A More Useful MK14
 

It's been suggested that I post the mods I made (way back) to convert my MK14 to a more useful system.
So here goes.

As supplied, MK14 has a 12-bit address bus. So entering any address, eg XF20, selects the same ONE location (of 4K), whatever the value of X.

Issue 5 original address decoding is shown in (A). Gates are shown with SoC IC number. No unused gates remain.
This puts PROMs at X000-X1FF. And leaves X200-X7FF unused.
(An earlier MK14 modified per SoC instructions will have different pin numbers.)

All MK14s have X800-XFFF occupied. Half of this is 'repeats'.
In particular, address lines connected directly to the '8154 present it at X8xx, XAxx, XCxx, and XExx.

With some basic Boolean and a few extra gates the memory map can be completely tidied.
(This was my first 'hands on' micro system; I avoided loading the unbuffered address bus further.)
Additional circuit is shown in (B), taking inputs from original circuit at A,B,and C. Gates are shown with their 74LS- number.

1. Break original decoding at D-E and insert the new circuit D-<AND>-E.
This puts Keyboard/Display at XDxx (used by SCMPKB and SCIOS). X9xx is free.

2. Output I - to CS1 of '8154, pin 34. A8 still to CS0, pin 35. A7 still to M/IO, pin 33.
This puts RAM/IO at XExx. X8xx, XAxx, XCxx are now free.

3. Output H - to Extra RAM enable. Now moved to 0Cxx. 0Bxx free.

4. Now two 1K blocks are available, X400-X7FF and X800-XBFF. Outputs F and G enable these.
PCB positions IC2 & IC3 and IC4 & IC7 can be modified to take 2114s.
(System RAM has to be at XFxx - IC6 pairs with IC5 by sharing CE from IC17.)

5. X200-X3FF remains. Breaking X-Y in original circuit and connecting B-Y presents PROMs(twice) at X000-X3FF.
A 1K PROM/EPROM can then be used here for Monitor + utilities.

Re: A More Useful MK14
 

This is great information - thanks for sharing a photograph as well - always interested in how people have modified their MK14's especially BITD. I have to ask what is the other larger board on the right with the IDC header?

Re: A More Useful MK14
 

I have to say that's not nearly as untidy as I was expecting it to be. :)

Unlike some here, I only have the one MK14 (original, issue II) and it certainly did get hacked about in the early stages of its life, although not nearly as much as Circuitryboy's. It did have the single-step circuitry floating out to one side and at various times the MK14 was in a proper enclosure so the keypad and display were extended offboard.

For at least the last ten years I've been more concerned with trying to restore it to close to its original configuration so I wouldn't ever carry out such extensive mods to it now. If I ever come to own one of the more modern replicas, they are built on sturdier PCBs and can probably take a bit more punishment. I would be less averse to experimentation on one of those.

Re: A More Useful MK14
 

Thanks.
The other board is a piggy-back which makes things really useful.
It latches the top four addresses and flags, plus tristate buffers on address and data.
(Pic not good - my regulation 200K JPEG got strangled down to 55!)
The red MK14 socket is now 42-way (yes, is the answer to everything). The p-b detects non-Page0 address; a signal back to main board forces MK14 to Page0 only.
And the 40-way connector let's you bus off to the outside world.

Re: A More Useful MK14
 

Ah interesting - that is what I was looking to do (the signalling back to the board) to allow the MK14 to run NIBL BASIC - Nick Toop probably did something similar as he authored a few items on NIBL basic on SC/MP while at SOC.

Re: A More Useful MK14
 

On top of all that I think Circuitryboy's MK14 must have had a VDU connected to it at some point - the PCB having had a 4.00MHz crystal fitted, as stated in the Tesla Programmer thread, is usually the giveaway - unless that was done to make the calculation of software timing loops less hair-tearing?

Re: A More Useful MK14
 

Well you're right and wrong about the VDU!
I could see attempting the SoC version would not be a value-adding endeavour. Instead I built a version of the PE VDU, feeding video to an old Dacoll green screen.
SoC supplied the PAL crystal because of cost. I wasn't concerned about overclocking but I did want easy timing calculations. (Also the PAL alloy can was huge.)

Re: A More Useful MK14
 

My MK14 did have a SOC VDU connected to it for a while. (I still have it, but not connected to the MK14) As a fundamental learning aid it was useful, in the way that it demonstrated the direct relationship between bit states in memory and pixels lit / not lit on the screen (in graphics mode) or character codes in screen ram and characters on screen (in character mode). In terms of actual usefulness though, it used up so much of the 'standard' RAM that there was very little RAM left over in which to write code.

On your greatly expanded machine I would have thought it would be quite usable, since the main problem with it was insufficient RAM to support it.

The other thing I didn't like about it was the way it drastically slowed the machine down by halting it whenever it was reading data from the RAM. I had an idea at one point to put a dual-port RAM IC between the MK14 and the VDU, arranged such that the VDU could read from the dual-port RAM without having to use any of the MK14's RAM and without ever having to halt it while doing so - in the meantime the MK14 could write to or read from the RAM via the RAM's other port. I got as far as buying the dual-port RAM IC but no further than that.

My MK14 did have a small 4.00MHz crystal fitted during the 'VDU' years but as part of the recent 'restoration' drive I obtained and fitted an original 'big can' 4.33MHz crystal - apart from anything else, it makes 'Music Box' and other timing-dependent original programs run correctly again.