When I was getting the Minstrel ZXpand II ready for the previous post, I needed to test on a ZX81 and get some photos.
I have many ZX81s, but I couldn't find any of them, they are all neatly packed away in boxes - somewhere. Somewhere amongst many many other boxes.
I had found a TS1000, which I used for testing and the photos, but I kept looking for a ZX81.
I eventually found a box with one very tatty ZX81 case and three boards.
Three boards, at least one of them is going to work, right?
Well, no, none of them did.
They were all issue one boards, two with full sets of chips.
And one with three empty sockets.
The third board had empty ULA, ROM and Z80 sockets and the alternate RAM option, a pair of soldered 2114 chips.
The RAM was a pair of NEC 2114. 209 could be a date code, 8209, 1982, week 9?
The 7805 is marked 8234 and the UM1233 Modulator is dated 8220, so this looks like a mid 1982 board.
The heatsink is marked "No B-£", which suggests a problem with the membrane, one of the 8 half rows is missing.
I tried that with known working chips and it did absolutely nothing.
Investigating further, there was nothing on any of the busses, which turned out to be due to there being no 5V rail. Dead 7805?
I tried it with 5V from the bench power supply, and it did a bit more, but as with the other boards, black screen, no contents. The ULA was a working chip, as was the ROM and Z80, could be bad RAM?
Looking over the board, the Z80 socket looks to have previously been desoldered. It looks the original type, so was it removed and replaced, or desoldered from another board?
The ZX81 is a very minimal design with just the 4 (or in this case 5) main chips, but they all need to be working to get a picture, if any of them are bad (or not talking to each other), then you get nothing.
You lose.
Good day, Sir.
A common theme with all the chips seemed to be bad contacts, When checking continuity from pin to pin, I was getting lots of intermittent or missing contacts.
The legs of the 2114 RAM chips were showing signs of corrosion, so it is quite possible the contacts inside the IC sockets are in a similar state if they have been empty for a while. There are not the best sockets anyway, single wipe, so you get one chance of a good contact with the pin.
When Sinclair went to the issue three boards (issue two is lost in the mists of time, we assume Rex ate it), they soldered all the chips in. I am sure there was an element of reliability as well as cost saving in that decision.
Desoldering Time
I thought I would start by desoldering that Z80 socket to fit a better one, In doing so, I found some evidence of damage to the pads, several of them were loose or missing. Someone had a bad day the last time it was desoldered. Based on the smell of the old flux, that was probably quite a long time ago. The smell of 80s flux is quite distinct. You don't get that these days, it has probably been banned.
Well, at least that explains why some of the signals were not getting through, there was no through hole plating left on several of the pins, with what turned out to be pointless blobs of solder on the bottom of the PCB not making contact with the top.
I know some PCBs were pretty poor quality in those days, but I had not had a problem with ZX81 boards myself.
As a test, and because it seemed like it needed to be done anyway, I tried removing the other sockets and the two RAM chips.
They all came off cleanly, no damage to the pads or tracks, other than the pre-existing damage to the Z80 socket.
I can't just replace that socket, there was nothing to solder to with the missing pads, previously, some were just optimistic blobs of solder.
I could solder a Z80 chip in directly and flow the pads on the top of the board, but it just seems wrong. I don't think I've ever soldered in a Z80 without a socket?
(ED. I think I did one on a Minstrel kit where the buyer specifically wanted only the ROM socketed for some reason. It looked nice, but felt wrong.)
I prefer to use turned pin sockets for various reasons, one of which is in situations like this where you might have missing tracks or through hole plating, it is useful that the pins are (just about) accessible on the top side of the board.
I had to temporarily remove a resistor array for better access, but with the 6 damaged pads, I was able to get solder to flow on the top side.
To do this, I would hold the board steady and then apply heat from the iron to the pin from the bottom of the board, then with my third hand and second pair of eyes, I would feed in solder to the junction of the pin and the pad on the top side of the board until it flowed.
There was one damaged trace on the bottom, the /M1 line (which is ironic given how many Z80s with bad M1 lines we see on Spectrums). I had to add a small bit of wire to reconnect to the trace on the bottom of the board.
The ZX81 has two sets of outlines for the RAM, either two 18 pin sockets for the 2114s, or a 28 pin socket for a 24 pin 1K or 2K RAM chip.
I am not sure what the plan was, I have never seen a 28 pin RAM chip with that footprint, pin 1 is connected to /Refresh, not sure why you would need that unless there was some kind of DRAM?
(also side note, I went down a rabbit hole of wondering why the only connections to /Refresh were the Z80, the edge connector and that pin. No connection to the ULA? On the ZX80, /Refresh is important for muxing the address lines and generating several of the control signals. The ZX81 didn't have the mux chips, just more resistors to overdrive the address lines from the ULA when it wanted to. It must presumably have been able to work out when to do that without /Refresh ?)
I think this board is heading towards being a test board, so I may as well fit both sets of sockets and a jumper for L1/L2.
The L1 / L2 links are used to set pin 19 of the 24 pin outline to A10 for a 2K chip or 5V for a 1K chip. I am not aware of any ZX81s that came with 2K, but I think most of the TS1000s did.
I initially tested with a 6116 2K RAM chip (the same type used on the VIC20 CR), and the jumper set to L2, for 2K RAM.
I had confirmed the 7805 was dead, so I fitted a switching regulator replacement.
With all the sockets replaced, and the new regulator, time to give it a go.
OK, looks promising, but I don't think I can keep using RF output, my TV isn't very good at tuning in and I don't have the time or space to dig out an analogue TV.
I fitted a temporary composite mod, to save having to use RF. Not the best circuit, more on that later I am sure.
Ah, that's better.
(at least for a few minutes, then it faded out, blame the composite mod and ULA)
I tried some more of the original chips, none of the ULAs worked, I was hoping the 2C210E would, as that is the better type, but unfortunately it didn't.
The best of them give a sort of rolling screen with a K but no keyboard response.
I will stick with my test one (8135) for the moment.
I did try the 2114 chips, and they worked after cleaning up the legs, always handy to have more of those.
Since that was working, I did a few extra things to the board.
The keyboard connectors were showing signs of corrosion, green pins on the outside, and probably the inside, which was also full of bits of broken of membrane tails.
Time for some new connectors there.
I also fitted a connector for a Minstrel keyboard, so I can use one of those on the test bench instead of a membrane.
I am also using a Minstrel Expansion Bus powered backplane to get 9V into the ZX81 board, so I have a power switch (inline with the cable) and a power LED, and I can monitor current consumption on the bench power supply.
I ran a few of the usual tests, all the keys worked, and I was also to confirm it was seeing 2K of RAM (18432 - 16384 = 2048 bytes)
Along with my usual test screen for checking the composite mod.
I got used to typing that in blind sometimes when the screen has gone too dark, but I still want to check what the output is doing.
The signal did not seem to change when it went dark. It is well out of spec in several ways, so I can't blame the TV for not coping.
The top trace is the ULA output, about 4V peak to peak, with no back porch to set the black level, and a very high black level of about 60% rather than 30% of the signal height as it should be.
The bottom trace shows the output to the monitor, only about 300mV peak to peak when it should be 1V.
I did quite a bit of fiddling with different circuits, I will have to come back to this.
I did find some PCBs for two different versions of a ZX81 composite mod / back porch board I had designed in the past. I don't think either of these was good enough, I can see I designed a few versions after these with different circuits, but I don't think I even get those made.
Further Testing
I was able to do what I started out to do, test the new Minstrel ZXpand II.
That worked nicely.
I was able to load 1K programs, like 1K Chess (amazing to think the fitted chess into 1K).
And with 2K fitted, I could load my character test program.
I fitted ZIF sockets to go through the chips from those boards and see what was working. Several of the chips from the other boards were dead, but if I can find another working ULA, there is enough for one more board.
Next Time
That's a good start, one board and some known working chips, time for another one?
Next time I will be looking at one of the other boards, and getting that repaired, upgraded and ready to go back into a ZX81 case.
But for now, Rex lies in wait as I can't load 3D Monster Maze with 2K of RAM.
All items are still available, and I can ship worldwide.
Including the Minstrel 2 and 3 which are available in versions which will fit in a ZX81 case as a drop in replacement for the ZX81.
The Minstrel 2 is ZX80 level, no NMI compute and display (slow) mode, but 16K of RAM.
The Minstrel 3 is ZX81 compatible, with slow mode and 32K of RAM to allow high resolution graphics.
The Minstrel 2, 3 and Minstrel 4th (Jupiter Ace / ZX80 / ZX81 / Lambda 8300 / RC2014,) are available with built in keyboards
All versions are available in kit form, or pre-built, and optionally the new Minstrel ZXpand II SD card and joystick interface.
International shipping is a bit complicated at the moment, so if you want to order from outside the UK, please contact me using the contact link at the top of this page.
See here for more information:
Sorry to have to do this, and I know this is going to affect sales, but please understand it is the only option I have right now.
Patreon
If you enjoy posts like this, you can support me via Patreon, and get access to advance previews of blog posts, and exclusive posts. Paid Patreon supports have already got the second part, and a part three will follow if there is any interest
You also get progress updates on new projects and other behind the scenes updates, as well as access to my Patreon only Discord server for even more regular updates, and to discuss your own projects.
