MK14 EPROM Burner

[ScottishColin]

See pages 52-53 of the September 1979 copy of 1979.

https://archive.org/details/Personal...erWorld1979-09

While I was browsing, I came across this and thought it looked interesting.

However, those who know me on here know that I'm a little limited sometimes in my electronics capabilities.

Does this look like something someone with entry level skills could have a go at?

Thanks.

Colin.

[Phil__G]

Hi Colin, as theres no PCB its a bit more complex as you'll need to work out a layout on PCB or Vero, not difficult but more to it than first appears.
Its interesting in the way that it interfaces directly to the address & data busses, unlike most programmers, but limited memory means that programming has to be done in separate chunks.
Nonetheless its a great example of making the most of what was available "back in the day" ™
Alternatively, Richard over on the Nascom forum has recreated the almost-identical "Bits & PCs" eprom programmer with a nice PCB, but it would need the 8154 which the PCW one doesnt.
https://groups.io/g/Nascom-Computers...ammer/99002824
If its a learning project or an exercise, its all good either way. If its because you want an eprom programmer, both Richards & the PCW one have only one benefit over most in that they do 2708s, which may be a requirement, I dont know.
2708s are awful multi-rail eproms that must be powered up in the right sequence, but can be easily replaced by the much more friendly and convenient single-rail 2716 or even more conveniently the 28C16 EEPROM.
I have a lot of 'liberated' 2708s from boards where I've replaced them to simplify power supplies. Also bear on mind the project needs a track cut on the MK14 which you might not want to do.
If you have an 8154 and want to program 2708s, then personally I would go with Richards Bits & PCs board and have a bit of fun writing MK14 software for it
(well, to be honest I wouldnt - I'd use something with enough ram to fill the eprom in one sweep...)
Cheers
Phil

[Realtime]

Originally Posted by ScottishColin

See pages 52-53 of the September 1979 copy of 1979.

https://archive.org/details/Personal...erWorld1979-09

I can't access that link for some reason. Just in case anyone has the same problem the article is also available here:

https://worldradiohistory.com/UK/Per...9-09-S-OCR.pdf

[SiriusHardware]

I think this would have to be done on the basis that you were doing it for fun or education, Colin, as your STAG programmer probably outperforms that magazine article programmer in every possible way. Even with a Realview present to add 1.5K of memory, that still isn't enough to program even a humble 2716 in one pass.

[Realtime]

Originally Posted by SiriusHardware

I think this would have to be done on the basis that you were doing it for fun or education, Colin, as your STAG programmer probably outperforms that magazine article programmer in every possible way. Even with a Realview present to add 1.5K of memory, that still isn't enough to program even a humble 2716 in one pass.

RealView provides 2 banks of RAM. 0x200-0x7FF and 0x1200-0x17FF. So although not contiguous it would allow data for a 2Kx8 EPROM to be stored and programmed.

[SiriusHardware]

I was thinking that for the sake of simplicity the host system would ideally need a continuous run of memory at least as long as the memory capacity of the largest EPROM likely to be programmed.

To be fair, even the official SOC DM74S571 PROM programmer required that the PROM be programmed in two runs of 256 bytes.

[ortek_service]

Originally Posted by Realtime

I can't access that link for some reason. Just in case anyone has the same problem the article is also available here:

https://worldradiohistory.com/UK/Per...9-09-S-OCR.pdf

I've just tried the original archive.org link, and it does work OK for me - even on WinXP with now quite old last version of Chrome to support that. Although it did take a while to load the webpage (even with my 36Mbps download speed). There is a choice of formats there, but the only pdf format - direct link here: https://ia801900.us.archive.org/25/i...09/1979-09.pdf is a rather large 72MB - taking a while to load, and it hasn't been OCR'd.

A bit of Googling for this 'Selective PROM Copier', has eventually found a reference in a PCW Volume 3 feature index, on page 76 (of 172) of the 1980-July issue: https://worldradiohistory.com/UK/Per...0-07-S-OCR.pdf
- Where it also mentions that PCW Vol 1 and Vol 2 features were Indexed in PCW 1980 Mar and Apr issues.
So those may be worth a look-at, to see what interesting projects etc. were published back then, that might have been missed (I only started often looking at / buying PCW from around 1984/85 and I don't think I managed to pick up many old original paper copies before that)


Whereas the worldradiohistory OCR'd pdf is a much more compact 6MB, downloading rather-quicker, so may be the better one to use.

I've extracted the article's pages from that (plus appended to the end, the original PCW contents page + cover page, for reference), with this now rather smaller file attached here, to keep with thread.

Quickly looking at this article (a pity they didn't include a photo at least of the veroboard-built prototype, to assist with constructing it similarly), I noticed at the end that they promised they'd shortly be publishing another project from the same author - A PROM copier, which may also be of interest, to see what circuitry was used.
This is also stated to allow selective-copying, to bypass bugs, but I'm not quite sure how that would help unless you hadn't got a copy of the code elsewhere. And you'd still need to edit a few bytes, to replace the bug / possibly inserting more bytes and extending the code.
So unlikely to be too helpful for reusing the original non-eraseable PROM's, unless bug is near the end, and you could maybe overwrite it and the following code with NOP's, before appending new code (Assuming NOP OpCode was 00h, and bits programmed to a '0' on the PROM's being used - Whereas I thought, unlike (UV-E)PROM's, most fusible link ones were opposite to this).

A bit of Googling for 'Selective PROM Copier' PCW 1980, eventually found it in the 'Feature Index to Volume 3' in the July 1980 issue page 74 - Page 76 of 172 in pdf file, available here: https://worldradiohistory.com/UK/Per...0-07-S-OCR.pdf
Where it also mentions that they'd previously published Feature Indexes to Vol.1 and Vol.2 in Mar & Apr 1980 issues.
So these may well be worth a look at, to see what else was published back then that may have been missed
- I only starting getting the odd PCW magazine in 1984/1985, and I don't think I ever found many older paper copies before then.

[ortek_service]

Originally Posted by ortek_service

>>
>>
I noticed at the end that they promised they'd shortly be publishing another project from the same author - A PROM copier, which may also be of interest, to see what circuitry was used. This is also stated to allow selective-copying, to bypass bugs, but I'm not quite sure how that would help unless you hadn't got a copy of the code elsewhere. And you'd still need to edit a few bytes, to replace the bug / possibly inserting more bytes and extending the code.
So unlikely to be too helpful for reusing the original non-eraseable PROM's, unless bug is near the end, and you could maybe overwrite it and the following code with NOP's, before appending new code (Assuming NOP OpCode was 00h, and bits programmed to a '0' on the PROM's being used - Whereas I thought, unlike (UV-E)PROM's, most fusible link ones were opposite to this).

A bit of Googling for 'Selective PROM Copier' PCW 1980, eventually found it in the 'Feature Index to Volume 3' (July 1980 issue page 74), as being in Vol.3 Iss 3 (Mar 1980 Issue)

I've now looked at this 'Selective PROM-Copier' article - available in issue, here: https://worldradiohistory.com/UK/Per...0-03-S-OCR.pdf
and I found that it is actually for UV-Eraseable PROM's (Not Fusible-links PROM's), and is actually an add-on for the originally-published SC/MP EPROM Burner, being also titled "The Memory Lingers" - Part 2 (with a rather-long gap since part 1 in Sep'79 !)

I've also now extracted the pages from this, and have attached it here + Also a copy merged with original article, to keep this all together in one place, for convenience. This later part-2 additional, does at least come with a (assembly language) program (more of a routine, as have to call it with addresses in registers) listing.

Co-incidentally, I note that there's an EPROM-programming by-post service advert there (£9+VAT for a 2708, £18+VAT for a 2716 so quite expensive in 1980! - Although that's mostly for entering the code from a printout listing, as rather-less at £2.50+VAT, if you supply a Master EPROM to copy from). I've recently seen some places that still offer to program Bipolar-PROM's / EPROM's they sell, for no / very-little cost (with you providing a programming computer-file).

[ScottishColin]

Hi all - it was just for a bit of fun really. As Sirius says, my Stag programmer is far more capable but I thought I might try it and see.

I have breadboards so I might give it a go, although a photo would have been useful for me.

Colin.

[SiriusHardware]

Going back to what Phil said about how we used to have to use whatever was available to us, I would very much like to have known of the existence of this MK14 eprom programmer project at the time it was published. I definitely would have built it.

Around then I was engaged in building a radio transceiver and I needed a way to generate some non-sequential byte-wide codes to feed to the PLL frequency synthesiser IC. Normally this job was handled by a highly specialised switch which had the binary codes built in as part of its mechanical construction and the switches weren't available to buy, so I hit on the idea of using BCD up-down counters driving 2 x display drivers to show the selected channel and an EPROM to translate the BCD codes into the corresponding PLL codes, the only problem being that I didn't have an EPROM programmer.

I bought an EPROM thinking that if I had to, I could program it manually by applying the address and data for each location with switches and generate VPP by stacking batteries and use a relatively simple push button triggered monostable circuit to generate the ~50mS programming pulse, but I found myself defeated when I realised that what I had bought was a 2708, with its weird supply requirements. That added one layer of complexity too many and I gave up and just used a DIP switch for channel selection.

If had been aware of this project and done just a little bit more research I would have bought a 2716 instead.

I could have used a diode matrix instead, but for the number of channels involved that would have been an awful lot of diodes.

[SiriusHardware]

Although there is a hardware mod shown I think that is to remove the OS images which would be present on issue II through to issue IV MK14s from the address range 0400-07FF. It looks as though the EPROM (or part of it) is mapped to 0400 -> in the SC/MP address range.

On original issue V, Czech replica issue V, 'Slothie' issue VI and the 'Oddy' I don't think this hardware mod would be necessary since the PROMs are already removed from that address range on those versions of the MK14.

However, the circuit does require a connection to the SC/MP NHOLD pin which is one of the few SC/MP signals not routed to the rear edge connector on the issue VI - so a flying lead connection would be needed.

[ScottishColin]

Well that's a bit more cheery if confirmed - i don't really want to be cutting tracks (and at best adding a switch).

Colin.

Originally Posted by SiriusHardware

Although there is a hardware mod shown I think that is to remove the OS images which would be present on issue II through to issue IV MK14s from the address range 0400-07FF. It looks as though the EPROM (or part of it) is mapped to 0400 -> in the SC/MP address range.

On original issue V, Czech replica issue V, 'Slothie' issue VI and the 'Oddy' I don't think this hardware mod would be necessary since the PROMs are already removed from that address range on those versions of the MK14.

However, the circuit does require a connection to the SC/MP NHOLD pin which is one of the few SC/MP signals not routed to the rear edge connector on the issue VI - so a flying lead connection would be needed.

[SiriusHardware]

Better still, if you think you can do without the ability to program 2708s and just make do with 2716s then the circuit can be simplified considerably, see attached a very crude hack-down of the original circuit.

Here we have all the circuitry required for the abominable 2708 removed, and also the circuitry involved in removing the unwanted PROM images removed, since those images are not present on your issue VI. Nor is the -5V supply needed any more.

This leaves only one additional burden which is rather understated even on the original circuit diagram, the need for a +25V VPP supply.

You -could- chain three rechargeable '9V' NiMH batteries together in series to 'make' a crude 25V supply - at nominally 8.4V each when charged, the series voltage would add up to 25.2V which is not far out at all.

[Phil__G]

Dupe, sorry

[Phil__G]

Unless it was for an accurate repro or an as-new repair, I'd avoid 2708s like the plague, horrid things but fortunately completely avoidable

Re programming voltages, for as often as it was used I found three PP3's clipped together was very convenient on my homebrew 2716/2732 programmer for the Eaca Genie.
I did the 50ms pulse in software, no monostable, which would make G's circuit even easier!

[SiriusHardware]

If you do that though, you end up having to make substantial changes to the support program, whereas at least the simplified (2716 only) version of the circuit still using the monostable to time the programming pulse will work with the magazine article software as written.

Question for Realtime: I can't remember whether it is possible to disable the 0200-7FF portion of the RAM on Realview while leaving the 1200-17FF portion enabled? The programmer as designed places the EPROM, or one half of it if it is a 2716, at addresses 0400-07FF, therefore the RAM in that area can't be enabled when this project is present on the system.

[SiriusHardware]

Phil, when using a PP3 battery 'stack' for VPP did you find that it was important to connect the +25V after powering up the EPROM from 5V and disconnect the +25V before powering down the EPROM, or did it not seem to matter?

I have to admit that instinct would tell me it should be +5V on, +25V on, +25V off, +5V off, but if you know from practical experience that that is not the case then that would be useful to know.

My first working homebrew EPROM programmer was driven by my Spectrum and I distinctly remember having a normally off relay controlled by the host to connect VPP only when it was needed and to disconnect it at all other times.

[SiriusHardware]

My first 'edit' of the diagram for the SC/MP based 2708 / 2716 programmer to simplify it for 2716 only (post #13) was done in a bit of a hurry and I left a couple of extraneous details in. You can play spot the difference if you like:

[Mark1960]

Originally Posted by SiriusHardware

If you do that though, you end up having to make substantial changes to the support program, whereas at least the simplified (2716 only) version of the circuit still using the monostable to time the programming pulse will work with the magazine article software as written.

And also the hardware, as the monostable drives nhold to hold the address and data during the programming pulse. Without the monostable you would need to latch the address and data.

I’m not sure I would trust the monostable not to trigger on power up, also probably shouldn’t hot plug the eeprom, especially on the address and data bus of a live system. So the sequence should probably be power off the MK14, fit the device to be programmed, turn on the MK14 power, make sure MK14 is running incase of faulty eprom, enter the data to be programmed in ram, apply 25v, program eprom, turn off 25v, turn off MK14, remove programmed device.

[SiriusHardware]

And then do it all again to programme another sub block of the EPROM

If the lower extra RAM 0200-07FF on Realview can be disabled while leaving the upper extra RAM 1200-17FF enabled, that would at least allow a 2K 2716 to be programmed in two 1K runs.