MTX new Replacement mainboard ??

Modern, Memotech inspired, hardware projects
stephen_usher
Posts: 205
Joined: 27 Nov 2016 19:58

Re: MTX new Replacement mainboard ??

Post by stephen_usher » 06 Jul 2018 20:06

1024MAK wrote:Designing a new main board is indeed possible. But due to it's size, it's likely to be expensive and will probably need several revisions before a version that does not need modification wires is produced, making it expensive :o

Mark
Of course, the alternative would be to have a large but extremely simple adapter board with the port connectors at the back (very simple to get right first time, and possibly single layer double sided (as you need that for the edge connectors)) and have the actual computer on a small daughter board which plugs in on top.

There are other solutions involving 3D printed parts which would allow even smaller PCBs but more cables and more hassle.

User avatar
1024MAK
Posts: 569
Joined: 24 Dec 2012 03:01
Location: Looking forward to summer, in Somerset, UK

Re: MTX new Replacement mainboard ??

Post by 1024MAK » 07 Jul 2018 12:20

Sorry, have read through the part of my post quoted above, I realise that I forgot to edit what I said before posting. So a clarification: it's not the size alone that is the problem. It's it complexity and size.

I don't think anyone really wants a new board that uses 4116 DRAM chips, or has the existing power supply circuitry (as on the existing MTX500 boards). It would be nice to have one or two of the new developments on the main board. And I see no point in using DRAM chips for the main CPU RAM (a single 128k byte SRAM chip is a better option).

So we can't just take the existing schematic, enter it into a schematic capture program, and then produce a PCB layout from it. As firstly, there are bound to be errors (either from the original Memotech drawings, or human data entry errors). If we are using more modern chips, it's wise to test these changes before getting boards build. Especially if some of the circuitry has to be changed. And of course, that's before any modern developments are included.

So doable, yes. But also time consuming and it's likely at least one, possibility two or three prototype PCBs will be needed before we get to a production version. That's what I was trying to say. And that's why I think it would be more expensive compared to repairing an existing Memotech MTX500 board.

Of course, some Memotech MTX500 boards may well require a lot of work to fix. But as long as the PCB is not (too badly) damaged, this remains the cheapest option.

I would offer to help with repairs, but my life has got rather busy at the moment.

Mark

stephen_usher
Posts: 205
Joined: 27 Nov 2016 19:58

Re: MTX new Replacement mainboard ??

Post by stephen_usher » 07 Jul 2018 21:05

Ah, got the wrong end of the stick, I thought you were talking about a complete replacement fitting in the original case and giving the same functionality but using modern components. i.e. a fully physically compatible modern re-implementation.

Martin A
Posts: 417
Joined: 09 Nov 2013 21:03

Re: MTX new Replacement mainboard ??

Post by Martin A » 07 Jul 2018 22:56

This thread reminded me of something I tried a couple of years ago: the MTX538.

The distance between the rails on the MTX case is exactly Euro Card PCB size, So I tried squeezing a much of an MTX a possible onto a single Eurocard and came up with this:
Never finished
Never finished
abandoned.jpg (84.32 KiB) Viewed 736 times


Power would have been a 7.5v 1A supply driving a single 5v regulator positioned to use the existing holes in the MTX case.

There would have been a V9938 on a daughter card sitting over the 2, 64k x4 bit drams. The output is a 20 pin header of the same format as the one on the MTX plus except it's driven by transistors and not a CMX1645 to save space.

Memory wise there's provision for a 128k Flash containing the system roms but only a 32k 0.3" ram (also to save space).

For I/O I managed to squeeze in the keyboard, CTC, sound chip and an IDE interface. Memory and I/O decoding needed 2 22V10 GALs.

If I were trying that now, I'd probably try using one 84 pin CPLD and try to ditch the GALs and 74 series logic.

Physically it just about fitted under the keyboard, though the clearance for the CPU and CTC was very tight. It might have been possible to scrunch the chip sockets even tighter to fit a 128k 0.6" wide SRAM. However the wiring was already pretty awkward with the existing spacing.

After issues with the 9938 and Bill coming up with the propeller 80 column card at Memofest 2016, I stopped work and build the MTXmini that Mark photographed at Memofest 2017. viewtopic.php?f=25&t=365 That used a card and a half to fit everything in. But was MUCH easier to wire and has a larger RAM chip required for CPM.

A double eurocard design (so 200mm x 160mm) would be able to accomodate pretty much everything and still only be 2/3 of the size of the production mainboards. Though I don't imagine hand wiring one would be too popular :D

I think there's a photo of the part populated MTX538 board in Andy's Memofest 2017 photos.

Post Reply