My next challenge: Memotech MTX500.

[stephen_usher]

It's all brand new and tested by the vendors.

[PhilC]

I would still check all address lines through the chips as the majority of 8 bit failures I'd dealt with was due to ram.

When working on an 800XL youd also get pulses for say 10 seconds on the address lines and then nothing, that was due to the peripheral interfaces. And obviously, if you can burn yourself a new rom its worth a try.

[stephen_usher]

I've checked the address lines and they're humming away nicely.

I removed the memory and the electrical activity (and the system) is identical.

I have noticed that the IORQ line looks strange, high-ish but with some square noise. The only thing I can see connected to that line other than the CPU is the Z80CTC and nothing changes if I remove that. I've tried replacing the Z80 and there's no difference.

I'm a bit stumped.

[PhilC]

Maybe post a circuit diagram?

[stephen_usher]

See pages 228 and 229 of the Operating Manual http://primrosebank.net/computers/mtx/d ... Manual.pdf

(I bet you wished Atari had produced a manual like this for the ST. )

[stephen_usher]

Sometimes the machine resets into a strange state where the sound chip seems to become a bus analyser.

One of these times I managed to video the data line output, probing each of the data lines in turn. Note how the sound changes as I touch each of the data lines with the probe.

Also note some of the strange output on the 'scope. The behaviour changes each time the machine is reset using the reset keys on the keyboard and doesn't follow the memory chips if I swap them about.

Other than the CPU, Z80CTC, video processor and ROMs the only chips on the data bus are three 74LS244 buffers and another LS chip running the input form the parallel port. I'm guessing that one of the 74LS244 chips is faulty, producing the strange voltages/put-up/pull-out on the bus.

Here's the YouTube clip: https://youtu.be/5VVGaIJvrm8

[stephen_usher]

Well, I successfully de-soldered the three 74LS244s, installed sockets and tested without any of them in the circuit, isolating everything from the main data bus except one 74LS373, which is for the parallel port input, the ROMs and the Z80 (I removed the Z80CTC and TMS graphics processor)... No change.

So, it's either the 74LS373 or the ROMs... Probably.

I'm going to see if I can get a spare set of ROMs from somewhere. Unfortunately they're non-standard TI ones with a strange pin-out apparently.

[stephen_usher]

Well, I'm still working on this.

After at one point somehow getting 12V on a couple of address lines (I have no idea how to this day as there are no shorts) I've had to replace the CPU, Z80 CTC and many of the other LS chips. Still no go.

I've possibly traced the problem down to the PAL chip which generates the ROM select lines, except this one isn't most of the time and when it does it's a very short spike down to nearly 0V rather than a nice square signal. Thankfully there's a GAL which is pin compatible with the PAL chip on the market and the programming information is available in the later version of the machine manual. (Don't you miss manuals which contain full circuit diagrams, chip information and PAL code for all motherboard versions and peripherals?)

Given the 12V issue I'm assuming that the ROMs are probably at least partly fried.

Unfortunately the ROM sockets as built don't support EPROMs so I'm having to do a bit of a modification as detailed here http://www.primrosebank.net/computers/m ... eplace.htm.

I have the new EEPROMs, programmer, sockets etc. I'm just awaiting the arrival of the GALs (coming from Mouser in the USA so taking a while) and some wire wrap wire, which is surprisingly difficult to find.

[stephen_usher]

Oh, I forgot to add that I'm having to fix the PAL video daughter board after the 'scope probe ground wire crocodile clip jumped off the modulator shield I was using as the ground and bounced over the solder side of the board. Now the display doesn't sync properly. Hopefully it's not the 1889 chip as those are not easy to get. I've sourced all the other components.

The probe ground cable is insanely short, which is why it got pulled.

[stephen_usher]

I had a breakthrough on this project today.

Having traced the ROM select problem back to one of the flip-flops in a 74LS273 I swapped this over and suddenly the machine came to life, well mostly. The PAL video adapter board was still shot and it was obvious from the display that at least one of the video RAM chips was faulty.

At least the diagnostic ROM I'd downloaded and burnt showed that there were no main memory errors.

Anyway, after changing the two CD4016B chips on the PAL board the video display output worked again. Phew!

So, now I've ordered some 4116 RAM chips and I'll replace all the old VRAM and put them in sockets. (I know Coldfire doesn't like them but it's best on old kit as there are only so many times you can heat the PCB before it dies.)

Anyway, here's what it looks like running the diagnostic ROM main menu:

IMG_1455.jpg (256.89 KiB) Viewed 5019 times

The character corruption makes it almost look like a Latin language of some kind at times.