Attachments

  • amiga 1000 error.jpg
    amiga 1000 error.jpg
    180.5 KB · Views: 6
  • atk cia.jpg
    atk cia.jpg
    97.3 KB · Views: 6
If you run an 'official' diagnostic and something is reported as failing, that would be the place to start anyhow.

This looks like the time of day (TOD) and it looks like things are running 'slow' as opposed to not working.

I wonder if there is a clock frequency that is out of specification feeding the CIA?

Dave
 
I will need to have a look at the schematics and bust the oscilloscope and see what it clock is being fed to the cia. I did try and swap the CIAs around but the error remained at CIAA.

Thanks Dave, I'll report back with my findings
 
Just to be sure: You are running the Amiga 1000 with Kickstart 1.3, right? Because you can not run WB 1.3 on older Kickstarts. That's the case with any classic version (except 2.1, which runs on 2.0x). You will always get errors like that when doing certain things otherwise. Also, did you try a different copy of the WB 1.3? This one could just have e.g. a damaged .info file, which confuses the icon.library when trying to show the contents of the disk.

The TOD timer being out of spec should not cause this. But you could simply test by swapping the two. They are identical.

daver2 said:
If you run an 'official' diagnostic and something is reported as failing, that would be the place to start anyhow.
Click to expand...
Amiga Test Kit is nothing "official". Could even be totally normal to get these results on an A1000. In fact, that failing test was only added in 2021 as: "Base CIA timer test on true detected VBL frequency".
 
Last edited:
Timo W. said:
Just to be sure: You are running the Amiga 1000 with Kickstart 1.3, right? Because you can not run WB 1.3 on older Kickstarts. That's the case with any classic version (except 2.1, which runs on 2.0x). You will always get errors like that when doing certain things otherwise. Also, did you try a different copy of the WB 1.3? This one could just have e.g. a damaged .info file, which confuses the icon.library when trying to show the contents of the disk.

The TOD timer being out of spec should not cause this. But you could simply test by swapping the two. They are identical.
Click to expand...

Thanks Tim, yes I have been using KS1.3

I agree that I don't think the timing error is a problem. I wanted to eliminate this as a possible problem though. I've seen the TODCIA signal comes from either TICK50 or /VSYNC. If it's coming from TICK50 (I'm in australia so 50hz as opposed to 60hz 110v) then that might be the 1% error that atk is reporting? This is selected by J2, but not sure exactly where it is on the board...

I tried removing the insider 1mb ram expansion, and after a power up it didn't fault when I opened the disk drawer. So maybe a fault with the insder board? I put the insider board back in, and it opened the drawer this time, but then immediately faulted.
 
In ATK it passes the memory test (I ran the test for about an hour with no problems). There are some dip switches on the board, but I can't find any documentation so don't know what they do.
 
Maybe give the RAM a test with MBRTest-2. It does a way more complex pattern test (make sure you enable all optins on the left side).
 
I tried MBRTest-2 (edit, oops, 1.3 - I'll see if I can find v2) but when I go to open the disk drawer I get the system request fault! I got the newer ATK 1.19 (was using 1.16) and still have memory passing but the CIAA timer fails (1% error).

Second edit - I just read the readme file (using winuae), and it looks like mbrtest will only test fast ram. The insider 1mb shows as slow ram so don't think it's able to test it.
 

Attachments

  • mbr2 system request.jpg
    mbr2 system request.jpg
    167.5 KB · Views: 3
Last edited:
MBRTest-2 can test all types of RAM. It's the only RAM tester I'm using ever since.

//edit: attached
 

Attachments

  • MBRTest-2.zip
    44.2 KB · Views: 3
Good news and bad news is it's all good so far! I'll let it keep running for a while and see if a fault pops up...

Might investigate that timing issue just to make sure that doesn't affect anything. I think CIAa is responsible for the FDD, so a timing issue with that might have something to do with it...
 

Attachments

  • mbr2test.jpg
    mbr2test.jpg
    219.6 KB · Views: 4
eight088 said:
I think CIAa is responsible for the FDD, so a timing issue with that might have something to do with it...
Click to expand...
Unlikely, since the error was gone once you removed the RAM expansion.

Both CIAs and Paula are doing floppy. CIA-A is the least important one here. But as I said already: simply swap the two CIAs and see if that has any effect. I would assume the error in Amiga Test Kit moves from CIA-A to CIA-B, but nothing else will change.
 
Timo W. said:
Both CIAs and Paula are doing floppy. CIA-A is the least important one here. But as I said already: simply swap the two CIAs and see if that has any effect. I would assume the error in Amiga Test Kit moves from CIA-A to CIA-B, but nothing else will change.
Click to expand...

I've already swapped them and no difference (post 3). I have a friend who has a retro chip tester so will try and test the memory chips in that and see what happens.
 
Ah, sorry. Did not read that. This means the CIAs are fine. So there's either an issue with the VBL frequency (but then, other results should be affected as well), or it is just how it is with ATK on the A1000. Since ATK is still in active development, you may want to contact the author about that.

If the memory tests ok, you could try playing around with the jumpers. Who knows what they are good for. :D
 
You must log in or register to reply here.
Back
Top