Matchbox sized 6502 / Z80 / 6809 Co Pro

discussion of games, software, hardware & emulators relating to the Acorn Atom
User avatar
danielj
Posts: 5364
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby danielj » Wed Jan 20, 2016 1:31 pm

65C102 won't work. You can get them for about $5/go form utsource 65C02P2/3 or 4 will do the job, but there's a $10 minimum order from them, plus p&p (about $4). I have a sneaking suspicion that Dave Hitchins may have some?

d.

User avatar
hoglet
Posts: 6629
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby hoglet » Wed Jan 20, 2016 2:43 pm

hoglet wrote:In the Matchbox Co Pro, were getting short on space in the FLASH ROM (we're actually now hitting the limit of 12 designs).

One of the things we're considering is creating single versions of the 6502, Z80 and 6809 designs that combine internal memory with some kind of banked access to external memory. The idea would be to run the processor as fast as internal memory would allow, and then add a couple of wait states to slow things down when external memory was accessed.

It would be nice if we could adopt a common approach across the 6502, Z80 and 6809 if that made sense

As for OSs, I think there is interest in:
- CPM on the Z80
- FLEX on the 6809
- Your Fuzix on all of the above

Do you have any suggestions about how we organize the banking? i.e. how much of the memory map, how many independent blocks, block size, whether any kind of protection attributes would be useful, what the MMU registers should look like?

Full response in the other thread
EtchedPixels wrote:If I wanted to make the 8bit one totally standard I'd go with 8x8K with 8 ports giving the top bits for that 8K bank. CP/M and MP/M can handle it, Fuzix can use the pages in pairs, OS/9 and NitrOS9 should be happy bunnies.

I think 16x4K would also be feasible, if that would be more natural for Fuzix. With 1MB of external memory, you end up with an 8-bit value in each of 16 "MMU" registers. Somehow we need to incorporate the 64KB faster internal into the physical memory map, possibly by just loosing 64KB of the external memory.

The reason I'm starting with 1MB of external memory rather than the full 2MB is to avoid having to make every external RAM write cycle a read-modify-write (the 32 bits SRAM interface doesn't have separate byte enables).

It would be fun to get Fuzix running on the 6502 (even though I guess the Z80 would be an easier target).

Alan, could you say a bit more about the state of 6502 platform support in Fuzix, and how the banking is done in the tgl6502 that you are using?

Dave

EtchedPixels
Posts: 22
Joined: Mon Feb 23, 2015 11:39 pm

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby EtchedPixels » Wed Jan 20, 2016 3:19 pm

hoglet wrote:I think 16x4K would also be feasible, if that would be more natural for Fuzix. With 1MB of external memory, you end up with an 8-bit value in each of 16 "MMU" registers. Somehow we need to incorporate the 64KB faster internal into the physical memory map, possibly by just loosing 64KB of the external memory.

The reason I'm starting with 1MB of external memory rather than the full 2MB is to avoid having to make every external RAM write cycle a read-modify-write (the 32 bits SRAM interface doesn't have separate byte enables).


At that point you need a cache, which may well be the right thing for a system with a mix of fast local and slow external RAM. Your cache is 32bits wide so your external bus cycles are always 32bit. But that's another distraction 8)

It would be fun to get Fuzix running on the 6502 (even though I guess the Z80 would be an easier target).

Alan, could you say a bit more about the state of 6502 platform support in Fuzix, and how the banking is done in the tgl6502 that you are using?

Dave


16 x 4K doesn't actually help anything I am aware of. In fact there is a tradeoff because the more banks you have the slower it is to bank switch. 8-16K seems to be the sweet spot trading off wasted memory for efficiency on 1-4MHz 8bit processors.

The tgl6502 is just 4 x 16K banks. The port has bit rotted slightly but just needs bringing back into sync and the signal handlers fixing. Bank switching on the tgl 6502 is interesting because you have to switch the low 16K bank, your stack and your zero page all in one. I had to debug it with instruction traces because if it goes wrong you've got pretty much no idea how you ended up where you did.

Alan

User avatar
BigEd
Posts: 1499
Joined: Sun Jan 24, 2010 10:24 am
Location: West
Contact:

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby BigEd » Wed Jan 20, 2016 3:50 pm

Would it help if page zero and the stack were always mapped from the same (internal, fast) memory, such that the first slot for banking is actually a little short?

EtchedPixels
Posts: 22
Joined: Mon Feb 23, 2015 11:39 pm

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby EtchedPixels » Wed Jan 20, 2016 4:15 pm

I think that would make it worse as you'd have to copy it all somewhere each taskswitch !

User avatar
BigEd
Posts: 1499
Joined: Sun Jan 24, 2010 10:24 am
Location: West
Contact:

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby BigEd » Wed Jan 20, 2016 4:33 pm

Oh I see - you prefer to swap it, but the TGL6502 holds it separately from the banking. (For performance presumably, the same reason I thought it was preferable. So, we can trade application performance for task switching cost?)

User avatar
hoglet
Posts: 6629
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby hoglet » Wed Jan 20, 2016 4:39 pm

BigEd wrote:Would it help if page zero and the stack were always mapped from the same (internal, fast) memory, such that the first slot for banking is actually a little short?

We can implement pretty much anything we need here, and pretty quickly once I'm clear on the design.

I think it would make sense, performance wise, if page zero and the stack were always in internal memory.

I would imagine each Fuzix process would need it's own zero page / stack, i.e. these regions need to be moveable, maybe via an 8-bit address register that could re-map this to any 256 byte page in internal memory. It might make sense to have separate address registers for zero page and the stack.

So, so far it's looking like:
- Initially 1MB of external RAM (i.e. 128 8K pages)
- Zero page address register (256 byte page in 64KB internal RAM)
- Stack address register (256 byte page in 64KB internal RAM)
- 8K Page 0 address register, controls mapping of 0x0200-1FFF
- 8K Page 1 address register, controls mapping of 0x2000-3FFF
- 8K Page 2 address register, controls mapping of 0x4000-5FFF
- 8K Page 3 address register, controls mapping of 0x6000-7FFF
- 8K Page 4 address register, controls mapping of 0x8000-9FFF
- 8K Page 5 address register, controls mapping of 0xA000-BFFF
- 8K Page 6 address register, controls mapping of 0xC000-DFFF
- 8K Page 7 address register, controls mapping of 0xE000-FFFF

Maybe bit 7 of the page address register would control whether the mapping was to internal or external memory.

So, another question is regarding the top end of memory....

In the Acorn world we have:
- tube hardware registers at 0xFEF8-0xFEFF
- MOS entry block 0xFF00-0xFFF9
- 6502 hardware vectors at 0xFFFA-FFFF

What would this need to look like in the Fuzix world? Would the top block be common to all processes?

I'll stop there for now. So many questions :D

Dave

User avatar
hoglet
Posts: 6629
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby hoglet » Wed Jan 20, 2016 4:57 pm

EtchedPixels wrote:The tgl6502 is just 4 x 16K banks. The port has bit rotted slightly but just needs bringing back into sync and the signal handlers fixing. Bank switching on the tgl 6502 is interesting because you have to switch the low 16K bank, your stack and your zero page all in one. I had to debug it with instruction traces because if it goes wrong you've got pretty much no idea how you ended up where you did.

Is this the right place to be looking for the 6502 Fuzix Port? (I'm assuming the tgl6502 is the only 6502 target at the moment):
https://github.com/EtchedPixels/FUZIX/b ... /tgl6502.s

Seems like there are address registers at:
0xFF8A - controls mapping of 0x0000-0x1FFF
0xFF8B - controls mapping of 0x2000-0x3FFF
0xFF8C - controls mapping of 0x4000-0x5FFF
0xFF8D - controls mapping of 0x6000-0x7FFF
0xFF8E - controls mapping of 0x8000-0x9FFF
0xFF8F - controls mapping of 0xA000-0xBFFF
0xFF90 - controls mapping of 0xC000-0xDFFF
0xFF91 - controls mapping of 0xE000-0xFFFF

I guess the linker map is also something we should understand:
https://github.com/EtchedPixels/FUZIX/b ... 2/ld65.cfg

Dave

EtchedPixels
Posts: 22
Joined: Mon Feb 23, 2015 11:39 pm

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby EtchedPixels » Wed Jan 20, 2016 5:34 pm

hoglet wrote:
BigEd wrote:Would it help if page zero and the stack were always mapped from the same (internal, fast) memory, such that the first slot for banking is actually a little short?


- Zero page address register (256 byte page in 64KB internal RAM)
- Stack address register (256 byte page in 64KB internal RAM)



Those two would be sweet for 6502, and harmless for everyone else: 6809 has a direct page but the CPU controls it already via the DP register.

In the Acorn world we have:
- tube hardware registers at 0xFEF8-0xFEFF
- MOS entry block 0xFF00-0xFFF9
- 6502 hardware vectors at 0xFFFA-FFFF

What would this need to look like in the Fuzix world? Would the top block be common to all processes?
Dave


I think that's a more general question. Fuzix at least doesn't care too much (just not in the middle of memory!). Some of the 6809 operating systems assume elements of the I/O space allocation.

The only tricky bit I can see is the NMI handling. Getting the hardware vectors copied into multiple banks is easy enough (just map them low and write them before using them). The more fun question would be whether there is a case where we have to handle an incoming interrupt (especially a tube NMI) reliably when we might be mid bank or stack switch.

If you have to change bank and stack pointer you can't do it atomically together so there is going to be a point in time where pushing the return address will be very bad indeed.

EtchedPixels
Posts: 22
Joined: Mon Feb 23, 2015 11:39 pm

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby EtchedPixels » Wed Jan 20, 2016 5:35 pm

hoglet wrote:
EtchedPixels wrote:I guess the linker map is also something we should understand:
https://github.com/EtchedPixels/FUZIX/b ... 2/ld65.cfg

Dave


Yes but don't get hung up on it. It's just how it was laid out on that machine !

Alan

User avatar
hoglet
Posts: 6629
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby hoglet » Wed Jan 20, 2016 5:57 pm

EtchedPixels wrote:The only tricky bit I can see is the NMI handling. Getting the hardware vectors copied into multiple banks is easy enough (just map them low and write them before using them). The more fun question would be whether there is a case where we have to handle an incoming interrupt (especially a tube NMI) reliably when we might be mid bank or stack switch.

If you have to change bank and stack pointer you can't do it atomically together so there is going to be a point in time where pushing the return address will be very bad indeed.

The only thing that NMI is used for on the Acorn Tube platform is handing host initiated data transfers in/out of parasite memory (e.g. during loading or saving of a file to the host using OSCLI/OSFILE, or a low-level reading/writing a disk sector on the host using OSWORD 7F). So these don't just happen out-of-the-blue, and they are always preceded by a data transfer request (via tube R4), which causes an IRQ interrupt.

We would, I guess, need to write a new device driver that mapped to OSWORD 7F calls over the tube, which then read/write the storage device attached to the host.

Dave

EtchedPixels
Posts: 22
Joined: Mon Feb 23, 2015 11:39 pm

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby EtchedPixels » Wed Jan 20, 2016 6:27 pm

I think everyone needs approximately the same set of low level driver bits

- read disc sectors
- write disc sectors
- check disc present
- CP/M 3 would also like "media change" events
- keyboard / serial / etc events
- console / serial out
- read real time clock / system clock if present
- enabling/disabling interrupts/events for these and timer

Doing CP/M well also involves a whole crapload of trying to read sectors that may not be present, figuring out the disc desnsity and the like. MSDOS put the media type information on the disc in a known place. CP/M unfortunately didn't so while it can support multiple media types it's a barrel of laughs doing so.

User avatar
jgharston
Posts: 2764
Joined: Thu Sep 24, 2009 11:22 am
Location: Whitby/Sheffield

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby jgharston » Thu Jan 21, 2016 5:47 pm

hoglet wrote:I think 16x4K would also be feasible, if that would be more natural for Fuzix. With 1MB of external memory, you end up with an 8-bit value in each of 16 "MMU" registers. Somehow we need to incorporate the 64KB faster internal into the physical memory map, possibly by just loosing 64KB of the external memory.

I'll vote for 16x4K as well. It fits naturally, you get 256 pages in 1M of ram, so a single whole 8-bit register for each bank, and each bank within 64K is addressed by a single nybble, &0xxx, &1xxx, &2xxx, etc. Plus, at long last, I'd have a Z80 system with my Z80 memory mapping system that I've always been meaning to build and can finish porting CP/M 3 to it.

With a Z80 system it's also natual to use as the I/O address the same address as the memory bank that is being paged, so OUT (&0xzz),nn for the memory bank at &0xzz, OUT (&1xxx),nn for the memory bank at &1xxx, etc., with the bottom 8 bits of the I/O address being the banking I/O port at the top 8 bits of the I/O address being the bank. (link)

Code: Select all

$ bbcbasic
PDP11 BBC BASIC IV Version 0.25
(C) Copyright J.G.Harston 1989,2005-2015
>_

User avatar
jgharston
Posts: 2764
Joined: Thu Sep 24, 2009 11:22 am
Location: Whitby/Sheffield

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby jgharston » Thu Jan 21, 2016 5:59 pm

EtchedPixels wrote:- CP/M 3 would also like "media change" events

That requires a soldering iron, as you need a disk drive that provides a media change line, and you'd need to modify the Beeb to read the media change line, and decide how and where in the I/O map it's read. The Acorn CP/M 2 BIOS just uses TIME and assumes that after 2.56cs the last accessed disk is no longer live.

EtchedPixels wrote:Doing CP/M well also involves a whole crapload of trying to read sectors that may not be present, figuring out the disc desnsity and the like.
Just call OSWORD &7F and let it deal with it. 'Read Sector IDs' tells you what sectors are present (see the BBC Flex code (link)). If you have double-density disk hardware you can use OSWORD &7F to read double density. If you have single density hardware, then its simple - you can't read anything other than single density.

Code: Select all

$ bbcbasic
PDP11 BBC BASIC IV Version 0.25
(C) Copyright J.G.Harston 1989,2005-2015
>_

User avatar
daveejhitchins
Posts: 3692
Joined: Wed Jun 13, 2012 5:23 pm
Location: Newton Aycliffe, County Durham

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby daveejhitchins » Thu Jan 21, 2016 6:06 pm

danielj wrote:65C102 won't work. You can get them for about $5/go form utsource 65C02P2/3 or 4 will do the job, but there's a $10 minimum order from them, plus p&p (about $4). I have a sneaking suspicion that Dave Hitchins may have some?
Sorry, not one I have . . .

Dave H :D
Parts: UM6502CE, GAL22V10D, GAL16V8D, AS6C62256A, TC514400AZ, WD1772, R6522, TMS27C512, AT28C256
Products: ARA II, ABR, ATI, AP6, MGC, AP5 . . .
For a price list, contact me at: Retro Hardware AT dave ej hitchins DOT plus DOT com

User avatar
flynnjs
Posts: 757
Joined: Tue Jul 06, 2010 9:33 pm

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby flynnjs » Thu Jan 21, 2016 7:02 pm

I'm going to order a batch from UTsource and hope they're genuine.
I've heard a lot of reports of duff/fake parts from them.
I'll get 10 to start with.

User avatar
danielj
Posts: 5364
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby danielj » Thu Jan 21, 2016 7:08 pm

Touch wood, everything I've had from UTSource has been fine...

d.

User avatar
hoglet
Posts: 6629
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby hoglet » Mon Jan 25, 2016 6:12 pm

Hi Jonathan,
jgharston wrote:Disassemblies so far at http://mdfs.net/Info/Comp/DRDOS/

Now I'll go up and read the last few posts ;)

Some of the disassemblies you had up here of DRDOS seem to have gone missing.

I'm particularly interested in the one of segment 06bb.

Dave

User avatar
jgharston
Posts: 2764
Joined: Thu Sep 24, 2009 11:22 am
Location: Whitby/Sheffield

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby jgharston » Mon Jan 25, 2016 9:13 pm

hoglet wrote:Hi Jonathan,
jgharston wrote:Disassemblies so far at http://mdfs.net/Info/Comp/DRDOS/
Now I'll go up and read the last few posts ;)
Some of the disassemblies you had up here of DRDOS seem to have gone missing.
I'm particularly interested in the one of segment 06bb.

Looks like I forget to set the public 'R' attribute. Should be ok now.

Sorry, won't be doing that as investigating why software doesn't work in a particular environment and making it work is plagiarism.

Code: Select all

$ bbcbasic
PDP11 BBC BASIC IV Version 0.25
(C) Copyright J.G.Harston 1989,2005-2015
>_

User avatar
fordp
Posts: 923
Joined: Sun Feb 12, 2012 9:08 pm
Location: Kent, England

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby fordp » Wed Jan 27, 2016 10:46 am

jgharston wrote:Sorry, won't be doing that as investigating why software doesn't work in a particular environment and making it work is plagiarism.


Damn, I spend a lot of time fixing bugs for a living.

I am not a big fan of crossword puzzles as they are already solved before you get them. Software is my puzzle and doing something that has never been done before is my motivation.

Computer preservation is an honourable and fine pursuit even if the crazy copyright laws mean that what we preserve may not be technically legal to use for many years.

Keep up the good work guys ;)
FordP (Simon Ellwood)
Time is an illusion. Lunchtime, doubly so!

User avatar
1024MAK
Posts: 6795
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby 1024MAK » Wed Jan 27, 2016 2:44 pm

@Ford

I'm sorry to say that this thread is suffering fallout from events elsewhere on the forum.

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

User avatar
fordp
Posts: 923
Joined: Sun Feb 12, 2012 9:08 pm
Location: Kent, England

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby fordp » Wed Jan 27, 2016 9:59 pm

1024MAK wrote:@Ford

I'm sorry to say that this thread is suffering fallout from events elsewhere on the forum.

Mark


Indeed,

Luckily we have a talented and great set of people round here. I hope we can all crack on with our fun and rewarding computer preservation hobby. Forums are notorious for people interpreting our words in a way they were not intended.

Still I have the fastest BBC Micro in history in my cellar ;)
FordP (Simon Ellwood)
Time is an illusion. Lunchtime, doubly so!

User avatar
1024MAK
Posts: 6795
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby 1024MAK » Wed Jan 27, 2016 11:38 pm

fordp wrote:Luckily we have a talented and great set of people round here. I hope we can all crack on with our fun and rewarding computer preservation hobby.
Absolutely :D
fordp wrote:Forums are notorious for people interpreting our words in a way they were not intended.
Yes, but thankfully there has not been too much trouble on *.
Let's hope things get back to normal from now on.

BTW, I do try to keep up with the threads that you, Dave and Jonathan post in, even if the ARM assembly source code goes far above my head :lol:

Keep up the good work 8)

fordp wrote:Still I have the fastest BBC Micro in history in my cellar ;)
Hey, we have to give the popcorn suppliers time to restock... :lol: Calm down :lol:

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

harrowm
Posts: 121
Joined: Sun Nov 30, 2014 11:07 pm
Location: Singapore

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby harrowm » Thu Jan 28, 2016 4:04 pm

Can you point me to the thread that has further details ?

I don't understand the legal precedent.

Regards
Malcolm

paulb
Posts: 784
Joined: Mon Jan 20, 2014 9:02 pm

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby paulb » Tue Feb 02, 2016 10:54 pm

harrowm wrote:Can you point me to the thread that has further details ?

I don't understand the legal precedent.


I think it may have something to do with one forum contributor modifying and distributing another forum contributor's code and then making some private remarks about that code which accidentally became public, thus annoying the latter forum contributor (more with the nature of the remarks than anything else, as I understand it). But as long as the licensing applied to people's code permits modification and redistribution, there shouldn't be any legal issues.

Social issues on the other hand... :roll:

User avatar
1024MAK
Posts: 6795
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby 1024MAK » Tue Feb 02, 2016 11:05 pm

paulb wrote:I think it may have something to do with one forum contributor modifying and distributing another forum contributor's code and then making some private remarks about that code which accidentally became public, thus annoying the latter forum contributor (more with the nature of the remarks than anything else, as I understand it). But as long as the licensing applied to people's code permits modification and redistribution, there shouldn't be any legal issues.

You forgot to sprinkle "alleged" in front of a number of your words here :roll:

Right enough, else we risk the thread becoming locked. So back to the topic please.

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

harrowm
Posts: 121
Joined: Sun Nov 30, 2014 11:07 pm
Location: Singapore

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby harrowm » Wed Feb 03, 2016 7:19 am

Got it. So to get back on topic, what about a copro based on this:

http://www.adapteva.com/andreas-blog/wh ... next-chip/

Malcolm

iainjh
Posts: 286
Joined: Mon May 14, 2012 11:18 am

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby iainjh » Thu Feb 11, 2016 11:39 pm

Hi, a few posts ago I reported continuing instability with my datacentred Model B, matchbox Co-pro and *XADFS and *YADFS (adfs's 1.32 and 1.33 both loaded into swram i think)

Jason kindly sent me a CMOS 6502 that stabilised matters in basic but I couldnt still use the datacentre or access it's IDE with the co-pro active. I wanted to try loading ADFS not into swr and see if that was the issue.

well;

Bas thecorfiot burnt me an ADFS (ide patched 1.33) onto eprom and..it works

*fx 151,230,0
control break
*adfs
*.
dir's the cf card on the ide

I can even *run !boot

which loads the retroclinic games menu, and run co-pro elite straight off the CF, and it works.

(with *ADFS, shift-break obviously(?) tries to boot from floppy not the !boot on the cf/hard drive. *XADFS and *YADFS allow shift-break off the CF directly, not sure if theres a way around that?)

so now Ive bought myself an eprom programmer and I'll collate myself a set of roms I need (1770 dfs, ramfs 1.01, this adfs etc) to keep adfs in rom and not swram. for me this seems a way forward..


cheers all for help :)

User avatar
hoglet
Posts: 6629
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby hoglet » Fri Feb 12, 2016 7:43 am

Hi Iain,

I'm glad you, Jason and Bas managed to get this sorted.

I wonder what the problem with the sideways RAM was? I'm able to run ADFS from sideways RAM without any problems. It sounds like something is managing to corrupt ADFS. Is the sideways RAM otherwise reliable?

It might be worth digging a bit further. e.g. after you experience the ADFS error, SAVE a copy of the ADFS sideways RAM, so it can be compared against the original. Sometimes there are clues in the pattern of corruption.
iainjh wrote:its not got a ramrom board, the main board has been modified by retroclinic to allow, the flysheet says: 10 sideways rom slots (up from standard 4), the machine contains 3 roms. first is os 1.20, then 2 'combiroms' with 2 sw ram slots at 15 and 11.

'Do you have any more details on this? The flysheet?

Does it allow the sideways RAM to be write protected?

Mark (retroclinic) any comments or thoughts? (if you are following this...)

Dave

User avatar
fordp
Posts: 923
Joined: Sun Feb 12, 2012 9:08 pm
Location: Kent, England

Re: Matchbox sized 6502 / Z80 / 6809 Co Pro

Postby fordp » Fri Feb 12, 2016 7:46 am

iainjh wrote:
so now Ive bought myself an eprom programmer and I'll collate myself a set of roms I need (1770 dfs, ramfs 1.01, this adfs etc) to keep adfs in rom and not swram. for me this seems a way forward..


cheers all for help :)


I got myself a programmer this week, but I have also ordered 5 32K EEPROMs from EBay so once I have the right code in one I should not need the programmer again :D
FordP (Simon Ellwood)
Time is an illusion. Lunchtime, doubly so!


Return to “acorn atom”

Who is online

Users browsing this forum: No registered users and 1 guest