Maplin Z80 SBC

[mikew8760]

Hi,
I stumbled across the tread on this site concerning the above when trying to find some info on it. So I was lucky to find the find the code for the EPROM, but having now popped it into the socket I have discovered there is a board fault! I happen to have a Nicolet Z80 NICE, so that helps, but not much, as I have long since lost the board schematic.
I wonder if any of those who posted on the original threads (now closed) might have a schematic to share with me?
I notice that I had made a slight mod to the board, as there is a bodge wire from the 'selector' links and pin 1's of two IC's are lifted and wired together for some reason. I wonder if there was a wiring fault on the pcb?

Any help gratefully received.

[SiriusHardware]

This site has many issues of British electronics magazines online and available to read, including the one featuring the Maplin Z80 CPU PCB. Circuit diagram included.

https://www.americanradiohistory.com...1985-06-08.pdf

[SiriusHardware]

And here is the magazine with the article (including diagram) for the companion keypad / display PCB.

https://www.americanradiohistory.com...cs-1987-03.pdf

(Page 28 onwards).

This is the original one with 'proper' keypad switches, like the one on mine. A later (presumably lower cost) version used a membrane keypad instead.

[SiriusHardware]

As mentioned in one of those earlier threads, beware the NiMH battery on these - rather than leaking electrolyte it seems to weep some kind of corrosive gas which damages tracks, through-hole VIAs and screen printing in a fair sized radius around the battery. If you don't replace it, at least remove it before it does any damage.

If you no longer remember what the mod is for then it might be best to record the details of it carefully and revert the PCB back to as-designed, as per the circuit diagram, as a starting point for getting it working.

Which are the ICs which are involved in this modification, is it the RAMs by any chance? The original RAMs were 6116 types, I am wondering if you have modified it to accept some other type of RAM IC in what were originally 6116 sockets.

[mikew8760]

Thanks for all your kind reponses. I've looked at the archived magazines and looks to be just what I needed, although the schematic(s) are a bit 'blurry' and difficult to read.
I never did fit a rechargeable battery, so no problem there.
Regarding the bodge wire, I see on the schematic given that IC14 pin 1 is connected to IC15 pin 1, and nothing else. On my board there is a track from IC15 pin 1 to IC14 pin 12, which is just plain wrong! I wonder if mine was an early version, later corrected, although I have no memory of this.
I did make up a keyboard and LED's to connect to the 8279, but this is on an old RS 'breadboard'; no idea where the circuit came from though, as it's not the same as the real one.
And, finally, having faffed about with the NICE, I've come to the conclusion that it's faulty, since although it seems to work it cannot find any ROM or RAM. Grrr!
I shall burn a few EPROMs with test code and set my Logic Analyser on it!

[SiriusHardware]

Oh, well at least you've got good gear for this sort of thing. I would have killed for one of those at one time. It doesn't sound like you're going to need much help once you have some readable diagrams.

I have the original diagrams that I pointed you to in paper form which I could scan if those online ones are really hard to read - you can zoom a long way into them and I find them quite easy to read, but then I only ever access the 'net using a computer, you may possibly be using a phone to view them?

I'm also sure that I have the same two articles as a single article in a Maplin projects book - not sure which number, but the website I pointed you to does not seem to have an online copy of it yet - it relies on people donating the 'missing' copies so they can be scanned and put online.

[SiriusHardware]

One further thing, you mentioned that your keypad / display module was homebrew and did not follow the official Maplin circuit. If you are using the M12 eprom code which you probably found in one of the older closed threads, the circuit needs to be more or less identical to the official circuit otherwise the keys will not be mapped where the M12 eprom code expects to find them in the row / column matrix. Furthermore, the allocation of display segments to bit weights on the official circuit is completely unconventional.

Conventional would be

bit 0 = segment 'a'
bit 1 = segment 'b'
bit 2 = segment 'c'
bit 3 = segment 'd'
bit 4 = segment 'e'
bit 5 = segment 'f'
bit 6 = segment 'g'
bit 7 = segment 'h' (dp)

On the official Maplin system keypad / display PCB the allocation of bits to segments is NOT hardwired that way and so the display routines / character to 7-segment lookup tables in the M12 eprom are configured accordingly, to compensate for the non-standard scheme adopted by the keypad / display hardware designer.

Somewhere in one of the older threads there is a commented disassembly of the M12 eprom, near the end of that disassembly is a description of the non-standard allocation of bits to segments.

[mikew8760]

Hi,
Thanks again for taking the time and trouble to reply. I have now discovered that the main board is OK, the problem seems to stem from 'dodgy' UV EPROMS. I have quite a collection of old devices which can be erased and re-programmed, but simply do not work. The logic analyser only sees the prom memory space as all FF's, and scoping the data lines shows them as being marginal, with low level outputs and slow(ish) rise and fall times. I had expected the problem to be with the prom CS and/or OE, but substituting a 6116, and lifting the WR pin and connecting it to the WR line works fine. I'm puzzled by this, but it seems to indicate some degradation of the proms, which is odd since they erase, program, and verify OK. I've ordered a 28C16 eeprom to replace the 2716; how much easier if flash memory devices had been available 40 years ago!!
I shall have a go at reverse engineering my somewhat bodge job of the kbd/display unit; I used an LED display from an old calculator with lots of transistors (16!) to drive it. The other problem is that I seem to have robbed one of the only two ic's and not sure what it might have been. The remaining one is a 74LS155, so perhaps the other is the same.
BTW, I have two Logic analysers, an old Paratronics beast I got as a 'dumpster' find, but it's so big and heavy I never use it now, had cost someone £20,000 in 1980, and a £10 one about the size of a matchbox, which just about does the job. Isn't technology wonderful?

[SiriusHardware]

I'm glad you are making progress.

If you want to try to 'keep it real' I have some New-Old-Stock believed working 2716s, in fact I think I programmed one with the M12 code a while ago intending to send it on to a fellow forum member but it wasn't required in the end.

I also still have my working Maplin SBC system so if there are any concepts, ideas or components you need trying out in a working system let me know.

I find it quite mysterious that your EPROM programmer reads the programmed devices back OK and yet the system itself can not - maybe the programmer reads devices at a more slow-but-sure speed.

[mikew8760]

Hi.

Yes, the behaviour of the EPROM's has me puzzled, but I have yet to try a few tests to pin down the nature of the 'fault'. I had thought it might the address decoding, but the LS devices all check out OK. As I mentioned before, I put a 6116 in the EPROM socket with the WR line connected to the '139 output and that all worked fine.

My EPROMS are all quite old, perhaps over 40 years old, but are OK on the programmer; the only thing that makes me suspicious is the Vpp voltage. I don't have any manufacturer's info on the 2716's, and some are of unknown origin, so not sure about the correct Vpp to use. I actually have some that will not program at all and some that I have used 21.5 volts to persuade them to accept code. They then read OK on the programmer, but not on the SBC. I have wondered if the Nicolet NICE may be at fault, but everything else works, although the so-called manual I downloaded from somewhere is quite poor; e.g. it has a command EPP, Enable Print Point, but nowhere describes how or what it is?

Thanks for the offer of EPROMS, but if and when I get a couple of flash devices that should be more convenient.

[mikew8760]

Short (?) update on my adventures down the rabbit-hole!

Regarding what I thought were 'dodgy' EPROMS turned out to be a bit of a red herring. Having taken the trouble to look at old data sheets, I find that programming requires a Vpp in many cases of 25 volts and a pulse width of 45 - 55 ms. My all-singing, all-dancing programmer is deficient in this aspect. It only provides a highest Vpp of 21v (actually measured at 20.6) and a pulse of 1000us (1ms) max. I'm surprised and disappointed, as I would expect the designers of the programmer to have taken these requirements into account; perhaps more modern chips will program at these levels?

I found that it was actually possible to program my EPROMs by repeatedly hitting the 'program' button. It would fail at the first attempt at address 0, then 1, then 2 etc. Clearly the low Vpp and very short timing would not allow sufficient charge to be acquired by the floating gate.

But I'm still having problems with the old Paratronics NICE; it seems old age has produced some intermittent connections. So I have been forced to get out the big guns and try my old Microtek MICE II emulator!!

[SiriusHardware]

Ah, then the 2716s I offered to let you have probably wouldn't have solved your problem. Although in brand-new never used condition, they are Intel i2716 dated '77, so they almost certainly require 25V VPP.

My high-end device programmer from the mid '90s (£800+ at the time) does 25V devices OK, like you I am surprised that any device programmer made post-period did not allow for the possibility of 25V devices. Staying with 2716 EPROMs I think someone pointed out that there is one version in particular, possibly Texas/Ti, which has non-standard power requirements even in read mode.

The whole issue of VPP voltages varying from one brand to another is a bit of a minefield - moving on to slightly later/larger devices, you start to see some which only require 12.5V VPP rather than 21V VPP.

[mikew8760]

Anyway, I thought changing to a parallel Flash EPROM would bypass this little problem. I got a couple on ebay, but neither of them worked! Surprise, surprise!! Begrudgingly given a refund, but still a disappointment; they looked genuine Atmel parts, seemed not to be the re-cycled variety, but arrived without any ESD protection. The ones I ordered from China are unlikely to arrive, I guess; can't think why?

[SiriusHardware]

I'm not sure what you are attempting to do: If all you need is something, anything, programmed with Maplin's original M12 code then, as mentioned earlier, I think I already have a 2716 programmed with it. Yours if you want it. I might even know where it is.

If your aim is to write some custom code for some purpose then obviously you'll need a more flexible solution which you can reprogramme yourself. I may possibly have a small number of 28C16s which were unused stock that I saved from being thrown out from work years ago. I'll have a look tomorrow.

Edit: forgot to mention, if your programmer doesn't mind programming 2732s or 27C32s, programme your code into the upper half of one of those and it will look like a 2716 as far as the system is concerned. (I hope I'm remembering that correctly).

[SiriusHardware]

Found a few CSI branded CAT28C16 - these are 'brand new', original, never used stock in the original tube with their pins still splayed. There are more than I will ever need, so you are welcome to a handful if you want them.

[mikew8760]

That sounds like an offer too good to miss! A couple of them would be ideal, just let me know how much you want for them. Do you take PayPal!?

Best regards.

[SiriusHardware]

There is no charge, as explained previously I intercepted them on the way to the WEEE bin a few years ago when work tried to throw them out. Originally that was because we still needed them for repairs even though we no longer made the equipment they were used in (nobody ever considers the needs of the repairs dept, to my constant annoyance). However it has now been quite some time since we've had one of those items back, so I think the ones I am holding on to will meet any future need.

When you've been a repair technician as long as I have you develop a kind of spidey-sense, and mine started tingling today when I remembered that you had tried to program two completely different devices with the same programmer and had problems with the first and no luck with the second. So... I have just taken the extra step of loading up a 2048-byte binary file which I have lying around and programming it into four of these devices and verifying them before writing them all back to FF again - they all definitely do work.

If you find that you can't program these ones either, you may therefore be looking at an issue with your programmer, rather than the devices.

For information, they are CSI (Catalyst?) CAT28C16AP with a speed marker of -20 which I suppose is probably 200nS.

If still interested, drop me a PM with where you need them to go to.

[julie_m]

I'm sure I heard a tale of some hapless person who managed to blow up several EPROM programmers with a single faulty chip. It could have been just an urban legend, though .....

[SiriusHardware]

Good ..or possibly bad... thought, Julie. With any luck, you are wrong about that. As mentioned, I have tested the devices I am going to send to Mike, so let's hope his programmer plays nicely with them.

[mikew8760]

The 28c16's were just the job, and programmed fine on my MiniPro programmer. I have been spending some time on other projects, so have just returned to the Maplin SBC.
I have given up on the Nicolet NICE. It's a dinky little piece of kit, but something has happened to it during the umpty years in storage, seems like a poor connection, but I cannot find any such.
So, I dusted off my 40 year old Microtek Z80 MICE, plugged it in and, amazingly it works! However, as usual, there is a snag. The comms is RS232, intended for a dumb terminal. I wish to use it with, say TeraTerm or PuTTY via USB, but the MICE is sufficiently old to expect 'true' RS232, with +/- 12v signalling. Of course there are myriad so-called USB to RS232 adapters, but all seem to be TTL signal levels - the MICE simply refuses to talk to such low-life devices. I suppose I could modify the MICE circuitry, it uses 4188 and 4189's, but I don't really wish to go down that rabbit hole.
So, how to acquire a USB to pukka RS232 adapter? I did have one, which had a Sipex chip to create the +/- 12V, but Sod's Law, the Rx channel don't work!
Does any know of a cheap adapter that will fill the bill? At the moment I'm considered making up a custom jobbie, with a couple of joule thieves to obtain the 12v. Again I could buy a MAX232 chip, (or even those adapters on ebay that also offer RS485, which are expensive and which I don't need) but where's the fun in that?