R-Pi meets MK14

[SiriusHardware]

This actually concerns both a brand new computer and an antique one, but since the antique one featured is more likely to excite interest on this website than the modern one is, I have posted it here in this category.

This is a little Youtube clip of my Raspberry Pi and a custom built keypad interface being used to create, assemble, and upload code to the original strange little uncased, single board, cheap British computer.

http://www.youtube.com/watch?v=j1WSy_ueX34

[MrElectronicman]

Absolutely fascinating, I love my R-PI's too and I am definitely not a programmer, but I like to dabble anyway. Thank you for the insight and incentive to do a bit more.

[SiriusHardware]

Thanks for the interest - although I described the MK14 as British (which it was), it was actually a reworking of National Semiconductor's 'National Introkit', which was an early microprocessor demo / development / educational system.

Although I've owned the MK14 for 30+ years it has only been in the last year that I have finally managed to pull together a complete code development system for it. (I'm nothing if not patient...)

[dominicbeesley]

Impressed...not found a "proper" use for mi PI yet but that is inspiring!

D

[G8VAT Wolfman]

It must seem like Magic to the MK14...

[SiriusHardware]

One of the great ironies of the situation is that even the PIC16F887 processor used on the serial hex downloader / keypad interface has considerably more on-device power in terms of code memory, peripherals and I/O than the entire MK14 has. That's before you even start looking at what the Pi can do.

Of course, it's absurd nowadays to really think of using the MK14 for any serious purpose - programs are held in volatile RAM and disappear as soon as you switch the machine off. But I still have great affection for it because it put a computing device into my hands when I could not have afforded one any other way (The more serious 'Nascom One' and similar machines were priced way out of my schoolboy reach). I was (then) the kind of youth the Raspberry Pi is aimed at now, so the MK14 was very much the Raspberry Pi of its day.

And by selling quite well, it convinced Clive Sinclair that there was a market for his later more consumer-oriented cheap home computers which launched the British home computer boom of the eighties - In the meantime, Chris Curry, (also at Science of Cambridge when the MK14 was conceived) went off to form Acorn, who went on to introduce BBC model B computers at hands-on level into British schools. So although not particularly impressive itself, the MK14 is arguably one of the most important British computers ever made.

[ScottBouch]

Brilliant stuff, I love this combination of era's!

I have a 100 year old Drummond Round-Bed metalworking lathe that I'm planning on converting to CNC, that'sll be a nice mix of technologies!

I've only managed to datalog signals in my central heating boiler so far using my Pi..

Bearing in mind that all signals in my boiler are 240V mains, I had to build an opto-isolated interface circuit to keep the pi safe! a nice man on teh RaspberryPi forum wrote the code for me to log the time upon a change of state of any of the GPIO, and to record the states of all GPIO at that time.

End result, we're having a new boiler installed tomorrow at great cost, but it was worth a shot, and an interesting little Pi project!

Part 1: http://youtu.be/zI-46zi-kWE
Part 2: http://youtu.be/8MPLQEQiMEA
Part 3: http://youtu.be/E3ajZ3zJmto
Part 4: http://youtu.be/t-n8pnv9YPw

Also used it in the car with xbmc: http://youtu.be/F9FwQRZsdEE

This Pi is going to be relegated to media playing use behind our TV, but i'll be buying a model B Mk2 to use for fun!

Cheers, Scott

[MrElectronicman]

Hi Scott, I was wondering if your signals interacting with the Pi had any relationship with the Pi power supply

My first experience with the Pi was that the USB keyboard and mouse input would some times chatter, and I found out that the PSU was somewhat critical.

The first PSU was an old phone charger, however after changing to a better PSU, the problem disappeared.

[ScottBouch]

Hi MrElectronicman,

See my comments on www.raspberrypi.org on this subject!

I originally suffered from sticky keys like you describe, then discovered it was a PSU issue. I have had the Pi working happily at 4V though..

I now use a 2A USB hub to power the Pi, plus if I'm using an external HDD, another 2A supply for the Pi, leaving the Hub's current for the HDD. I have got a 10A 5V switch mode supply for if the going ever gets too tough!

I've not used my scope yet on it, but I think a cheap 1A supply, may start to get ripply near 1A, this may get the processor's knickers in a twist! I need to prove this though, as it's just speculation.

Cheers, Scott

[SiriusHardware]

Although power supply problems have been a large contributor to problems with the R-Pi, there is an acknowledged ongoing minor issue with USB which occasionally causes dropped characters, including the key codes and key-release codes sent by keyboards. If you take your finger off a key and the R-Pi doesn't see you do that, it starts auto-repeating the key it thinks is still held down. It can be annoying, especially if it takes off when you are in the middle of editing carefully formatted Python code.

There is a thread about the R-Pi in the 'modern technology' subforum for further discussions about that machine.

[MrElectronicman]

Although we are going OT, I just got a wireless mouse and keyboard and it worked for a few hours connected directly to the Pi, then it started to chatter as described above, so I switched it to the powered hub and it's been working OK for a few days now.

I found that any PSU that puts out less than 5Volts is suspect, and the PSU's I have that put out 5.25Volts do not have any problems in my somewhat limited experience.

[AC/HL]

One post deleted (eBay rule C3)

[Karen O]

I was one of the lucky Mk14 generation and I learned my microprocessor coding skills on one of these venerable machines. Sadly, mine got thrown out in a house move many years ago - something I deeply regret now.

A few years ago I implemented a cycle perfect emulation of the SC/MP on a PIC microcontroller, which I then used to make a Mk14 emulator:

http://www.geopodium.com/files/Karen/micro.htm#PIC14

I went on to make a NIBL computer, and even a complete NIBL machine with integral keyboard, video output and cassette interface (picture attached)!

That was all before I started the medication

I've since discovered that there is an error in my implementation in that jump offsets should not partake of extension register offset substitution. I'll fix that when I get a Round Tuit.

[SiriusHardware]

Yes, Karen, I remember seeing your PIC-MK14 project a while ago, very impressive!

There have also been two distinctly diverse resurrection projects that I am aware of, one a repro-MK14 using modern programmable logic to take the place of the original 1970s logic, as seen here:-

http://www.mymk14.co.uk/

(That site is a great resource for all things MK14, although the owner sadly no longer seems to be active)

Also, I'm aware of at least one person who has reverse-engineered the original MK14 even to the point of making his own PCB and the specific components of the keypad to make one which is as nearly original as possible. Here he is building it in this youtube clip:

http://www.youtube.com/watch?v=vhSAMRwGL-A

...And here he is laser cutting the upper frame for the keypad. I wish he would make one for me, since mine no longer has its original keypad. :-(

http://www.youtube.com/watch?v=vhSAMRwGL-A

Originally Posted by Karen O

I went on to make a NIBL computer, and even a complete NIBL machine with integral keyboard, video output and cassette interface (picture attached)!

That's really neat! How did you go about making the keyboard membrane? Is it a re-labelled ZX81 membrane or similar, or did you make it from scratch?

[Karen O]

See attached photo of my little computer before the keyboard legend sheet was added (which is just a laminated printed image).

[SiriusHardware]

Simple, but effective.

We build / maintain these things, but do we actually use them for anything? Have you ever used your little NIBL system for a specific purpose?

I have other retro micro systems which I keep in working order because I like them, but they won't actually get used for real purposes, as there are so many single chip microcontrollers around now which outperform them in every way.

In the case of the MK14 I think it is unwise to power it up too often because every power supply inrush could be the one which finally kills off a vital component (although I do have a spare SC/MP and INS8154, just in case). As you may remember from your own MK14, the bipolar PROMs in particular run very hot - my MK14 (as shown in the first video clip in this thread) consumes over half an amp (at 5V).

[Karen O]

I wrote a calender generating program for my PICX80 (see attached images). I also wrote a Babbage difference engine program for my Mk14 emulator.

The PICX80 is slooooooow. It spends most of its time maintaining the video output. Also, the program is stored in an I2C accessed EEPROM. The whole effect is like waiting at an ASR33 for a response from a distant mainframe!

[SiriusHardware]

Originally Posted by Karen O

I wrote a calender generating program for my PICX80 (see attached images). I also wrote a Babbage difference engine program for my Mk14 emulator.

I did notice that as well. I remember thinking that was a mighty big step forward from using 'Moon Landing' or 'Digital Clock' to demo the system.

Originally Posted by Karen O

The PICX80 is slooooooow. It spends most of its time maintaining the video output. Also, the program is stored in an I2C accessed EEPROM. The whole effect is like waiting at an ASR33 for a response from a distant mainframe!

The video output is done by software as well?!

One of the other items I have for the MK14 is its optional VDU board, although it hasn't been connected to the MK14 for more than 30 years. It used 512 bytes of the fully expanded Mk14's 600+ bytes, leaving a mere 128 bytes for program code and severely limiting the usefulness of the system. Also, it spent a lot of time halting the CPU while it helped itself to screen data from the system memory, reading the same data over and over again for every TV frame, because it had no onboard RAM of its own.

For more than three decades the VDU card remained a rather forlorn, static museum piece, but last year (2013) I decided to take pity on it and lashed up a companion controller PCB for it using a PIC as both the controller and the RAM. In character mode, the VDU card reads a character from the screen RAM every 2 microseconds so the PIC has to read the VDU's 9-bit address bus (to pick up the address that the VDU is trying to read from) and then fetch the corresponding data from the PIC's internal RAM and present it to the VDU databus, all well inside that time frame. I had a lot of trouble getting it to run at a high enough speed, eventually resorting to using a 40Mhz (10Mhz * 4) clock for the PIC. I couldn't even spare the time overhead that would be introduced by a 'decrement and jump if not zero' loop, so where I needed 16 read address / fetch data / output data cycles I had to repeat the code 16 times inline.

..so, I find the idea of trying to generate a full video output with a PIC frankly quite daunting!

[Karen O]

Hi,

My PICX80 was one of my greatest feats of PIC cramming. The screen memory is in the EEPROM and is character-based with newlines, etc. (kinda the way the ZX80 did it, I believe). The font memory is in the high order bits of the top 4k (the low order bits contain the NIBL ROM).

Using the EEPROM as screen memory has a longevity issue, but it is only written character at a time in a circular fashion so it isn't as bad as you might think. The PIC USART is used for serialisation of the pixel data and can do back-to-back transfers.

So, maintaining video with a PIC can be done and even character generation through a font table but it takes all the PIC's time. Also, I only attempted about 25 characters per line so that took some pressure off.

NIBL has an important place in my affections. It was my first exposure to the inner workings of a language interpreter and got me interested in how language interpreters and compilers worked in general.

[julie_m]

..... and it uses BBC-style memory operators! Except it uses @ where Acorn used ?. The calendar will only be correct from 1 March 1900 up to 28 Feb 2100, since it misses that multiples of 100 aren't leap years unless they are also multiples of 400.

I'll get my coat .....