Tiny SC/MP system !!!

[Phil__G]

I just finished manually typing in an old paper document, "A Guide to Kitbug" whilst making minimal edits to cover KitbugPlus. Its at the bottom of the SC/MP page
Typo spotters welcome!

[Phil__G]

I've long assumed "A Guide to Kitbug" is a National Semis publication but I've since pondered the incorrect reset statement, a mistake that NS wouldnt make.
Digging deeper, I found that its actually from Kemitron.
Credit where credit is due, I'll amend the pdf

[audiokit]

Today I finally found some time to change my wire wrapped SC/MP intro kit board.
Phils KitbugPlus uses more memory then the original Kitbug so I had to extend the ROM from 0000 - 01FF to 0000 - 07FF and of course start RAM at 0800.

This can be done by changing the chip enables for the ROM and the RAM from address 9 to address 11.

After that, address 9 and 10 from the SC/MP have to be connected to ad9 and ad10 of the eprom (2716 in my case)

Actually that was all the rewiring I had to do.

At first I did not have a CR after using the "D" command (see picture) but after a mail to Phil he pointed out that I had to configure Teraterm Rx/Tx as Auto/CR. After that all went well.

This version is a nice addition to the existing Kitbug. being able to show a whole block of memory and download newly made binary code from SBASM with the "S" command to test it on the real SC/MP.

Well done Phil

[Phil__G]

Sorry for the late reply, just got home from a 10 day camping trip
It's great to see KB+ being used on different hardware Benny, thank you

[Phil__G]

Heres the 'John Conways Game of Life' I did for the PICL, adapted for KB+ on a real SC/MP

Use the command S800 to load, use Teraterm 'send binary' with no delays, takes about 8 or 9 seconds, transfer activity can be seen on the FTDI LED.
When the LED stops, press reset and enter G800


.

[Phil__G]

A starting-pattern variation:

[Phil__G]

My "Vintage Computing Christmas Challenge 2022" (VC³ 2022) entry.
Load the .bin using S800, wait for the FTDI LED to finish, reset, G800

[Phil__G]

Music program from the SoC manual
Load the .bin using KB+ command S800, wait for the FTDI LED to finish, reset, G800
Audio can be heard on either F1 or F2 (F0 is the serial terminal)

[Phil__G]

...always remembering to set Teraterm file to 'binary' in the File, Send menu:

[Michael Haardt]

How much RAM does Kitbugplus need? Elektor SC/MP systems have IO at 0x700 and while
I could reconfigure that, of course it would be great if less is needed.

Are you aware of this monitor program recently recovered?

https://github.com/ekuester/SCMP-INS...C-Interpreter/

I just got NIBLFP running on my system. Unlike previous versions, it is now started at its load address at 0D000. Loading a monitor that has a serial load function would make more sense, of course.

Michael

[Phil__G]

Hi Michael, it just uses 0xFFF down to 0xFF6 for the registers, and a few more bytes below that for variables and the stack, I generally regard FE0 as the 'top' of memory to allow some stack space.

Like NS Kitbug and SCIOS, it relies on wrap-around so 0xFFF is an offset of -1 from zero.
The actual amount of memory over and above this, it doesnt care, I have 1k on my Aitken SC/MP, 2k on the micro-introkit board, 8k on the PICL - any size as long as it can see memory at FF0-FFF, thats all it needs.

I hadnt seen the monitor in your link, I had a quick scan and it looks quite complicated code, I will need to sit down and spend some time reading.
I did immediately notice something strange though:

      21/       0 :                     ; ILLEGAL OPCODES
      22/       0 : =0x20                WRITECH = 0x20
      23/       0 : =0x21                READCHR = 0x21

These are Karens own unique pseudo-ops for the PICL, used to relieve the SC/MP of serial I/O and move the task onto the PIC.
I did use these in my original KitbugPlus because it was running on the PICL, but I'm not sure why they would be in some native SC/MP code,
I'm not aware of any SC/MP code anywhere that uses them except for the KitbugPlus monitor I posted for the PICL... maybe all will be clear after a proper read.

How are you finding FP basic? is the speed ok?

[Michael Haardt]

Using memory outside the area of 0x700-0x7ff is fine for Elektor systems, just that space is used for IO. As I said, I could change it, but if that is not used I do not have to.

I just got NIBLFP running, which required making a new boot ROM for my system, so I did not check it much yet. Seeing floating point on the SC/MP is pretty amazing.

All this code was recovered from decades ago and up to now only ran on an emulator to fix recovery errors and build problems. The emulator contains the same opcodes as PICL for tty communication. I am just building a native SC/MP system for the author, because I still had components from building mine. Seeing NIBLFP running natively after all the time was pretty cool.

Getting to a setup that is both compatible with Elektor stuff, but also starts a monitor that can load code as Intel hex would be pretty cool, because BASIC could be loaded over serial that way.

Michael

[Phil__G]

So are Karens pseudo ops not original after all? someone defined them before Karen did in 2007 ?
I had always assumed she created them herself, I thought they were a great idea
In the PICL source she comments that they are 'new pseudo instructions' (2007) but doesnt actually claim them and I'd always just assumed...
do you know who is ukuester?

I tried Intel hex and found the necessary processing was impossibly slow, it needed delays after every byte to allow the SC/MP to process it and after many attempts at streamlining the code i gave up and KitbugPlus now has a very simple binary loader. Its quick & reliable at 1200 baud with no delays at all.
Here is a demo, the windows capture doesnt see pop-up windows so after the 'S800' command you dont see me doing 'File, send' and choosing the 'Life' file but still it gives an idea:
https://youtu.be/0FdaDoIqXpU

[Phil__G]

Wow I just had a read through and Erich's code is a lot to take in, it seems to have everything, cassette i/o, disassembler... really complete but very complex.
I'm not sure but I think it uses a similar technique to the "Strong Arm of P3" technique (described in the Haviland book) as a subroutine handler...
All very clever, way beyond my skills and too much for me to take in for now but thanks for posting the link Michael
Cheers
Phil

[Realtime]

Originally Posted by Phil__G

Its quick & reliable at 1200 baud with no delays at all.
Here is a demo, the windows capture doesnt see pop-up windows so after the 'S800' command you dont see me doing 'File, send' and choosing the 'Life' file but still it gives an idea:
https://youtu.be/0FdaDoIqXpU

Nice demo of 'Life'. Looking forward to building a micro-introkit soon. (btw, your video doesn't show up in searches on youtube or on your channel. I think the reason may be that it's status is 'unlisted' so only viewable if you have a link to it. I guess that can be changed in the video's properties)

[Phil__G]

Hi Ian yes if a simple video is intended just to illustrate a thread like this then I upload them as 'unlisted', otherwise you get random punters complaining that they dont know what its about...

[Phil__G]

I've been working on KitbugPlus's serial I/O with a view to upping the baud rate.

My Aitken SC/MP runs the original NS Kitbug at 300 baud, which is much nicer than the old 110. At 1200, KB+ is much better than 300. 2400 is doable, maybe 4800, 9600?

A minutes cycle counting revealed that 9600 just isnt possible using the standard sense/flag I/O - SIN and SOUT would probably manage it but senseB/Flag0 are standard for NS Kitbug and NIBL so I want to stick to that, which means 9600 is out of the question for now.

I did get it kinda working at 4800 but now the SC/MP started to need a delay between characters to allow it to process. Since Teraterms minimum delay is 1ms, I found there was almost no difference in overall throughput between a reliable 2400 with no delays, and an iffy 4800 + 1ms.

Thing is, as you try ever faster bit rates the inbetween processing time becomes more & more significant and the bit delays that actually 'tune' the serial routines tend towards nothing.

To cut a long story short, I've been running it at 2400 full duplex (ie proper GECO) for a few hours, and so far its been perfectly reliable with no delays at all, the Sload command works great and does actually load twice as fast as 1200.

I'm sure a more capable programmer could get 4800 to run perfectly but 2400 baud and a doubled fast-load speed is fine by me, quite a bit better than the original 110 !!!

I'll update my SC/MP page sometime today, here's a slightly better demo where you can see the menu choices:

https://youtu.be/WX3yQuYx314

Cheers - Phil

[Mark1960]

I followed the link you posted in #34 and saw the link Timbucus posted to practical electronics february 1979, page 60, which I don’t remember reading earlier.
https://worldradiohistory.com/UK/Pra...9-02-S-OCR.pdf

This uses hardware to create trap instructions. Apart from thinking they used too many ICs, I was wondering if all the gaps in the instruction table were treated as NOP by the INS8060. Also what happens for the two byte NOPs, does the second byte still get fetched and is it ignored. Has anyone tried these undefined instruction codes?

[Phil__G]

I've updated KitbugPlus again with a 'C' block copy command for moving memory around.
Can also be used to clear or fill memory. Has the 1200 or 2400 baud console and Serial loader.
Documents, source, hexfiles etc all updated on the SC/MP page of http://philg.uk via the KitbugPlus.zip download.
Cheers
Phil

[Phil__G]

And another KitbugPlus update, this is just a change to the Copy command, it now handles overlapping source & destination blocks, similar to the 'intelligent copy' command in NAS-SYS. I'm quite content with it now, cant think of anything else it desperately needs, its still under 1.5k bytes but I think that will do for now. I would love to know if anyone has tried it. As far as I'm aware its just me & Benny (Audiokit)
Files & updated docs are in the KitbugPlus.zip file on my scamp page.
Cheers
Phil