Panasonic FS-A1F and 128kb RAM upgrade

Página 3/8
1 | 2 | | 4 | 5 | 6 | 7 | 8

Por ro

Scribe (4867)

imagem de ro

02-09-2022, 08:36

And using openmax with that panasonic machine? What does that do?

Por sdsnatcher73

Prophet (3810)

imagem de sdsnatcher73

02-09-2022, 18:12

Emulation is incomplete for A1F. There is something slightly different in the A1F compared to other MSX2 machines with mapper. On an unmodified A1F (which has a mapper of 64kB) the mentioned software (e.g. PA3) detects a larger mapper (probably 512kB but it is not reported). Other software is able to detect the correct size (e.g. RAMTEST and TESTMAP).

On openMSX the malfunctioning software also detects the correct size (so 64kB).

I have reported it to openMSX team but have not heard back from them.

Por ro

Scribe (4867)

imagem de ro

02-09-2022, 21:07

Well, perhaps I could meet up with retroFan. He lives close by. Do some investigation on the matter. Since all mentioned games and music disks run on our o.s. I might be able to pin point the issue

Por Manuel

Ascended (19227)

imagem de Manuel

02-09-2022, 21:14

There is not much we can do without further investigation. Please keep us updated ro!

Por sdsnatcher73

Prophet (3810)

imagem de sdsnatcher73

02-09-2022, 21:52

I think it should not be too hard to at least figure out where the code gives back an incorrect value. On RetroFan’s machine with 128kB the code that checks the size of the mapper returns a value higher than 128kB. Same on a machine with 64kB.

Finding what causes this in the hardware design might be difficult, then again we have people in the community that could assist. If they are not close to someone with an affected machine I can provide one for purpose of the analysis.

As I mentioned there may be other machines that have this issue. I’ve seen the TESTMAP issue under DOS1 also on National FS5000 and Canon V30F. I will try PA3 on these machines as well.

For openMSX I guess the exact hardware issue (which may still be within ASCII spec) is not that relevant just the circumstances the code fails under should be enough to implement the same outcome.

Por itozzz

Resident (59)

imagem de itozzz

04-09-2022, 01:02

Hello, a few days ago I received an fs-a1f from Japan, I was in the process of restoring it and I took the opportunity to put some sockets and expand the ram to 128k but before doing so I check with the msx diagnostics program it tells me I have 512k, can someone tell me why?
20220903_230314.jpg
20220903_232650.jpg

Por sdsnatcher73

Prophet (3810)

imagem de sdsnatcher73

04-09-2022, 09:01

It is exactly what we are trying to find out in this thread. The issue is that some detections (apparently also MSX Diagnostics) identify incorrectly that the machine has 512kB. Other software is able to detect correctly the 64kB or 128kB (depending on original machine or upgraded machine).

Por Manuel

Ascended (19227)

imagem de Manuel

04-09-2022, 11:54

So, can someone please figure out which detection method is used in these programs?

Por sd_snatcher

Prophet (3609)

imagem de sd_snatcher

04-09-2022, 15:01

Manuel wrote:

So, can someone please figure out which detection method is used in these programs?

The programs that are detecting 512KB of RAM are doing the quick and dirt detection method that is prone to failure. They're just checking the amount of bits active in the Memory Mapper registers (5bits, in this case).

Por sd_snatcher

Prophet (3609)

imagem de sd_snatcher

04-09-2022, 15:22

Wierzbowsky wrote:

I ran into a weird problem after upgrading the Panasonic A1F board to 128kb. TestRam and MemTest show no issues, but TestMap shows this error:

Any thoughts?

This error is caused by a mismatch between the number of active output bits (*1) in the memory mapper registers and the actual number of existing pages. People who are building their omegaMSX are getting similar errors when they mess up with the memory configuration jumpers.

TestRAM is very strict on its testing. Too sad it's not very clear when it comes to explain Memory Mapper register size mismatch. I wonder if the source code is available somewhere.

Under normal conditions, this mismatch should cause no problem. But a difference between the number of active mapper register bits and the number of existing RAM pages is known to cause poorly written software to fail.

It's similar to the problem that some software have with split-RAM configuration.

For developers wanting to avoid this bug: The MSX Acid3Test covers this situation.

Página 3/8
1 | 2 | | 4 | 5 | 6 | 7 | 8