Problems with Datacentre (I think!) with certain games

for bbc micro/electron hardware, peripherals & programming issues (NOT emulators!)
User avatar
leenew
Posts: 4016
Joined: Wed Jul 04, 2012 3:27 pm
Location: Doncaster, Yorkshire
Contact:

Problems with Datacentre (I think!) with certain games

Post by leenew » Thu Aug 25, 2016 4:53 pm

Hi,
I wonder if anybody may be able to help?
We are trying to achieve the goal of having every game on the bbcmicro.co.uk site playable online on JSBEEB, also downloadable and playable in BeebEm or B-Em, also playable from a physical disc on a standard beeb, and lastly, playable vis an SSD image via datacentre.
*most* games are fine, which is a relief!
However, we are struggling with the disc versions of the Level 9 text adventures at the moment.
The first one we are trying to tackle is "Return to Eden".
BillCarr2005 has the original disc versions, which are protected, and need de-protecting, whilst at the same time, retaining the ability to load and save the game properly.
BillC has identified that the save-game saves directly back to disc on certain sectors and he has sent me an image that he says works on his machine via datacentre.
However, it fails to load on my machine.
Could anyone with a datacentre try the image on their machine and try to play the game, and specifically save/reload the game and see if it behaves as expected so we can narrow down where it is going wrong?
Mark Haysman seems to think datacentre is unable to emulate the direct writes to the 8271, but if this is the case, why does it work on 1 datacentre and not the other?
I am not very good with technical talk regarding disc sectors and OSWORD &7F calls etc. etc, so if BillCarr could chime in with the techie talk please, that would help.
If the game detects it is being loaded from a physical disc, BillC has coded for D0 to be displayed on screen, and if datacentre is detected, DC should be displayed on screen.
Hope this can be resolved.
Thanks,
Return to Eden LOADER at FF0900.zip
(64.67 KiB) Downloaded 56 times
Lee.

User avatar
hoglet
Posts: 8825
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by hoglet » Thu Aug 25, 2016 5:11 pm

I can try this over the weekend at Cambridge (my Beeb is packed away now).

As far as I know, all the level 9 adventures use for saving and loading is OSFILE. There is no low level disk access at all.

I've also ported all of these to the Atom without any problems:
https://github.com/hoglet67/AtomSoftwar ... ter/level9

The version of Return to Eden I ported used memory down to E00, and so this could be problematic with some disk file systems, as E00 is the file system workspace (on the Model B).

Is the issue showing up on both the Model B and the Master?

Does *DTRAP ON help?

Dave

User avatar
leenew
Posts: 4016
Joined: Wed Jul 04, 2012 3:27 pm
Location: Doncaster, Yorkshire
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by leenew » Thu Aug 25, 2016 8:56 pm

OK,
Hi Dave,
DTRAP *does* help, (see below) but we are trying to avoid it wherever possible, and only use it as a last resort.

with NO DTRAP, the game hangs before the game loads in.
My beeb is a BBC Model B with an Internal datacentre MK2 with RAMfs v1.04
It has an 8271 DFS 1.20.
Also fitted is an IFEL ROM/RAM card in which I have erased all ROM images for the purposes of this test.
My Master does not have a datacentre connected to it unfortunately.

If I create an 80 track disc from the image, it works perfectly, and I can load and save to the different slots.

If I run from datacentre with DTRAP ON, the game *does* load. However, the save-game is faulty. It appears to work, but does not RESTORE you to the correct position, but merely leaves you where you were before you typed RESTORE.

I believe BillCarr has different results on his beeb/datacentre???
Thanks,

Lee.

User avatar
hoglet
Posts: 8825
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by hoglet » Thu Aug 25, 2016 9:36 pm

The LOADER is *LOADING A and B, then running the following code:

Code: Select all

09A2  A9 8C       LDA  #&8C       ).
09A4  A2 0C       LDX  #&0C       ".
09A6  20 F4 FF    JSR  osbyte       t.
09A9  A0 00       LDY  #&00        .
09AB  B9 00 05    LDA  &0500,Y   9..
09AE  99 00 0E    STA  &0E00,Y   ...
09B1  B9 00 06    LDA  &0600,Y   9..
09B4  99 00 0F    STA  &0F00,Y   ...
09B7  B9 00 07    LDA  &0700,Y   9..
09BA  99 00 10    STA  &1000,Y   ...
09BD  88          DEY              .
09BE  D0 EB       BNE  &09AB      Pk
09C0  A9 40       LDA  #&40       )@
09C2  8D 00 0D    STA  &0D00     ...
09C5  A0 80       LDY  #&80        .
09C7  B9 00 04    LDA  &0400,Y   9..
09CA  99 00 0D    STA  &0D00,Y   ...
09CD  B9 80 04    LDA  &0480,Y   9..
09D0  99 80 0D    STA  &0D80,Y   ...
09D3  88          DEY              .
09D4  10 F1       BPL  &09C7      .q
09D6  4C 00 71    JMP  &7100     L.q
OSBYTE 8C is effectively *TAPE 12.

This is overwriting the file system workspace (E00-10FF) and NMI (D00-DFF) with it's own data.

What's not clear to me is how SAVE and RESTORE are subsequently meant to work to disk.

This doesn't look too hopeful to me.

Dave

User avatar
leenew
Posts: 4016
Joined: Wed Jul 04, 2012 3:27 pm
Location: Doncaster, Yorkshire
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by leenew » Thu Aug 25, 2016 9:56 pm

Hmmm...
Well, bearing in mind, this is from an original disc version of the game that *does* save/restore to disc, AND this image *does* work when put back to disc, this is quite strange.

I really need BillC to reply here.
You must also bear in mind that this is I think attempt 5 by BillC to get it working on my machine, so the loader may have a couple of "bodges" as it stands.
Let me see if I can find a snippet of a mail conversation between BillC and myself that may explain it a little better...

Lee.

User avatar
leenew
Posts: 4016
Joined: Wed Jul 04, 2012 3:27 pm
Location: Doncaster, Yorkshire
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by leenew » Thu Aug 25, 2016 10:02 pm

** This is from a conversation where LOADER was at &C00 (it is presently at &900)

"Here's the general breakdown of LOADER, which loads at &C00

C1A The OSWORD &7F call to obtain the disk information (used for knowing which drive it's on and whether it's Datacentre or not)
C23 OSCLI"LO.A"
C2A Manipulate result from previous OSWORD &7F call at C1A
C3F Displays "DC" & Datacentre code
C61 Figure out which drive
C75 Displays "D?" for drive"

and another:
SAVE and RESTORE with the Datacentre uses an OSWORD &77 control block, just the same as OSWORD &7F, and it only intercepts &4B for WRITE (save) and &53 for READ (restore) - that's part of the Datacentre code. Maybe different datacentres utilise different OSWORDs?
Is the same error occuring if *RUN !BOOT is issued from the > prompt?
Or if *RUN LOADER is done independently?
Finally, if "File not Found" was being shown after DC and D0 had been displayed, could you ?&980=&60 (or ?&C80=&60 on the old loader) to stop before the final file is loaded with "LO.B" and *CAT to see if the correct drive is still being selected and/or the DIR isn't messed up somehow!

and another later snippet:
"I kept the code at &C00 - &CFF because it's put there by the loader program when loading from the original disk using OSWORD &7F.
Code is then loaded at 400 - 7FF (file A) and 1100 - 7BFF (file B), filling all available memory (except the MODE7 screen)
The code from 400 - 7FF is then relocated to D00 - 10FF. (Game code then occupies D00 - 7BFF. The save game is 400 - 7FF)
This leaves 800 - CFF untouched after relocation. Since C00 - CFF isn't working correctly on your DFS and/or datacentre, maybe the loader would be comfortable at &900... i've attached a disk image with this change. If this fixes the problem, it could also be possible to rewrite LOADER into A00 and have "Datacentre" instead of just "DC" and "Drive :?" instead of "D?" using the extra space!
It's possible that the crash is because of how the !BOOT is being run... Elixir has a similar crash on Watford DFSs... does LOADER display the same symptoms if *RUN !BOOT is issued instead of SHIFT-BREAK?"


** */!BOOT does not help in my case either!


Lee.

User avatar
hoglet
Posts: 8825
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by hoglet » Fri Aug 26, 2016 6:21 am

leenew wrote: The code from 400 - 7FF is then relocated to D00 - 10FF. (Game code then occupies D00 - 7BFF. The save game is 400 - 7FF)
One possible issue is the loader overwrites DF0-DFF, which are the ROM private workspace bytes, with the following data:
000DF0 : 4C 80 0D 4C 5E 0D 86 71 5C B3 61 2C AA 8A 78 82

These values will likely be machine specific, and will depend on the order of the ROMs.

For debugging it would help to know what ROMs are in which slots in the working machines.

But it just seems wrong to do this.

I'm also surprised that SAVING/LOADING is via sector level OSWORD calls.

Do you have a copy of the original disk image, before any changes were made.

Dave

User avatar
lurkio
Posts: 2351
Joined: Tue Apr 09, 2013 11:30 pm
Location: Doomawangara
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by lurkio » Fri Aug 26, 2016 6:50 am

Hi, Dave. I've been part of the discussion with Lee and BillC.
hoglet wrote:I'm also surprised that SAVING/LOADING is via sector level OSWORD calls.
I belive the original game actually did this. According to BillC:
billcarr2005 wrote:Due to the nature of how the save games are stored, the S.? [savegame] files can't be moved from their positions on track 1-10!
BillC was referring specifically to his first attempt at converting the FSD image to SSD, which I'm attaching to this post:
hoglet wrote:Do you have a copy of the original disk image, before any changes were made.
I believe I can supply that. See second attachment:
Last edited by lurkio on Fri Aug 26, 2016 10:19 am, edited 1 time in total.

User avatar
hoglet
Posts: 8825
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by hoglet » Fri Aug 26, 2016 8:01 am

Thanks for all the context, it's very helpful.

I think the version I ported to the Acorn Atom must have been the tape version, because as far as I remember that used OSFILE for save/restore.

I imagine the reason OSWORD 7F was used for save/restore was so that the game could use E00-1100 for game data. i.e. most of the active file system is not used.

I think the best plan (which I think you had already!) is to stick with the use of OSWORD 7F, and debug why there seem to be variations between different Data Centres.

As soon as I get setup in Cambridge, I'll try to reproduce the crash.

Debugging it will be a good excuse to get ICE-T65 running, as this is exactly the sort of issue that cannot be debugged in a current emulator.

Dave

User avatar
billcarr2005
Posts: 1429
Joined: Fri Sep 09, 2005 3:01 pm
Location: UK
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by billcarr2005 » Fri Aug 26, 2016 9:16 am

Maybe some clarification, or possibly just further information / rehashing which confuses things :lol:

LOADER checks to see whether the game is running from a floppy disk and display D? according to which drive.
If a Datacentre is detected DC is displayed first, followed by which drive.
If a Datacentre is detected, the original code is overwritten with code that uses an OSWORD &77 call.
OSWORD &7F *is not* used to save to disk on the original game, since this requires use of page &D00.
&FE81 is accessed directly to load and save.

However... with Lee's Datacentre, it's not even getting so far as to load A and overwrite the original code with the datacentre equivalent, so that's not really the issue.
If somebody with a Datacentre could (kindly) check whether the disk image above works correctly and if not, with enough details, perhaps the reason for the error can be found. It could be a fluke that it works on my setup or something different with Lee's setup that is preventing LOADER from executing as expected.

For the record, the BBC I've got the Datacentre running from has OS1.2, DFS 0.9 (higher priority than the DC) and Datacentre ROM.
hoglet wrote: Do you have a copy of the original disk image, before any changes were made.

Dave
The SSD with the (possibly) working Datacentre code contains the unmodified A and B files - they had to be renamed from the original 400 and MAIN to save a few bytes in the LOADER - the modification for the Datacentre is contained within LOADER.

User avatar
hoglet
Posts: 8825
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by hoglet » Fri Aug 26, 2016 2:53 pm

Hi Guys,

I'm set up now in Cambridge. I just tested the following config:
- BBC Model B with a 1770 Disc Controller
- RamFS 1.00
- Return to Eden LOADER at FF0900.zip

This shows DC and then D0 on boot and successfully loads the game.

But As you mentioned, restore doesn't seem to have any effect.

I'll try to do a bit more debugging of this now.

Dave

User avatar
hoglet
Posts: 8825
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by hoglet » Fri Aug 26, 2016 3:42 pm

Hi Guys,

I've just found one issue...

When doing a RESTORE the OSWORD 77 Command Block (at D97) is:
00 00 04 00 00 03 13 01 00 24 00 44 0D A5 A5 20

This breaks down to:
Drive 00
Start Address 00 04 00 00
Num Params 03
Command 13
Params 01 00 24

The command of 13 is incorrect.

DataCenter OSWORD &7F only implements 4B (write sector) and 53 (read sector).

If I change 13 to 53 it looks like restore works.

There is the same issue with SAVE, in that the command is 0B and it should be 4B.

Is this LOADER code you guys have written (and have the source for)?

Dave

User avatar
leenew
Posts: 4016
Joined: Wed Jul 04, 2012 3:27 pm
Location: Doncaster, Yorkshire
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by leenew » Fri Aug 26, 2016 3:45 pm

Ooh good work!
I too am further forward...
I have unplugged the IFEL ROM/RAM board and the DFS ROM, and the game now loads from datacentre :D
RESTORE still does not work though...
Plugging in a DFS 1.21 (My DFS image was on the IFEL) and it does not load again!
DFS fighting with datacentre?

Lee.

EDIT: I think BillC wrote LOADER?
If it isn't on the original FSD, then he did anyway!

User avatar
lurkio
Posts: 2351
Joined: Tue Apr 09, 2013 11:30 pm
Location: Doomawangara
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by lurkio » Fri Aug 26, 2016 4:26 pm

I have access to a DataCentre now. It's an external one, and I have a Model B and a Master 128 I can connect it to.

I won't muddy the waters by posting my own findings in detail right now, but could you all please try booting the RtE-900-loader SSD from a DC, play around in the game a bit, and then do Ctrl-Break and try to re-boot the game from the same DC RAM-drive without re-importing the image?

It won't re-boot for me. I get an error. ("Bad string", I think it was.) That was on the M128 with 1770 DFS, RamFS 1.00 and DC.

:?

(On the Model B, I couldn't get the game to boot from DC at all -- it would just hang on the title screen without printing DC or D0 or anything, like Lee's earlier experience -- until I used the ADT ROM and *UNPLUGged the DFS 1.2. Then I could *RUN !BOOT from the DC and play around in the game. And I could even re-run the game from the same RAM-drive without reimporting the image -- as long as I made sure to *UNPLUG the DFS1.2 first and then manually *RUN !BOOT. (I've got RamFS 1.00 in SWR socket &C, and DFS 1.2 in SWR socket &E.))

User avatar
billcarr2005
Posts: 1429
Joined: Fri Sep 09, 2005 3:01 pm
Location: UK
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by billcarr2005 » Sat Aug 27, 2016 1:29 pm

hoglet wrote:
...
The command of 13 is incorrect.

DataCenter OSWORD &7F only implements 4B (write sector) and 53 (read sector).

If I change 13 to 53 it looks like restore works.

There is the same issue with SAVE, in that the command is 0B and it should be 4B.
Because the usual loader is directly accessing the hardware with FE81, the command sent to the controller isn't always &53, it needs to change according to the drive being accessed. I presume when any OSWORD&7F block is passed to the disk controller, that the value is changed accordingly. The value of D9D, for the Datacentre version, is changed with

Code: Select all

0D80 LDA &A2
0D82 STA &D9D
This allows the same code block to be used for both SAVE and RESTORE.

A2 is set at

Code: Select all

7AE1 LDA #&4B for SAVE
7AEA LDA #&53 for RESTORE 
The initial values of these locations *are* &13 and &0B because they should then have &40 added to them during the loading, which works for both the low level (8271 disk) and OSWORD &77 (datacentre) versions.

However, it appears that some DFSs don't respond correctly when special register &23 is read during the loader, which would allow &40 to be added to the initial values.
Also, &4B and &53 would work all the time, presumably, for the Datacentre version.

If you could

Code: Select all

*LOAD LOADER
?&923=&60
CALL&91A
PRINT ~?&918
then it might shed some light on the matter...

User avatar
KarateEd
Posts: 4416
Joined: Fri Sep 20, 2013 9:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by KarateEd » Sun Aug 28, 2016 9:51 am

Hi Folks,

On my 32k Issue 4 Beeb With DataCentre installed:

I also have an ATPL Sideways RAM/ROM board installed with:

ROM 00 LOGO 2
ROM 01 LOGO 1
ROM 02 DFS 2.26
ROM 03 RamFS 1.00
ROM 08 EXMON II 2.02
ROM 09 HELP ("H) 1.2
ROM 10 The BASIC Editor 1.32
ROM 11 Disc Doctor 1.10
ROM 12 ARM 1.13
ROM 13 ADT 2.00
ROM 14 BASIC

Using the "KILL command, I disabled all but the 2 ROMS shown below.

With just DC RAMFS and BASIC ROMS active, DC and D0 are what I see. The game runs with "RUN !BOOT or "RUN LOADER equally well. The SAVE/RESTORE also does not work, as previously described.

If I leave the DFS ROM active with RamFS and BASIC, I get D3 and the game runs normally until I try to SAVE. At that point the game freezes up.

Ed......:-)
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

User avatar
billcarr2005
Posts: 1429
Joined: Fri Sep 09, 2005 3:01 pm
Location: UK
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by billcarr2005 » Sun Aug 28, 2016 11:43 am

Could you also try the above code which displays the result of reading the special register?
Results under emulation with Watford DFS 1.30

Code: Select all

Drive  Result
0      &48
1      &88
2      &68
3      &A8
Since the reading of the special register requires a DFS ROM, I'm supposing that LOADER requires a different branch if the result isn't as expected.

Also, it only needs &53 and &4B sent to the OSWORD &77 block, rather than the current situation of altering according to drive - which is required for the disk version.
KarateEd wrote: With just DC RAMFS and BASIC ROMS active, DC and D0 are what I see. The game runs with "RUN !BOOT or "RUN LOADER equally well. The SAVE/RESTORE also does not work, as previously described.

If I leave the DFS ROM active with RamFS and BASIC, I get D3 and the game runs normally until I try to SAVE. At that point the game freezes up.

Ed......:-)
Unless the reading of the special register is returning &4? then the SAVE/RESTORE will not function correctly - it's easily fixed to always be &40.
If no DC is displayed, then the program assumes that an 8271 DFS is being used, and alters the code accordingly, from drive 0 to drive 3, so unless you had a disk in drive 3, the game can't save to it.

I've just tried running under BBC Master 128 emulation and the result is &FE, which is translated to drive 3... but the original program was never designed to write to the disk using a 1770 DFS

Another alternative would just be to have two separate A files, to load depending on which system is being used... but the Datacentre would still need to be accurately identified, or just have two separate disk images, then no checking is needed!

User avatar
KarateEd
Posts: 4416
Joined: Fri Sep 20, 2013 9:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by KarateEd » Sun Aug 28, 2016 9:28 pm

billcarr2005 wrote:Could you also try the above code which displays the result of reading the special register?
Results under emulation with Watford DFS 1.30

Code: Select all

Drive  Result
0      &48
1      &88
2      &68
3      &A8
I get FE when I PRINT PEEK &918. Hope that helps.

I don't really have a physical Drive 3 but I'm going to do an experiment and put a diskette in my physical drive and see if anything happens.

Coming soon the answer to that.

Edit: No difference. It seems that if DFS 2.26 is active then it just isn't going to save anywhere. My physical drive doesn't have a side 3 anyway, probably why and the fact it's a 1770 chip......

Ed......:-)
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

User avatar
lurkio
Posts: 2351
Joined: Tue Apr 09, 2013 11:30 pm
Location: Doomawangara
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by lurkio » Mon Aug 29, 2016 3:17 pm

billcarr2005 wrote: If you could

Code: Select all

*LOAD LOADER
?&923=&60
CALL&91A
PRINT ~?&918
then it might shed some light on the matter...
On my Model B with RamFS 1.00 in SWR socket &C, and DFS 1.2 ("DFS,NET") in SWR socket &E, if I try to follow your instructions, then the machine hangs after I enter CALL&91A, so I can't get any further than that.

But if I first use ADT to *UNPLUG the DNFS ROM and then follow your instructions, the CALL returns, but the value of the contents of &918 is zero.

This is obviously all after initially importing the RtE image (with loader at &900) to the DataCentre, drive 0.

UPDATE

If I load the DFS 0.90 ROM into SWR in slot &D, and then use the ARM ROM to *KILL the DNFS in slot &E, and then follow your instructions, I get a value of &40. Also, the game seems to correctly boot, run, SAVE and RESTORE on DC!

Looks like DFS 1.20 (DNFS) is the culprit?

User avatar
KarateEd
Posts: 4416
Joined: Fri Sep 20, 2013 9:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by KarateEd » Mon Aug 29, 2016 7:26 pm

I'll give that a try tonight Lurkio.

Edit: Well, managed to burn the ROM incorrectly, am now wiping it and will try again perhaps tomorrow.

Ed......:-)
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

User avatar
KarateEd
Posts: 4416
Joined: Fri Sep 20, 2013 9:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by KarateEd » Wed Aug 31, 2016 1:35 am

Well, when burning the DFS 0.9 ROM, I'm burning it to 4000 - 7FFF. When I plug it in, it's not seen by my Beeb.

Any ideas, have I got the right address to burn it? The burn looks good btw.
DFS 0.9 start.png
DFS 0.9 end.png
Ed......:-)
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

User avatar
lurkio
Posts: 2351
Joined: Tue Apr 09, 2013 11:30 pm
Location: Doomawangara
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by lurkio » Wed Aug 31, 2016 7:17 am

KarateEd wrote:Well, when burning the DFS 0.9 ROM, I'm burning it to 4000 - 7FFF. When I plug it in, it's not seen by my Beeb. Any ideas, have I got the right address to burn it? The burn looks good btw.
With the caveat that I know nothing about burning ROMs, having never done it myself, can I ask if the &4000 address you mention is the offset within the ROM? If so, then why do you need an offset at all? Shouldn't you be burning at &0000 -- i.e. at the "beginning" of the ROM (or whatever the right term is)?

(In the Beeb's memory map, paged ROMs occupy the address space from &8000 upwards.)

User avatar
KarateEd
Posts: 4416
Joined: Fri Sep 20, 2013 9:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by KarateEd » Wed Aug 31, 2016 4:02 pm

lurkio wrote:
KarateEd wrote:Well, when burning the DFS 0.9 ROM, I'm burning it to 4000 - 7FFF. When I plug it in, it's not seen by my Beeb. Any ideas, have I got the right address to burn it? The burn looks good btw.
With the caveat that I know nothing about burning ROMs, having never done it myself, can I ask if the &4000 address you mention is the offset within the ROM? If so, then why do you need an offset at all? Shouldn't you be burning at &0000 -- i.e. at the "beginning" of the ROM (or whatever the right term is)?

(In the Beeb's memory map, paged ROMs occupy the address space from &8000 upwards.)
Well, I'm fairly ignorant on ROM burning too, it's quite new to me. Having said that, this is exactly what I had to do when I burnt ROMs for ADT, ARM, ExMon, etc., about 7 ROMs.

I was advised to do it that way because you could fit 2 ROMs in the EEPROMs I'm using and the Beeb would not recognize the 1st 1/2 of the EEPROM. I did it that way and it worked but maybe there is something different with doing the DFS chips, I really don't know.

What I do know is the DFS would fit in 1/4 of the EEPROM very easily so maybe it has to be burned up higher, like around &6000.

I suppose that's what I'm really asking here.... I'm just trying to help gather data for this game with a DC installed and instead I need help, how's that for irony?

Ed......:-)
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

User avatar
lurkio
Posts: 2351
Joined: Tue Apr 09, 2013 11:30 pm
Location: Doomawangara
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by lurkio » Wed Aug 31, 2016 5:14 pm

Ah. I stand corrected. Seems like a black art, this ROM burning!

Out of interest, what equipment and what software are you using to burn the ROMs? And are they ROMs or EPROMs or EEPROMs?

Also, did you get the DFS 0.90 ROM image from mdfs.net?

User avatar
KarateEd
Posts: 4416
Joined: Fri Sep 20, 2013 9:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by KarateEd » Wed Aug 31, 2016 6:16 pm

lurkio wrote:Ah. I stand corrected. Seems like a black art, this ROM burning!

Out of interest, what equipment and what software are you using to burn the ROMs? And are they ROMs or EPROMs or EEPROMs?

Also, did you get the DFS 0.90 ROM image from mdfs.net?
Below is the chip I'm using:
EEPROM.png
Hi Lurkio,

It is a bit of a black science that I certainly don't really understand. What I do understand is you should be able to copy ROM images down to the chip and make it work, provided you've placed the image of said ROM in the correct locations.

The burner I'm using is a MiniPro and the associated software from that company.

This burner really works well.... I burned 7 or 8 chips with it without fail, they are all visible and all check out and run as advertised.

I got the DFS from the *. site under ROMs

Ed......:-)
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

User avatar
leenew
Posts: 4016
Joined: Wed Jul 04, 2012 3:27 pm
Location: Doncaster, Yorkshire
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by leenew » Wed Aug 31, 2016 6:18 pm

Blind leading the blind here.... but I *think* you have to do 2 copies if you are burning an 8K file into a 16K EPROM?
Oh why can't an expert drop by when you need them? :roll:

Lee.

User avatar
KarateEd
Posts: 4416
Joined: Fri Sep 20, 2013 9:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by KarateEd » Wed Aug 31, 2016 6:32 pm

leenew wrote:Blind leading the blind here.... but I *think* you have to do 2 copies if you are burning an 8K file into a 16K EPROM?
Oh why can't an expert drop by when you need them? :roll:

Lee.
I only had to burn one copy for the other ROMs though. Is it because the DFS ROM is so small?

Ed......:-)
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

User avatar
leenew
Posts: 4016
Joined: Wed Jul 04, 2012 3:27 pm
Location: Doncaster, Yorkshire
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by leenew » Wed Aug 31, 2016 6:44 pm

I suppose we had better start reading this: viewtopic.php?f=3&t=8688

Lee.

EDIT: after a quick scan, it is mainly about fixing a suspected broken programmer, altough there are a couple of snippets regarding "offsets"...
around here: viewtopic.php?f=3&t=8688&start=60#p95690

User avatar
KarateEd
Posts: 4416
Joined: Fri Sep 20, 2013 9:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by KarateEd » Wed Aug 31, 2016 8:05 pm

That's useful. It helps me understand how to reach certain parts of the EEPROM with programming it. I think I've successfully burnt the ROM but I won't be able to check for a few hours.

I've succeeded in putting the ROM at $4000 up which in theory should work. This ROM burn looks better than the last one I used.... seems like 'normal' ROM info is in placel.

I'll let people know the result once I test this evening.

Ed......:-)
DFS 0.9-2 start.png
DFS 0.9-2 end.png
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

User avatar
KarateEd
Posts: 4416
Joined: Fri Sep 20, 2013 9:15 pm
Location: Squamish, BC, Canada
Contact:

Re: Problems with Datacentre (I think!) with certain games

Post by KarateEd » Thu Sep 01, 2016 2:38 am

Hmmm, still no luck.... oh well, wipe it again and try another tack.

Ed......:-)
Ed...... :-)

3 working Beebs, 1 RetroClinic Master, 1 normal Master, 1 A3010, 1 Pi2 RISC OS, 2 broken Beeb Motherboards, 1 Omnibus A7000+ server, 1 A7000+ Desktop, 1 PET, 1 C64, 1 C128, 1 Amiga 500 and 1 Roamer.

Post Reply