exxos's Atari, Amiga & retro forum

Well, looked through the MegaST service manual.


Only memory strikes me as something that could start cold good and then "GO BAD" of course I could be wrong.

I can replace the MMU and the Shifter easy enough one at a time.. see if that comes up with anything.

Is there a way to pull a ROW of ram at a time?? to test with only say 1mb or 2mb?? or will pulling any ram just stop the machine?

James

Is there any existing documentation on Errors received using a Diagnostic cart?

I am getting E9 or EB buss errors etc.. but no idea what those mean?

Now I can get the cart to run.. it will show only 1m of the 4m.. but will continue to work.. and I can test many items that work. So I think it's memory addressing.. but something that when the machine is truly cold works for a limited time..

I could try taking chips from other machines and replace a row at a time?

I have looked for bad joints, but not found anything that stands out.

I think Shifter is out.. it's not a video issue.. I can sub out a MMU easy enough.. will try that I guess.

James

@Bikerbob have a look at the Mega ST service manual for codes E8 and E9. I would look for you but I'm out at the moment. It's in the latter half of the manual.

MEGA ST service manual has loads of errors codes, such as the ones on page 3.17 etc

Looking at what @exxos posted I'd try reseating the glue and mmu but it's likely ram related.

exxos wrote: 29 Oct 2021 18:12 MEGA ST service manual has loads of errors codes, such as the ones on page 3.17 etc

Thanks Phil and exxos --- OK, I will now begin my searching and math :)

I6 Testing Bus Error handling
No (I6) error

Keyboard failed, connect RS232 terminal

Mega and ST Field Service Diagnostic Rev. 4.4
1990, Atari Corp.
1M RAM Keyboard rev. 0 60 Hz Version 2.6 USA NTSC

This TOS version is unlisted in this version of the diagnostic


R RAM O O.S. ROM C Color
K Keyboard M MIDI S Serial Port
A Audio T Timing D DMA Port
F Floppy Disk P Printer/Joy Ports H High resolution

J Hard Disk Write/Read (use SH204/SH205/Megafile)
G short BLiT test Y long BLiT test
L Real-time Clock
X Expansion Connector

Q Run all tests
Z Run internal tests (R,O,C,K,A,T,L,G)

E Examine/Modify memory
B Set RS232 rate
V toggle video output--50/60 Hz
? Help

Enter letter(s), and RETURN: R


Testing Ram
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *R2 RAM addre
ss=0801CA W:01CA R:00CA bad bits: 8
R2 RAM address=0801C6 W:01C6 R:00C6 bad bits: 8
R2 RAM address=0801C2 W:01C2 R:00C2 bad bits: 8
R2 RAM address=0801BE W:01BE R:00BE bad bits: 8
R2 RAM address=0801BA W:01BA R:00BA bad bits: 8
R2 RAM address=0801B6 W:01B6 R:00B6 bad bits: 8
R2 RAM address=0801B2 W:01B2 R:00B2 bad bits: 8
R2 RAM address=0801AE W:01AE R:00AE bad bits: 8
R2 RAM address=0801AA W:01AA R:00AA bad bits: 8
R2 RAM address=0801A6 W:01A6 R:00A6 bad bits: 8
R2 RAM address=0801A2 W:01A2 R:00A2 bad bits: 8
R2 RAM address=08019E W:019E R:009E bad bits: 8
R2 RAM address=08019A W:019A R:009A bad bits: 8
R2 RAM address=080196 W:0196 R:0096 bad bits: 8
R2 RAM address=08014A W:014A R:004A bad bits: 8
R2 RAM address=080146 W:0146 R:0046 bad bits: 8
R2 RAM address=080142 W:0142 R:0042 bad bits: 8
R2 RAM address=08013E W:013E R:003E bad bits: 8
R2 RAM address=08013A W:013A R:003A bad bits: 8
R2 RAM address=080136 W:0136 R:0036 bad bits: 8
Fail at cycle:000001

OK so from what I am getting from the manual 0-1fffff is bank 0 on a Mega4.

From what I am reading I am still not sure how I identify which row(s) are bank 0 OR 1 physically on the board and is there a way to narrow it to a chip?

My addresses would seem to be bank 0 - I am guessing at the start of it.. but dont know where I would look.

James

Thanks Exxos, very helpful.

As the board heats up.. or gets tested more, I error out before even being able to run the ram test..

I have pulled all ram chips and reseated - no change

I subbed out the MMU - no change

3 different CPU and put back the 6 chip TOS 1.4 - no change.

EB and E9 errors now.. basically I can get the Diag screen to come up.. and it goes funky (screen) and gives EB and E9 errors.

I guess I could swap the GLUE - Phil thought it could be an issue.

edit - GLUE - no change

James

.--EB Bus error. Generated internally by the 68000 or
externally by GLUE. Usually caused by device not responding.
Displays the address of the device being accessed.

OK, so again sorry if I am not as advanced as many here.. but is the address something I could look up? or would I need to know the internal workings of the machine. If I posted the EB errors (I need to let the machine cool now) would that be something easily figured out?

James

edit was able to copy this from the Terminal log

EB Bus Error Access Address: 423D02 Program Counter: FA3E76
EA Address Error Access Address: 01789D Program Counter: FA3EF0

E9 Bad instruction fetch