Panasonic FS-A1F and 128kb RAM upgrade

Page 4/6
1 | 2 | 3 | | 5 | 6

By sd_snatcher

Prophet (3549)

sd_snatcher's picture

04-09-2022, 15:31

sdsnatcher73 wrote:

It is writing FF to address 8000 and reads back 00.

This happens because Testram tests all pages that are supposed to be present. Once it reaches 128KB+1, it writes 00h and there's no RAM chip present to read the value back. The FFh is then read from the databus pull-up resistors.

It's is correctly reporting what it found. But IMHO it should give a more detailed explanation for the problem, otherwise users without more intimate knowledge of the architecture will struggle to understand what is going on.

By sdsnatcher73

Prophet (3516)

sdsnatcher73's picture

04-09-2022, 17:01

But testmap correctly identifies the mapper as 128kB. So why would it swith to the page 128kB+1?

Also openMSX emulates the 5bit register and the mentioned software works well even with this but not on real hardware. I am afraid the case of the A1F is more complicated.

By Retrofan

Paragon (1315)

Retrofan's picture

07-09-2022, 19:39

By the way, you also can test the 128 kB RAM upgrade with this BASIC listing from here See below.

On my machine it detects 128 kB without problem in DOS1 mode. Maybe this listing is useful to check how to detect the mapper in the correct way. First execute line 10 and then reload the BASIC listing and run it again (line 10 is now skipped).
If you don't do this you will end up in a Syntax error.

10 'poke &hf676,1:poke &hf677,&hc0:poke &hc000,0:new
20 DEFINT A-Z
30 PTR=&H8100:S=4
40 FOR I=4 TO 255
50 OUT &HFE,I
60 PTR=&H8100:FOR J=0 TO 127:POKE PTR,&H55:POKE PTR+1,&HAA:PTR=PTR+2:NEXT J
70 EF=0:PTR=&H8100:FOR J=0 TO 127
80 IF PEEK(PTR)<>&H55 OR PEEK(PTR+1)<>&HAA THEN PRINT "NG:";I:J=512:I=256:EF=1:GOTO 100
90 PTR=PTR+2
100 NEXT J
110 IF EF=0 THEN S=S+1:PRINT "OK:";I;S*16384/1024;"KB"
120 NEXT I
130 OUT &HFE,1
140 END

By Retrofan

Paragon (1315)

Retrofan's picture

07-09-2022, 19:56

Interesting: OpenMSX also detects 128kB, but the program doesn't end with Ok like on the real machine. It looks like it does crash never ending the program.

By Retrofan

Paragon (1315)

Retrofan's picture

11-09-2022, 14:22

MSXMEM reports 512kB RAM instead of 128kB RAM:
https://m.blog.naver.com/zeta6262/80154575265

By Manuel

Ascended (18938)

Manuel's picture

11-09-2022, 15:02

sd_snatcher wrote:
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).

Well, but that part is emulated in openMSX, and still it works correctly...

By Retrofan

Paragon (1315)

Retrofan's picture

12-09-2022, 17:48

sdsnatcher73 wrote:

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.

Did you try PA3 on these machines as well sdsnatcher73?

By sdsnatcher73

Prophet (3516)

sdsnatcher73's picture

12-09-2022, 20:45

Not yet, not all machines are fully repaired yet Wink especially those I will keep for myself have been on the backlog..

By Wierzbowsky

Guardian (3489)

Wierzbowsky's picture

22-09-2022, 23:33

Funny thing is that if you use Carnivore2 in slot 1 and run TESTMAP, the error will be shown only at start and the 128kb mapper tests will be working fine from then on. I somehow feel that the modding instructions for A1F are not complete and something else needs to be bridged or cut on the motherboard. But my modest skills in electronics don't allow me to figure this out...

By Retrofan

Paragon (1315)

Retrofan's picture

23-09-2022, 03:04

I have the same feeling from the beginning, but all found instructions on the internet are the same including Japanese websites. And sdsnatcher73 did report the same happens with the unmodified 64kB A1F.
However, how can you explain that the BASIC listing above does detect the 128kB correctly? Taken from a Japanese website as well. I'll meet ro today. Just try to find out.

It would be helpful if the other mentioned machines by sdsnatcher73 can be tested as well by himself or someone else with the same config.

Page 4/6
1 | 2 | 3 | | 5 | 6