MK14 schematic revisions

[Slothie]

Originally Posted by SiriusHardware

The keyswitches in my issue VI aren't super cheap ones, they are OMRON brand which I obtained through work - I mainly did that because the original ones I sourced had quite a robust 'click' whereas I wanted something which felt more like a soft-click keyswitch, so I got some low-Newton ones.

I have not had problems with key bounce, and mine is the exact same hardware as anyone else's apart from the switches.

Capacitors across switches may solve keys bouncing but if you do that then the maximum typing speed of a keypad injection uploader, should you have plans to use one, may be compromised. It may also have some effect on the display since the column drive lines to the keypad are also the drive lines for the common cathodes of the display.

Yes, thats why I wanted to experiment. I might just have to get some decent keyswitches - I might be able to get some that the keytops I have will fit to because I like them.
I do have a range of capacitors of various values, So I might see what the minimum value that seems to make a difference is - its entirely possible it won't work at all or have undesirable consequences as you say, but its worth trying given that its just a bit of work and 0 cost. If they don't slow the uploader too much then it might be worth it (talking of which, I should get the uploader put together!)
One reason I thought of trying it was that I managed to "cure" the bounce on my mouse keys by using a capacitor

[Mark1960]

Originally Posted by Slothie

So whats next, a liquid nitrogen cooled MK14 running at 25MHz?

Maybe an MK14 with three 8060s running at 8MHz

I think a 74hct109 could detect the halt flag to release reset on a second 8060, with any further halt instruction halting the first 8060. This could then be used with the second revision of the SoC monitor. When you press Go the running 8060 executes the target address while a second 8060 starts running the monitor. Just need to remember the monitor is not re-entrant.

On the switches, I find the tactile switches with button caps are a bit tall, so the cap tends to wobble side to side and doesn’t feel stable. I think I’ll try the lower profile tactile switches on the next board with a printed overlay. I want to try a second board with higher value pull up resistors and 74hct series as that might be better for playing with expansion.

[Slothie]

Well I've been experimenting on my MK14 and coming up with endless problems connecting to the top edge connector. I did find an 0.1" pitch edge connector that I could cut down to 32 ways, but the clips I was using to connect to the edge connector kept dropping off in a most annoying way. So I have decided to solder on a 32x2 dupont-style female connector onto the MK14 which will accept either a 32x2 card connector or individual dupont pin connectors depending on my needs. It might even plug into a DIN 41612 a+b style connector, but I'd have to see about that in the future if the occasion arises.
I probably wouldn't have done this to an oriiginal MK14, but for a replica that I wish to experiment with it makes sense because these connectors are plentiful, cheap and likely to be available for some time since they are popular with the Arduino crowd.
I did have to cut a 40 way connector down, but it all came out quite neat.

[SiriusHardware]

I use a conventional 32-way edge connector - I find that Arduino style 'jumper jerky' female-female leads are a snug push fit onto the pins on the back of the connector. The pins on mine are rectangular in cross section rather than square.

There's no doubt that an unkeyed edge connector is a perilous thing, though, you only need a momentary lapse of attention to fit one on one finger out or for something to knock it out of position while the whole thing is powered, so on the whole yours is a better, safer, if non-canon way.

DIN A/B connectors are pretty reliable too but it's quite hard to plug anything other than another DIN connector into one of those.

[Slothie]

At the time I was designing the Rev 1.2 PCB I searched high and low for a supplier of 0,1" edge connectors with end cheeks with the thought of putting a slot in the right hand end on the PCB to allow for them, but couldn't find a single supplier that did one to take dimensions from, and if you're cutting down a connector you're going to lose the end cheek...
I tried plugging dupont sockets onto the pins of the edged connector I had. They were rectangular, but quite skinny and a bit short so they just fell off at the slightest provocation. The pin clips I had are cheap ones and would pop off too.

[Mark1960]

Originally Posted by Slothie

It might even plug into a DIN 41612 a+b style connector, but I'd have to see about that in the future if the occasion arises.

The pins on a male DIN41612 seem to be a bit slimmer than the standard square header pins, though might be close enough to connect if the female header is shaved down enough at the cut end.

I think square header pins are based on the dimensions of wire wrap pins so the dupont connectors could be used with a wire wrap edge connector, if such things are still available.

The female header is a good solution and is what I was planning to use, though I think you made a neater job of cutting it to size than I normally achieve with side cutters and a stanley knife to trim the edge.

[SiriusHardware]

So now that you (Slothie) have the connections side of things sorted, what was it that you were trying to connect?

[Slothie]

Originally Posted by SiriusHardware

So now that you (Slothie) have the connections side of things sorted, what was it that you were trying to connect?

Well to start with, I was trying to get the tape interface to act reliably so I could do some speed tests, but when I've done that I'm making an Ortonview since I have all the needed parts in my spares bin so I can tinker with the firmware, and a RAM expansion, and get a grip on the SC/MP bus timings since the relationship to XOUT with NENIN, NENOUT seems to be a bit vague. Hopefully along the way I will get to grips with Bitscope....

[SiriusHardware]

So not really that much to do then?

For the OrtonView, remember we settled on the non-A version of the PIC16F877. The 'A' version works in character mode but for some reason, in Graphics mode, outputs every other video frame with a half-pixel offset to the right.

Also A8-A11 need Mark's fix of capacitors down to GND from those lines - in my case 200pFs, I think Tim was able to get away with a bit less (lower value).

[SiriusHardware]

Oh, you forgot the Arduino powered Tesla MH74S571 programmer. Good job I'm here to remind you...

[Slothie]

Originally Posted by SiriusHardware

Oh, you forgot the Arduino powered Tesla MH74S571 programmer. Good job I'm here to remind you...

Well can't do that till I get to my supply of Tesla PROMs. I'm collecting other parts though....

[Timbucus]

Originally Posted by Slothie

Well can't do that till I get to my supply of Tesla PROMs. I'm collecting other parts though....

I am sure someone with a supply of Tesla PROMS would lend you some until that time not to delay your rapid progress on such a device...

[ortek_service]

Originally Posted by SiriusHardware

Oh, you forgot the Arduino powered Tesla MH74S571 programmer. Good job I'm here to remind you...

On the subject of Arduino (mega-2560) powered / shield-for programmers / readers / IC-testers , I recently found this design that was published < a year ago: https://va3ngc.weebly.com/arduinoeprom.html#

Which looks very handy for reading 2708's (as used on Softy-I etc), as well as going right back to very first 1702 (/ 1701) EPROM's, on some more modern / easily reproducible hardware.
It does say: Basically any 24-pin EPROM or ROM that has VCC at pin 24, and Ground at pin 12 should work. Similarly, any 28-pin EPROM or ROM that has VCC at pin 28, and Ground at pin 14 will work.
But 1702's don't quite match this, even in read mode with pin 12 at +5V, Pin24: -9V and No Gnd! So you might have to put these in reversed (in which case it may also read MM5204's (used on the NS SC/MP Introkit etc) as they are similar-reversed, but with Pin24 at Gnd and extra -12V on pin23). And when I recently made a MM5204 to 2716 readout adaptor, I put the socket rotated 180degrees, as it simplified the interconnections crossing too-much.

I reckon it wouldn't take too much extra circuitry, to be able to program many EPROM's, as well - certainly >= single-rail 2716 (but tackling a 1702, with it's high negative programming voltages on many bus lines / supplies changing voltage / polarity could be a challenge....) - Already done with custom design, here: http://www.tronola.com/html/1702a_prom_programmer.html

Also, it should be capable of reading PROM's at least - Although may need to borrow some circuitry from SoC / Acorn PROM programmer, to be able to apply higher voltages on Data bus etc. to program these, whilst not having these voltage on lines back to uC for reading (but usually only 4bits to deal with, at least)

[SiriusHardware]

A quick information update: Thanks to Kan_Turk over in Ireland I've been able to try a pair of AMD 27S13A PROMs in my issue VI replica and they work fine. (K_T - you have a PM).

This is a complete list of PROM types which have now, to my definite knowledge, been tried and found to work in MK14s or replicas:

DM74S571 - National Semiconductor, as originally supplied in the MK14 kit.

MH74S571 - Tesla clone of above but with completely different programming regime / algorithm - quite cheap, but difficult to get programmed as only a handful of very high-end device programmers support them.

N82S131N - By Philips / Signetics.

27S13A - By AMD.

Honourable mentions to the DM74S571AN and N82S131AN which are just faster versions of the devices mentioned above and should work. The 27S13, which is the 'normal' version of the faster 27S13A, should work because the 'normal' versions of the other devices do.

[Slothie]

Originally Posted by SiriusHardware

A quick information update: Thanks to Kan_Turk over in Ireland I've been able to try a pair of AMD 27S13A PROMs in my issue VI replica and they work fine. (K_T - you have a PM).

This is a complete list of PROM types which have now, to my definite knowledge, been tried and found to work in MK14s or replicas:

DM74S571 - National Semiconductor, as originally supplied in the MK14 kit.

MH74S571 - Tesla clone of above but with completely different programming regime / algorithm - quite cheap, but difficult to get programmed as only a handful of very high-end device programmers support them.

N82S131N - By Philips / Signetics.

27S13A - By AMD.

Honourable mentions to the DM74S571AN and N82S131AN which are just faster versions of the devices mentioned above and should work. The 27S13, which is the 'normal' version of the faster 27S13A, should work because the 'normal' versions of the other devices do.

I will be sure to update my documentation, thanks Sirius. Still trying to work out how to draw a diagram of the keyboard that will print out dimensionally accurate using Libre Office...
Talking of keyboards, all I need now is a 3D printer...

[Mark1960]

Originally Posted by Slothie

I will be sure to update my documentation, thanks Sirius. Still trying to work out how to draw a diagram of the keyboard that will print out dimensionally accurate using Libre Office...

Maybe you could use kicad instead, possibly in a documentation layer of your VI board.

[Mark1960]

If the idea is to make a guide for the keytops you could make a pcb and mount it on spacers. Just remember that jlcpcb minimum inside radius is 1mm, minimum 2mm slot width. But you could make the holes for the buttons with H shape, overcut the corners, so there would be a slight notch at each corner of the holes for the buttons.

[Slothie]

Well Sirius had the idea of a print-out-and-keep page that people could laminate and put over small tact switches on the board since getting keycaps for good switches that can be labelled neatly is difficult. Its easy enough to write a program to generate a PDF that will print out OK (a few years back I wrote a python library to make PDFs) so I might just make do it that way as a separate download. Or I could find a tool to append the keyboard PDF to the PDF generated by LibreOffice, it just seems a bit of a faff.

[Mark1960]

Does it not work if you create a drawing in libreoffice draw and then insert in the libreoffice doc as an image without scaling. It’s a long time since I tried replacing excel with calc, so I’m not sure if this works.

[SiriusHardware]

Actually, I just meant a sheet that you could print and cut the authentic looking key legends out of as little squares to insert under the clear keycaps but if they are accurately spaced as well as sized then yes, an alternative would be to use the whole sheet in a similar way to the original key legend sheet.

I'm really very happy with the clear-cap switches - mine are low force ones and they don't seem to swivel / tilt around as much as the lower cost ones appear to.