SymbOS 4.0 released

[darkhalf]

@Prodatron unfortunately I've had issues with my current setup using symbos 4.0 version with my G9K setup. Finally found some time to isolate the problem

Hardware Configuration: CPC464 + Piotrs DDI4 interface (comprising of DDI clone + 512RAM) + Technobytes V9990 board + M4 board 

Software Config:  New M4 SD card and copied the latest 4.0 SymbOS-CPC40-MassStorage straight onto it

Issues: Only displays 'start' button and task bar when running SYMG9K 

Notes: This setup works fine with upto the last V3 revision (I was testing with 15 Oct 24 build prior) 

I've removed the DDI4 and run directly with the M4 and revaldinho's 512MB RAM expansion in different RAM modes and that is fine. 

So the incompatibility is with the DDI4 itself. I've got a DDI1 but borrowed the chip for my 464+ so can't test that one

Also another interesting error was originally I called the card drive label 'SYMBOS' and was getting CODE 33 (folder errors) and I think the label might have been causing this. I reformatted the card and called it something else and the problem went away

[Prodatron]

Hi @darkhalf , thanks for the report!

I've removed the DDI4 and run directly with the M4 and revaldinho's 512MB RAM expansion in different RAM modes and that is fine.

So the incompatibility is with the DDI4 itself.

So 3.x works with a DDI4 but 4.0 doesn't anymore, hm. I will have a look, what might be special with the DDI4...

Also another interesting error was originally I called the card drive label 'SYMBOS' and was getting CODE 33 (folder errors) and I think the label might have been causing this. I reformatted the card and called it something else and the problem went away

Wow, that's a special find!

I didn't know, that a "volume ID" entry in the root directory can have the same name like a file or subdirectory entry in the same directory. It's not possible for a file and a subdirectory having the same name, but for a volume ID it is.

So when SymbOS tries to open a subdirectory it goes through all directory entries and searches for the name. When it finds the name, it checks, if it is a subdirectory. If it isn't, you get this error message, as it assumes, that it is a file.
Now with the knowledge, that a "volume ID" can have the same name like a file or a subdirectory I will have a look if this can be fixed.

[Brocky]

hackaday finally found SymbOS!

https://hackaday.com/2025/07/19/symbos-is-a-funhouse-mirror-look-at-a-future-that-never-was/

[Gryzor]

hackaday finally found SymbOS!

https://hackaday.com/2025/07/19/symbos-is-a-funhouse-mirror-look-at-a-future-that-never-was/

Good thing I searched before posting it myself, though I think it warrants its own thread😁

[Prodatron]

hackaday finally found SymbOS!

https://hackaday.com/2025/07/19/symbos-is-a-funhouse-mirror-look-at-a-future-that-never-was/

Thanks for the link, very nice article!

It's cool that they mentioned the new SymbOS AI client for ChatGPT and other LLMs by @prevtenet; this information was only posted here on CPC-Wiki so far. As it's still beta I didn't put it on the website yet. So it's interesting that they know/read (?) this forum?

[darkhalf]

So 3.x works with a DDI4 but 4.0 doesn't anymore, hm. I will have a look, what might be special with the DDI4

@Prodatron I've done some more investigation:

The problem is this hardware/software combination: AMSDAP board + V9990 board + M4 board + DDI4 board with V4 of SYMBOS. 

It works fine if:
Go back to SYMBOS V3
Remove the DDI4 board
Remove the AMSDAP board (and run standard MX4 board)
Remove the M4 board (and copy the 9K files onto disk and run it from there)
Run in normal CPC video mode (only a problem in V9990 mode)

Also got it working fine with CPC464 standard backplane + V9990 + DD1 + 512MB expansion + M4 board

I've posted previously about AMSDAP and hardware incompatibility issues and SYMBOS, so I suspect  this is another variation 

Attached is a pic of my 464 running V3 fine with this setup

[tubs74]

ZX Spectrum Next Port

- 90% classic ZX Spectrum port: SymbOS ZNX is mostly a port for the classic Sinclair ZX Spectrum, except for using the advanced memory banking and display capabilities of the Spectrum Next.

Hi, does this mean there might be a classic spectrum (239/+2a,/+3) version in the works?

[Prodatron]

ZX Spectrum Next Port

- 90% classic ZX Spectrum port: SymbOS ZNX is mostly a port for the classic Sinclair ZX Spectrum, except for using the advanced memory banking and display capabilities of the Spectrum Next.

Hi, does this mean there might be a classic spectrum (239/+2a,/+3) version in the works?

Good question, that's something which is pending since many years. There is a finished screendriver for the ZX Spectrum (256x192, 2 colours), so the only missing thing would be the memory mapping. For this some special hardware is needed, which hasn't been done for the classic Spectrum before. I always hear different stories: Some say, that such memory mapping is not possible (swap the whole visible 64K with any part of a memory expansion), some say, it is. If it is, and we will see such a hardware, a port would be possible very quickly.

[andycadley]

Good question, that's something which is pending since many years. There is a finished screendriver for the ZX Spectrum (256x192, 2 colours), so the only missing thing would be the memory mapping. For this some special hardware is needed, which hasn't been done for the classic Spectrum before. I always hear different stories: Some say, that such memory mapping is not possible (swap the whole visible 64K with any part of a memory expansion), some say, it is. If it is, and we will see such a hardware, a port would be possible very quickly.

For the 16K, 48K, 48K+, 128K+ and +2 (the grey one) The lower 16K is always ROM (on machines with 128K it can be one of two ROMs). It is not possible to have RAM here.

For the +2A, +2B (the black +2s) and +3 the lower 16K is usually one of four ROMs, but there is a special "all RAM" mode that was basically intended for CP/M. But there are only a few limited paging configurations supported like that so it's not exactly flexible (especially because display RAM has to come from one of two pages) 

[d_kef]

In fact this is the only memory area (0000 - 3FFF) that can be easily used by a RAM expansion. The ROM in all ZX Spectrum models can be paged out using the ROM disable pin(s) in the expansion port. Then the interface can page in any 16K page of RAM much like a CPC does for memory modes >= C4.
Of course paging in/out a whole 64K bank is close to impossible without (heavy) motherboard modification. 

d_kef

[andycadley]

Yes, I should have been more precise. Without extra hardware only the +2A/B and +3 can put RAM in the lower 16K. With extra hardware, you can replace the ROM (it was intended for Interface 2 ROM cartridges, but you can use RAM instead) although I'm not entirely sure that works on the later models (compatibility there was always a bit funky). 

[Prodatron]

although I'm not entirely sure that works on the later models (compatibility there was always a bit funky).

Thanks for your answers!

Later models are using different pins for disabling the ROM:
https://sinclair.wiki.zxnet.co.uk/wiki/ZX_Spectrum_edge_connector#ROM_disable_pins

But if I got you both correctly, it's impossible to replace the upper 48K with external RAM? The Spectrum 48K, Plus, 128K never had memory expansions??

[andycadley]

The 128K spectrums can put any page in the top 16K (from C000 onwards) - the middle 48K is fixed.

I don't believe there were ever any further RAM expansions (there are some obscure 80K RAM expansions for the 48K but they're so rare as to be irrelevant).