Page 17 of 23

Re: Resurrecting an MTX500.

Posted: 16 May 2019 12:55
by stephen_usher
OK, so that theory goes out of the window.

I may have to get the other machine down from the loft and test with the other CFX card to make sure that it's not the device.

Re: Resurrecting an MTX500.

Posted: 16 May 2019 12:59
by Dave
Have you tried more than 1 game?

How are you loading it/them ?

Re: Resurrecting an MTX500.

Posted: 16 May 2019 14:03
by stephen_usher
I'm loading them via the CFX from CF card.

It's possible that the loading system in the CFX is corrupting them. I've tried "REV.RUN", which is Andy Key's "Reveal" game and that runs stably for hours. "STAR.MTX" starts but the star field slowly gets "suck stars" and some of the display numbers get written in the wrong place. Eventually (after quite a long time) it crashes. Of course, this could also be a bug in the MTX file loading routines. Most of the early Continental Software games run fine, but seeing as they run on an MTX500 they're only testing the lower part of main memory.

When I boot into CP/M probably 50% of the time when it crashes it starts printing "ink" coloured square blocks which start off fully filled and then get more randomly pixelated until the machine freezes. I only ever get these problems when running code hosted on the CFX.

I s'pose I could re-write the CFX EEPROM. It's still on version 147 so it may be worth upgrading. I'm guessing that it's the centre chip on the board. I need to get the programmer out anyway to create a new GAL for the RS232 board.

Re: Resurrecting an MTX500.

Posted: 16 May 2019 14:32
by Dave
Can you try star.com from CP/M?

I have seen some odd characters (like underscores) when I run star.mtx here, but not with star.com. It is possible that the .mtx file on my CFX master is corrupt

Re: Resurrecting an MTX500.

Posted: 16 May 2019 14:50
by stephen_usher
I'll try, assuming that CP/M runs for long enough to type the name.

Re: Resurrecting an MTX500.

Posted: 16 May 2019 15:06
by Dave
No issues starting or running CP/M here

Please also try Spectron.mtx from SDX BASIC

Re: Resurrecting an MTX500.

Posted: 16 May 2019 16:59
by 1024MAK
Corruption of the display is corruption in the video DRAM. So could it be problems occurring during I/O operations?

Mark

Re: Resurrecting an MTX500.

Posted: 16 May 2019 16:59
by Martin A
Something to consider, the RUN type games usually load and run at #8100.
REV.RUN is 14k in size so fits entirely into the paged memory area between 8000 and BFFF
STAR.RUN is 18k in size so overflows into the unpaged memory at C000 onwards
CPM has all of it's drivers in the top 4k of the unpaged area at F000 to FFFF
CTC interrupt vectors in MTX basic are at FFF0

IF the issue is "something" affecting the high memory then the under 16k RUN type games that fit entirely into paged memory would run OK (Nemo, Toado, Hawkwars etc) The longer ones that overflow into high memory would be expected to have issues (chess, Knuckles, Reversi)

Re: Resurrecting an MTX500.

Posted: 16 May 2019 17:06
by Dave
Hi Martin,

I've tried a couple today. Star.mtx seems to have the problems that Steve described but Chess.mtx seemed OK

As we know, the .mtx files are the most "variable", I did do some tests a loooong time ago, it is unfinished, but I posted a "rough ' ready" list of my tests in a link towards the bottom of the CFX page http://primrosebank.net/computers/mtx/p ... fx/cfx.htm

I did list Star.mtx as working, but that might be because I didn't run it long enough at the time. At that point, I was just concentrating on whether the games loaded or not - I did not see or expect any to crash after loading

Errata - actually, I just checked the file again. I didn't mention Star.mtx in the Excel file, I did note that Star.com was working, which it is

Re: Resurrecting an MTX500.

Posted: 16 May 2019 17:37
by stephen_usher
Thanks... I'll do some more testing myself.

I could try replacing all the memory chips as I have a spare set, though as yet they are untested.