Since then, however, the machine has spent a lot of its time stored in a cupboard because … reasons. The other day I dragged it out, set it up and was pleased to find it was working fine. And then I tried to fix it.
Let me explain.
I’d rashly decided to buy a new toy to go with it – a Gotek floppy drive emulator. But I had a nagging feeling that this motherboard has issues with the floppy disk interface. Somewhere in the dark recesses of my memory I recall trying to use floppy drives with it and them refusing to work. As I had other options, I never followed up on this.
Anyway, the Gotek arrived and didn’t work. I mean, the Gotek itself worked fine, but the Beeb wouldn’t talk to it.
One potentially suspect part was a 7438 logic chip that sits between the WD1770 disk controller and the socket. I figured I’d replace this. At the same time, I decided to replace the socket for the machine operating system (MOS) chip. Acorn used cheap, single-wipe sockets for this and the sideways ROM chips. The MOS chip itself refused to sit confortably in the socket. And a new multi-OS board I’d bought to replace the MOS chip wouldn’t work at all in this socket. I determined to replace this and the 7438 one with turned-pin sockets.
Which I did.
Bad to worse
Did the Gotek work after that? No. In fact, the BBC Master doesn’t work at all now.
After flipping on the power switch, I get a continuous, plaintive tone from the speaker and nothing much on the screen.
The continuous tone is the result of random data being fed to the sound chip. Normally, this is replaced by organised data coming from the system 6522 VIA chip – a moment’s silence followed by the Beeb’s signature ‘bip’. For that to happen, the operating system needs to be running normally. Clearly, it’s not.
How did this happen? Hard to say. I’ve checked that my new sockets have not caused any shorts, and that all the pins are connected properly to other chips on the board. I’m convinced that the sockets themselves have not introduced problems. It is possible, however, that during the rework process – perhaps as a side-effect of heating parts of the motherboard – that some other component that was in a fragile or marginal state chose this moment to die.
This is always the danger with 40 year-old electronics.
At first I despaired somewhat. But then I got annoyed and now I’m getting organised. A computer, after all, is just a collection of chips strung together with wires. There’s no black magic going on. How hard can it be to fix it?
Since the last time I had the Master go bad on me I’ve learned a lot about electronics, and have accumulated tools like an oscilloscope, and so I feel a little more confident about tackling this problem.
And I have a fallback. I still have my original motherboard. And yes, that still has a problem, but possibly a less severe one.
A cunning plan
- Visual inspection of the motherboard, checking solder joints & tracks – looking for dry or cracked joints.
- Check that all socketed ICs (there are hardly any) are fully seated. Take out, use contact cleaner & re-insert when in doubt.
- Check for any short between +5V & GND.
- If a problem, check caps for shorts.
- Check for good +5V, -5V & GND at main power connections.
- Check for good +5V & GND at all ICs.
- Check for hot chips.
- Check clock signals wherever they should be present.
- Check /RST (reset), /NMI (non-maskable interrupt) & /IRQ (interrupt request) signals. These should all be high by default.
- Check activity on Address & Data lines.
- Check every other signal on major ICs for missing or strange waveforms.
- Check for shorts/low resistance between Address lines & Data lines.
Some of these things I’ve already tried, to a greater or lesser extent, but I plan to re-do all of them in a methodical way.
In forthcoming posts, I’ll log my progress and what I find.
Wish me luck!