Resurrecting an MTX500.

About original Memotech hardware.
Post Reply
stephen_usher
Posts: 199
Joined: 27 Nov 2016 19:58

Resurrecting an MTX500.

Post by stephen_usher » 07 Apr 2019 18:45

Way back in 1988 I had my MTX500 hooked up to a BBC Micro via RS232, using the BBC as a terminal to a program on the MTX when suddenly the Memotech stopped working... dead!

On and off ever since then I've tried to revive it. The first things I did were to replace the Z80 and Z80 CTC as they were easy to find back then... No luck.

About 10 years ago I swapped out the video processor... nothing.

With the tools I had I couldn't get any further.

Anyway, recently I picked up an oscilloscope and with that I managed to bring a BBC Model B+ 128K back to life. So, now with the new tool it's time to look at the old Memotech again.

Of course, the first thing to look at is what do the address and data buses do when you switch the machine on. Well, not a great deal it seems, at least one the memory data bus. One or two of the memory chips may be dodgy but... When I put the scope on 74LS157 chip D6 pin 9 (MA1) I got this...
IMG_1409.JPG
IMG_1409.JPG (203.76 KiB) Viewed 626 times
That doesn't look right for an address line does it?! Especially when pin 10 of the same chip looks like this...
IMG_1410.JPG
IMG_1410.JPG (199.94 KiB) Viewed 626 times
So my first plan of action is to replace D6 and see if the fault changes or stays the same. Of course, I'll fit a socket at the same time.

If nothing changes then as it's on the memory bus side and there's nothing else there than the memory it's probably a memory chip fault so I'll replace all the memory.

I do have a set of OKI chips but I don't really trust them. Will any 6164 chips do? I have almost a complete set of 120ms ones I bought in case I needed them for the BBC B+ but only needed one in the end. I can get another (along with a socket) easily from RetroClinic.

Any thoughts?

Martin A
Posts: 414
Joined: 09 Nov 2013 21:03

Re: Resurrecting an MTX500.

Post by Martin A » 07 Apr 2019 21:31

Here's a display of the same chip/pin on a working MTX. It does look like you're not getting a true low.
Ads00031.jpg
74LD157 position D6 pin 9
Ads00031.jpg (33.71 KiB) Viewed 621 times
The chips don't have to be OKI, I've upgraded both of my MTX500s to MTX512 spec by replacing the OKI 32k parts with 64k, 150ns Samsung KM4164B parts. To make the upgrade work I also re-set the links at B7 and put a 16V8 GAL as a replacement for the PAL at A6.

The service manual with the link settings etc is on Dave's site. The PAL code is in the 2nd version of the operators manual (also on the site if you done have it) though will need "tweaking" to work with something like WINCupl.

If you have access to a chip programmer and can knock up a socket adaptor, it's also possible to bring up part of the system for testing without having any working ram. The test rom in the shot below was running on the MTXplus but if you need it, generating something similar for the MTX shouldn't be too hard, I'm pretty sure I've still got the source (somewhere).
Running ramless.jpg
Test rom running without requiring RAM
Running ramless.jpg (75.88 KiB) Viewed 621 times

stephen_usher
Posts: 199
Joined: 27 Nov 2016 19:58

Re: Resurrecting an MTX500.

Post by stephen_usher » 07 Apr 2019 21:43

Unfortunately I don't have any programmers for anything.

When I hoic out the 74LS157 I'll check the voltage on MA1 before I put the new chip in.

stephen_usher
Posts: 199
Joined: 27 Nov 2016 19:58

Re: Resurrecting an MTX500.

Post by stephen_usher » 08 Apr 2019 19:43

Having removed all the memory, MA1 is still bad, so it looks like it's definitely the 74LS157.

stephen_usher
Posts: 199
Joined: 27 Nov 2016 19:58

Re: Resurrecting an MTX500.

Post by stephen_usher » 08 Apr 2019 22:36

All but one memory socket now installed. The others and the replacement LS chip should arrive tomorrow.

User avatar
1024MAK
Posts: 563
Joined: 24 Dec 2012 03:01
Location: Looking forward to summer, in Somerset, UK

Re: Resurrecting an MTX500.

Post by 1024MAK » 08 Apr 2019 23:07

Good luck ;)

Mark

stephen_usher
Posts: 199
Joined: 27 Nov 2016 19:58

Re: Resurrecting an MTX500.

Post by stephen_usher » 08 Apr 2019 23:49

All ready for the sockets and chips... :-)
IMG_1411.jpg
IMG_1411.jpg (602.96 KiB) Viewed 605 times

stephen_usher
Posts: 199
Joined: 27 Nov 2016 19:58

Re: Resurrecting an MTX500.

Post by stephen_usher » 09 Apr 2019 19:29

Well, the memory and 74LS157 have been replaced but the machine's still not alive.

At least the address lines look OK now.

The machine produces a black screen and a continuous beep. The memory data lines most of the time don't show any data and often go continuously low of continuously high.

stephen_usher
Posts: 199
Joined: 27 Nov 2016 19:58

Re: Resurrecting an MTX500.

Post by stephen_usher » 09 Apr 2019 21:26

OK, IORQ looks weird, high-ish but with some square noise. Removing the CTC makes no difference. Replacing the Z80 makes no difference.

Otherwise I'm not getting very far. The address lines all look good and are going up and down, RD and WR seem to be running. Sometimes I get data pulses on the data lines sometimes not depending upon how the machine powers up.

Any pointers where to look next?

Both crystals are outputting the correct wave form. The TMS graphics chip is getting VERY hot but removing it makes no difference to the rest of the system electrically.

stephen_usher
Posts: 199
Joined: 27 Nov 2016 19:58

Re: Resurrecting an MTX500.

Post by stephen_usher » 09 Apr 2019 23:58

Address lines A0-A6 show activity, A7-15 do not. Could this be just the ROM start-up code getting so far within the first 256 bytes before crashing and restarting? But then I should see data on the data bus and I don't. (Well, I often see one edge and sometimes see multiple data on some data lines but never all of them.)

The data lines are sometimes continuously pulled high, sometimes pulled low but most of the time floating just under 2 volts (with a lot of noise) Almost as if both the memory and CPU are waiting for the other to put data on the bus or acknowledge it.

Post Reply