Sinclair ZX80 / ZX81 / Z88 Forums

Hello folks,

Long time no visit, it's been a while since I've had a 'new' ZX81 to play with but when I saw this one in the Loire Valley I couldn't resist it.

Sadly on its trip across the Channel some careless oaf dropped it and smashed one of the boards' expansion connectors. Some careful sticking and glue applications has restored things but power input is unknown so before I put 12V up it and possibly blow something up, assuming it worked before of course, I thought I'd put some pics up here first for you all to pick over and ask questions. History is totally unknown, the seller didn't know either.

First thing the power does is hit a 5V regulator so that's pretty clear, but there's also 2 RAM boards and the 'special' wire wrapped board that the external keyboard connects to and also appears to contain a serial connection. The RAM chips are D416C and 4116 types needing +12/-5/+5 so something has to provide that. Trouble is the only name on the RAM boards is 'MIC-Rob' who these days happens to be the name of some music type. No info there either.

The reason I think power input needs to be +12V or higher is the RAM boards only contain another 5V power transistor. The only source of -5V I can see comes from a CD4049UB inverter. I need to trace out a schematic before I do anything else.

The board itself is an issue 1 and fortunately still has sockets for RAM so I can test it on its own just by adding a pair of 2114s.

Anyhoo, pics.

More pics.

aaaand finally...

WoW!

Various circuit designs were used in the past to obtain a -5V supply from a positive supply rail. The power required by 4116 (and equivalent) DRAM chips on the -5V is not very much (it is a bias supply). 4116 (and equivalent) DRAM chips actually run most of their internal circuitry from the +12V rail.

And various ZX81 RAM expansion boards relied on the Sinclair "9V" PSU actually having a voltage of around 10.8V upwards (10.8V being 12V - 10%).

When you are ready to power up, immediately test that all the -5V rails are within 10% of -5V. If this supply voltage is missing, it can result in 4116 (and equivalent) DRAM chips self destructing. So if it is not okay, switch off immediately.

Good luck

Mark

Evening Mark,

I buzzed out the rails from the RAM chips, -5V comes off the emitter of a BC307 amplifying transistor and +12 is coming from the emitter of a BD237 power transistor which is fed by whatever the 9V rail is providing at the edge connector. The CD4049 inverter is also fed off this rail and does the RAM control signals. Could the stock power brick provide enough current for 16x 4116 plus the extra ROM board which I notice also has RAM?

Cheers

Hi Witchy,
We are interested in the ROM set…
The Fast load Monitor : "FLM"...
Can you dump them?

Those RAM boards could very well be off-the-shelf RAM expansions. With 16KB on each, I suspect one of those is modified to fill some other memory area than the 4000-7FFFh for a normal 16K pack.

That last expansion board is certainly some custom (or at least low-volume) job. Wire-wrap no less. Great find!

If I were to take a guess, my thoughts would go in the industrial / laboratory direction. Maybe used to control some machinery / measuring equipment, or some kind of data logger? Or it was used to prototype other ZX81 expansion cards?

Looking at the pictures closer: that broken-off strip serves to hold 2 rows of contacts apart, to form an edge connector. So if you pull it out of whatever it was stuck in, on power-up those rows of contacts (backside of "MIC-Rob" RAM board?) would make a massive short circuit on the ZX81's edge connector. So whatever you do, put some other strip of non-conducting material into the back of that board's edge connector!

witchy wrote: ↑Tue Jun 18, 2019 10:30 pmI need to trace out a schematic before I do anything else.

Might be a lot of work. But as a short-cut: I'd have a look at how the various boards & mods are wired, so that for each item you get an idea what purpose it serves / what part of the ZX81 circuit it modifies (power circuit as #1 priority). Maybe some of them are easy to -temporarily- disable. Maybe RF output was modded into composite, which would be good to know before hooking up a monitor. Etc.

From what little I could find, MSM5126 is a 2K*8 bit SRAM, so together another 4 KByte RAM there. Maybe some user-defined graphics function?

And perhaps: dump those EPROMs (out of circuit, using an EPROM programmer or whatever). Contents could provide a lot of info about what this machine was used for. Good to have those on file in case something 'blows up'.

Morning all,

The 2nd RAM board in the stack is very much home produced with patch wires and bodges using a small area of the board that seems deliberately placed for this reason, but the 1st one is more professional with solder mask etc and no bodges. It's possible that the 2nd one was prototyped off the first one to allow the address mods mentioned by @RetroTechie. I'd agree with the industrial application too.

One thing I didn't show was that I'd fixed the broken edge connector with only some minor adjustments needed to the 5V and 9V fingers

The order of boards is:

ZX81 with custom keyboard and small circuit for adjusting cassette levels.
16K board (solder mask)
16K board (home made with bodges)
Custom ROM board for keyboard with 25 pin flying socket.

There's a wire coming off the video feed to the modulator that ends up on the DIN5 socket at the back so that socket obviously provides power in and video out, this video connection also goes to the little circuit that feeds the cassette socket at the front of the case.

@XavSnap the first thing I did was dump the ROMs, my EPROM machine is off at the moment so next time it's powered up I'll get them and attach them here. I've not checksummed the ROM on the ZX81 itself but looking through the contents it's almost identical to the stock ROM, but if you think FLM is a fast loader then it must redirect the LOAD command to a routine on the ROM board at the back.

Moar pics:

OK, here are the ROMs. The 'ZX2 0CDA' is the on-board ROM and it differs in only 2 locations from the stock ZX81 ROM, the first one changes RAMTOP from 0x7FFF to 0xBFFF so that's 32K to 48K. The second location is 0x0879 which according to the ROM disassembly is in the COPY routine for the printer, this is replaced with C3 D0 30 (JP 0x30D0) which is probably in "ZX E000.bin".
Interested to see what you find!

In regards to the 4116 16k byte DRAM boards, I get the impression that the one that does not use a solder mask is the earlier of the two, as it does not use plated through holes. Whereas the board with the solder mask, does use plated through holes. Maybe a company started making the first version, and when they found they were selling a lot more, they moved the the more expensive (but with a saving in labour) plated through board design.

I don’t think the board without the solder mask is a copy, because why would anyone when doing a home made copy, reproduce the text from the existing board exactly?

The number of capacitors and diodes around the 4049 CMOS inverter, makes me think that one or two gates are being used as an oscillator (astable multivibrator). That then uses a capacitor to transfer charge, into another capacitor and hence produce a -5V supply. One of the 6.2V Zener diodes then helps with the regulation to give a -5V supply.

Mark