Yet another MK14 tape interface (ESP32).

[ortek_service]

Originally Posted by Mark1960

FF9-FFF are used to store users pointers and registers before and after calling users code. They are addressed using PC relative addressing from the low area of prom code to avoid using pointers as that would corrupt the user codes pointers. It seems the cassette loader was reusing the address for P1.

Thanks for info on what these are used for. So does that mean 0FF9-0FFF should never really have been used by user programs, if needing to return to the monitor? (And SCIOS v1 also used these?)

It might be useful to have an enhanced Memory-map diagram of:
http://81.174.146.201/acorn/Acorn%20...mory%20Map.pdf - that includes the parts reserved by the OS etc (like IIRC Siclair had in the original Spectrum manual)

[Mark1960]

Probably best not to use FF9-FFF for code as it would be overwritten by pointer and register values on return to the monitor, but could be used for scratch pad if not using the single step function of the monitor. It might be possible to avoid the overwrite by jumping into the monitor at a different address, but pressing reset might have the same effect, though I haven’t read through the code to check.

[SiriusHardware]

So is it better to leave the fast uploader scratch variables at 0FF8-0FFA or move them down into the 'system variables' area 0F00-0F11?

In all cases the use of either area - 0F00-0F11 or 0FF8 - 0FFF for user code carries with it the certainty of corruption of some of the bytes in those areas on return to the monitor - even if the program auto-runs, if the user resets the computer for any reason the monitor will corrupt those areas and it will not be possible to manually re-run the program, it will have to be reloaded and autorun again.

This is just a known hazard of placing user code in either area.

[Realtime]

The variables need to stay at FF8 because they are accessed using negative addressing relative to P2, which conveniently is zero after a reset. As relative addressing can only point +/- 128 bytes it’s not possible to point all the way down to the F00 area. If P2 had to be initialised to point to F00 then the fast loader code would exceed the available space.

[SiriusHardware]

Fair enough, they stay where they are.

[SiriusHardware]

You may ask, under what circumstances would you absolutely have to put code in 0F00-0F11 and / or 0FF8-0FFF?

Imagine an MK14 which has the full onboard 640 bytes of RAM.. and a VDU. You want to load and display a full screen (512 byte) graphic image on the VDU screen.

The only two 256-byte areas you can point the VDU at are 0F00-0FFF and 0B00-0BFF.

In this case you would format the Intel Hex file so that the pixel data bytes intended to go into screen RAM at 0F00-0F11 and 0FF8-0FFF are initially loaded into the I/O RAM instead.

You also load a small block shift program into I/O RAM which, when autorun, will copy the two blocks from I/O RAM to their correct locations within screen RAM, and then place the SC/MP into a forever loop so that it does not return to the monitor.

This is about the only way to display a full 512-byte screen image on an MK14 / VDU which only has onboard RAM.

[SiriusHardware]

Well, so here, as far as I know, are the world's first SCIOS V3 PROMs. I've gone with a different dot colour scheme to distinguish them from the blue / green dots of the late MK14 SCIOS.

U2 is red, U3 is yellow although I really wanted orange (Red=2, Orange=3) but try as I might I could not find anyone selling sticky 5mm orange dots, so yellow it is.

Although I don't have the offboard hardware to test their operation, I have had them installed in my Issue VI and I have used that to look through and check that the replacement code in the tape load / save area is byte for byte the same as in the .LST file attached in Coolsnaz2's post #168.

My programmer lets me load an 8-bit wide file and choose which nibble (upper or lower) to programme into the 4-bit wide PROM - for the benefit of others who may not have that feature on their programmer I have read the two programmed PROMs separately and saved them as two 512-byte files with the PROM code in the lower four bits of each byte, see the attached .zip file.

I'll put these in the post to Coolsnaz2 in the next day or so.

[ortek_service]

The free-download Dataman48Pro(+) etc, that you can run in Demo-mode without having the programmer, also has quite flexible load /save options.
And I recall previously using this to split PROM files into separate High/Low Nibble ones for Chris, when we wanted to verify against existing ones before committing to fresh PROM's.

[coolsnaz2]

Just a couple of pictures showing my MK14 issue VI board with SCIOS v3 PROMs installed and working. Thanks to all forum members who have helped me with this project. I can now load programs within seconds of powering on the MK14, so much better having the Fast Loader installed in the PROMs.

If you execute the Fast Loader on the MK14 and hit the button on the M5Stack Atom Lite the last 8 digits of the web server IP address are displayed. This helps you connect your web browser to the server if the router gives out dynamic IP addresses.

[SiriusHardware]

Nice to see that working!

Coolsnaz2 very kindly sent me some pre-programmed external hardware to try this setup out, although I will have to make myself a pair of SCIOS V3 PROMs to get the full benefit of it. I'm tied up with a visitor until some time next week so unfortunately I won't be able to play properly until then, but I'm looking forward to it.

Coolsnaz2, if you now feel that the M5STICK and ATOM side codes are 'stable' perhaps you could include them in a post here along with the documentation - you can attach any sort of file to a post as long as it's in a zip file that does not exceed the maximum allowed size for a .zip file.

It's up to you of course, but it would be good if all of the required code and documentation could be found in one place / one thread.

[Realtime]

Originally Posted by coolsnaz2

If you execute the Fast Loader on the MK14 and hit the button on the M5Stack Atom Lite the last 8 digits of the web server IP address are displayed. This helps you connect your web browser to the server if the router gives out dynamic IP addresses.

That’s brilliant. I keep finding the IP address is re-allocated each time I power on. Look forward to trying that in due course.

[coolsnaz2]

I will publish the ESP32 code and documentation soon on this thread, just need to complete some documentation to accompany the code. If anyone wants a pre-release copy, they can PM me.

ChrisOddy has obtained PROMs so he can start producing SCIOS V3, and is happy to supply them to forum members for £12 for a pair of PROMs incl postage, members will need to PM ChrisOddy with PayPal details.

[SiriusHardware]

Kind of Chris to offer that service - I don't think my own limited supply of Philips/Signetics PROMs would have been able to stretch to everyone that might want V3 PROMs so I'm relieved that Chris has been able to step in.

I haven't forgotten that I need to make myself a V3 PROM set so I can play with the modules you sent, but I haven't had a lot of time to myself lately.

[SiriusHardware]

A question for Coolsnaz2 or anyone who already uses his fast loading system with an M5stickCplus - which of the connections G32 and G33 go to which connections on the issue VI rear? Coolsnaz2 was good enough to email me a draft of the manual for this version of his loading system but unfortunately it is in docx format and I don't have MS Office on any machine here at home, so I can't open it.

[SiriusHardware]

Update: Coolsnaz2 has contacted me with the relevant details from his as yet un-finalised user manual for this version of his loader system.

M5 StickC Plus HY2.0-4P (Grove Connector) to MK14 connections.
Connect pin 1 (Black) to MK14 GND.
Connect pin 2 (Red) to MK14 5V.
Connect pin 3 (G32)(Yellow) to MK14 SIN edge connector 27.
Connect pin 4 (G33)(White) to MK14 Sense B, edge connector 28.

I can also now confirm that the separate PROM code files I posted in #187 for the V3 OS containing Coolsnaz2's fast loader code are good - I was sure they were, but I have now been able to test my own V3 PROMs which were programmed from those files.

[Brian of Romsey]

Originally Posted by SiriusHardware

... unfortunately it is in docx format and I don't have MS Office on any machine here at home, so I can't open it.

I would humbly suggest LibreOffice. It is free, well supported, in continuous development (so bugs are few and soon squashed). It can easily read MS formats like .doc and .docx. Available for Mac, Linux, and Windows. Download, install, and enjoy. Easy.

https://www.libreoffice.org/

Cheers, Brian.

[SiriusHardware]

Thanks Brian - I have a habit of keeping my various devices very clean with nothing on them that I do not regularly use. I'm aware of LibreOffice but have so little use for office type software in general that it would almost never be used. My go to 'Word Processor' is Wordpad - that's all I've ever used to write correspondence, etc for the entire time that I have been using Windows.

I've suggested to Coolsnaz2 that .PDF would be a better format for the finished document, as virtually all devices now have the 'native' ability to view PDFs via already included software (like Browsers, or some other integral utility.)

[ScottishColin]

I now have one of these devices. I have the Realview device too, so a simple way of fitting it would be to solder the cables to the back of the Realview device. I was wondering whether it would be feasible to create some kind of breakout board that would allow simpler connections to be made to the ESP32 and then attach to the Realview - perhaps made from a Vero board?

My thinking is that I would like to be able to separate these two devices without necessarily having to desolder them.

Thanks.

Colin.

[SiriusHardware]

Did you get a 'Grove' connector with the M5 which breaks out the GND, 5V and two I/O pins from the M5 on a four-way male pin connector?

[Silvester]

Re: M5 StickC Plus, You can get suitable proto boards:

https://thepihut.com/products/m5stickc-proto-hat
https://thepihut.com/products/new-mini-proto-unit

The first is probably the best to use because the Grove connection for 5V is OUT (so not suitable to link to MK14 5V). The other connector allows 5V IN (to preserve/charge battery). The only hitch is that the output pins will need to be changed (G32/33 on Grove, G25/26/0 on 8 pin).