Strange behaviour

About original Memotech hardware.
Post Reply
acsi
Posts: 71
Joined: 16 May 2019 08:49

Strange behaviour

Post by acsi »

Well I fired up my original MTX512 (MTX500 MB with CFXII) yesterday and got the following strange behaviour.

On Boot I got the CFX II boot screen (Even if I hold down a key to go to CPM or MTX mode). After several attempts I got into CPM which booted fine but if I tried to type in DIR I ether get nothing or strange characters (ones I didn't type), same when finally getting into MTX mode. It also reboots after a minute or so.

Needless to say I didn't get to do what I wanted to do and now have some investigation to do once I get a chance.
Bill B
Posts: 599
Joined: 26 Jan 2014 16:31

Re: Strange behaviour

Post by Bill B »

Keyboard cable out of place by one pin?
acsi
Posts: 71
Joined: 16 May 2019 08:49

Re: Strange behaviour

Post by acsi »

:oops: Yep. I thought surely not as it wasn't doing that before... was it? No I was definitely using it fine... I think?

Keyboard connector was loose and sitting 2 pins over. Not sure how that happened.

I haven't powered it up to test yet. I'll save that for another day as bow that I've opened the case I'll look at replacing the video ram chip that's giving me dots in certain resolutions as well as redoing how I've mounted the RGB card on the video chip.
acsi
Posts: 71
Joined: 16 May 2019 08:49

Re: Strange behaviour

Post by acsi »

RGB board relocated to the underside of the board. I had mounted pins onto it could be mounted in between the socket and the video chip but found there it was very slightly too high and pressed against the keyboard.
And for a first time piggy backing the ram chips actually got me to the faulty one. One removal and socket install later and I now have a perfect display again.
acsi
Posts: 71
Joined: 16 May 2019 08:49

Re: Strange behaviour

Post by acsi »

Nope somethings still not right.
It'll work fine for a bit then start randomly resetting. Sometimes to a pink screen, sometimes it finds the boot drive, sometimes it's just scrambled.

I've reseated the PAL chip as tapping it would sometimes make it happen (might have just been a coincidence however as it seemed to get better but then started resetting again).

Trying to load PotHole Pete from Tape with the CFXII removed I get this (always the same except for the 4 digits after AF on the top right) and it seems quite content to sit on this screen without further resets.
IMG_2643.jpeg
IMG_2643.jpeg (362.14 KiB) Viewed 20831 times
With the CFXII fitted
IMG_2641.jpeg
IMG_2641.jpeg (486.72 KiB) Viewed 20831 times
IMG_2642.jpeg
IMG_2642.jpeg (420.14 KiB) Viewed 20831 times
User avatar
Dave
Posts: 1293
Joined: 11 Aug 2012 18:16
Contact:

Re: Strange behaviour

Post by Dave »

First thing that springs to mind is a possible bad connection between the MTX and CFX-II boards due to the edge connector, but it would be good to try and confirm whether the problems are with the base machine first.

The loading problem could be a bad tape, do you have other tapes that you can try without CFX-II attached?

Can you try the machine with just BASIC? Does it run reliably? Does "tapping" any of the chips cause problems? (You could remove/reseat all of the removeable chips, CPU, CTC, ROMs, VDP, GAL - being careful not to bend any pins obviously!)

regards
Dave
acsi
Posts: 71
Joined: 16 May 2019 08:49

Re: Strange behaviour

Post by acsi »

Hmm, OK so I'd already cleaned the contacts on the edge connector and for whatever reason didn't think to try another tape (possibly because it was midnight).
Tried several other tapes with the CFX unplugged and they work fine. Not good news for my PotHole Pete tape :(
Plugged the CFXII back in and instantly unstable with resetting and garbage screens. So I removed the CFXII and pushed down all the chips, the Alterra Max clicked back into its socket along with most of the other chips moving back into place.
It now seems perfectly stable again. :D

Edit: However testing my Pot Hole Pete tape on another MTX (Fitted with a MFX) works perfectly?
Martin A
Posts: 806
Joined: 09 Nov 2013 21:03

Re: Strange behaviour

Post by Martin A »

I had a quick look at the Pothole Pete code.

The program counter in that panel screenshot is at #8141, when it encouneted a #FF byte that triggered Panel.

That's in the middle of some sprite tables, the CPU should never have been running code at that point.

If the tape had loaded correctly the byte at #8141 should have been #55, there is a #FF at #8142 however. That suggests the tape loading missed a byte when loading the data block. and likely near the start, or the first instruction in the code would have jumped over the tables instead of trying ot process them!

I can tell from the panel data at FFE0 that the CTC vectors are correctly set for the tape system. the two #53 #08 pairs are the ROM routine "Toggle" used for timing the length of pulses. So the OS side of things looks fine.

WHY that happened on one MTX and not the other I don't know. Putting it some to borederline signal levels , and the diffecences between boards, feels like a cop out.
acsi
Posts: 71
Joined: 16 May 2019 08:49

Re: Strange behaviour

Post by acsi »

So I had a bit more of a play trying adjustments to the volume and even replacing the CTC chip as I have a spare and the Z80 with a third one.
The only change I could get was where it would reset or come up with the pink screen with the PC being 9101 and FF.

I also could not get Kilopede to load but all the rest of my tapes did. Tried to load it on another MTX and it worked fine.

I know the tapes are around 40 years old and maybe this MTX has some extra noise in the circuit or the EAR jack contacts are not conducting as well (dirty, corroded or have lost the pressure)?

Edit: I tried to load Kilopede on the volume level that worked on my other MTX and it found the header but the name was random characters.
Martin A
Posts: 806
Joined: 09 Nov 2013 21:03

Re: Strange behaviour

Post by Martin A »

The edge detection circut on the MTS is based on the 741 op amp. See page 226 of the Phoenix manual.

The final stage is a CMOS exclusuve or gate to set the level for the CTC, the rest is analogue and I don't have a clue how it works.

I would hope that it was designed to be tolerant of component differences. But ...
Post Reply