The Speccy from Hell

[LambdaMikel]

Pretty unusual and misleading defect symptoms and challenging repair (for my Speccy skills, at least):

 

[johnx993]

What's your take on repairing Sinclair Systems?
I've only done a couple (plus some calculators), and I find them to be more prone to failure than other systems.
-J

 

[LambdaMikel]

What's your take on repairing Sinclair Systems?
I've only done a couple (plus some calculators), and I find them to be more prone to failure than other systems.
-J

Not much experience - I tried to salvage a lot of Timex 1000 & ZX81s (during Covid, I acquired them for 10 to 20 $ broken from eBay, and tried to repair them), and out of 10 or so, I managed to get 4 working ones. So your millage varies.

I really like the Speccy, but it can be challenging to repair, as my video shows. They were designed and built to a very low price point / budget consumer electronics. What can you expect. I am happy that we have replacement ULAs - with that, they will live forever!

Unlike the Timex / ZX81 PCBs, the Speccy PCBs seem to be of good quality - I replaced 8 chips, including CPU and ROM, and not a single lifted trace or pad!

The 12V for the RAM and the unusual vintage transistors are challenging, too.

Are they more error prone with higher failure rates than other 8bit machines? Probably. It's a lot of parts in a very crammed space, and some of the design choices were probably a bit too much on the edge of what is reliable. Look at that poor little 7805 with its giant heat sink that has to power *everything*, including the Microdrives...

Not talking about the QL at all... this seems to be an entirely different cup of tea. I have no experience with that.

The Microdrives etc. are a bit of a nightmare too, of course. I keep them as curiosities in my collection and enjoy them.

 

[cj7hawk]

Nice Work !

Historically, I too feel that they are more prone to failure that some of their contemporaries. The -5 and +12 volt regulation inside along with the 5V regulator caused heat problems, and the transistors were close to their working limit. Selection of RAM that had already failed in many cases often led to more failure. The expansion port also was a source of failure, given that peripherals had direct access to the bus, and there's some nasty-high voltages exposed there along with data lines.

Almost anything in it was prone to breaking. The ULAs, the z80, the RAM... And especially the Power Supply inside was just waiting for a mains-glitch to give it an excuse.

Nice work on the repair. Some are difficult to diagnose. I had a 128 that failed because it was waiting for a particular state of the AY-3-8912 audio chip on boot, and refused to go elsewhere... I used a CRO to figure out when the CPU was stopping by watching the address lines, and found it in a tight loop, tracked down the ROM disassembly, noted it was waiting on an external bit from the sound chip, and replaced the audio chip to get it working again. The second most difficult to fix was a Toastrack - no complex faults, just corrosion and naturally dead chips, though it looks like it was dug up from landfill, and when I got it, it was still filled with the landfill. Severe corrosion but some how no track faults, but nearly every second chip was dead. So in the end, I gave up on logical fault finding and just removed every chip and socketed it.

Some are harder to fix than others. I actually love fixing them, so squirreled away a few so I can play with them later and get them working again