Resurrecting an MTX500.

[1024MAK]

Never mind that, my brain is still trying to work out how Stephen got a cable to go inside the front of his ‘scope. Look, you can see it through the screen!

Mark

[Dave]

Cable?

that's nothing - looks like there's a whole desk in there !

[stephen_usher]

Hehe.

Good to know that you're getting a similar waveform. I've replaced the LS00 and LS04 anyway, just in case. Only 14 more chips to go (other than the VDU RAM, PAL and ROMs) until everything's swapped!

Well, after removing the 741 I popped the 4070B back in and removed the CTC. It started (black screen+beep) the CPU OK (well, until the 4070B finally burnt out). I'm getting a very intermittent and extremely short ROM select. OK, sometimes it is a proper drop for a couple of microseconds but usually it's just a spike, which doesn't always get down to 0V. This is being generated by the PAL isn't it? Could that be the faulty item?

[Martin A]

Neither of my boards has the original 14L4 PAL fitted, as they're both been upgraded from 32 to 64k.

However, this is the waveform for the /CEA rom select output while sitting at "Ready" from the replacement ATF18V8B-15 doing the memory decode.

The CEA output from the GAL

pin17.jpg (31.8 KiB) Viewed 6322 times

The MTX had a tendency to crash each time I put the probe on the output pins, so it's no so easy to work out what was happening when that trace triggered. The other 3 outputs had similar sort of noisy shape, with no obvious spikes.

[stephen_usher]

Well, it's nothing like that. Mostly a slightly noisy high with a rare spike towards zero, often not getting that far. Rarely it can produce brief nice drops, but they're nowhere as long as even the shortest of yours.

Hmm... Any recommendations for a good GAL/ROM programmer for Windows 10?

P.S. This is what the ROM select line looks like "on a good day"...

IMG_1451.JPG (630.42 KiB) Viewed 6316 times

[stephen_usher]

Yay! The Z80 CTCs (2, which was half of Farnell's remaining stock) and five CD4070Bs (minimum order) have arrived. So, more things I can replace this evening.

[stephen_usher]

One step forward, two steps back.

After fitting the new 4070B there are no negative voltages from 10E and it's running cool.

Still only very rare ROM select pulses, so that is possibly the issue, which would point towards the PAL.

The major step back is due to the crocodile clip on the scope probe jumping off the modulator case, which I'm using as 0V, hand sliding across the top of the PAL board! (The probe ground cables are insanely short.) This shorted something and now the display doesn't sync correctly. Aaaargh!

Anyway, I think that I'd better get a GAL programmer.

[Martin A]

Device programmers:
The first one I had was an ART Epp-3 that came from Maplins (remember them ?) but that's parallel memory only, and been obsolete for 10 years or more, the software is designed for windows 98 I think, but ran on XP.

For the MTXplus project we used the MiniPro TL-866A, that's got a massive list of devices,for memory logic and microcontrollers. I've had no issues here using the software with windows 10. That's been replaced by the TL-866II plus which seem to be on EBAY, Amazon and the like for £50 ish

The TL-866A had one notable omission - The ATV22V10 isn't supported, that seems to be the only 24pin SPLD officially still in production. To program those I ended up with a much more expensive VP-290. Again massive list of devices, and no real problems running under windows 10. Now out of production and replaced by the VP-299 which seems to retail at £120 or so.

Both programmers do the ATV16V8B which is a suitable replacement for the PAL14L4s used by Memotech, there's little advantage in getting the more expensive programmer.

None of these will program the big 84 "pin" CPLD's used in the MTX plus and CFX2 for those I have a USB blaster close for Altera Parts, and snappily named ATDH1150 for the Atmel ATF equivalents. The software for those is freely available, but Altera's is definitely the better, but it all runs on 10, but Atmel's parts are cheaper.

Though Atmel's free WinCUPL will compile for the ATF16V8, ATF22V10 and ATF150x CPLDs, it can be flaky on 10, and needs frequent saving of files just in case.

[Martin A]

Still only very rare ROM select pulses, so that is possibly the issue, which would point towards the PAL.

IF the memory is misreading, from either faulty decode, or duff rom, or address/data bus shorts, the MTX could be running random code from ram, which could account for the lack of rom select signals.

[stephen_usher]

Still only very rare ROM select pulses, so that is possibly the issue, which would point towards the PAL.

IF the memory is misreading, from either faulty decode, or duff rom, or address/data bus shorts, the MTX could be running random code from ram, which could account for the lack of rom select signals.

I've double checked, there are no shorts on the data or address buses. All the chips connected directly to the data and address buses have been changed, including all of those to do with the address decoding. The only chip left is the PAL.

Removing the ROMs completely and I still get an intermittent/weak/short ROM select pulse.

It's quite possible that the ROMs were fried when I had those glitches which took out the LS138s and LS157s, but the ROM select is iffy even without them.