Home Page Forums MAME MAME TMS-09xx/1xxx thread (was New Dumps)

There are 21Kx17K die shots linked off the blog (giant images), and I copied just the melody ROM area here: www.seanriddle.com/bx301acidmelody.jpg

And here's my visual dump of the melody ROM: www.seanriddle.com/bx301melody.bin

I think something's still not right. The game doesn't work at all(no changes in RAM, doesn't read K pins, doesn't write S pins). I converted bx301.bin to LFSR, then to raw, and lined up the bits to the die image. But it doesn't match. Could you upload the raw rom?

Here's the raw file I transcribed from the die shot: www.seanriddle.com/bx301raw2.bin

I converted the raw file like I did for the Konami games and got this file: www.seanriddle.com/bx3012.bin

Then I put the converted file into LFSR order and got www.seanriddle.com/bx301.bin, which exactly matched the electronic dump, after I shifted the bytes in pages 32-63 of the electronic dump.

Hmm, bx3012.bin converted to linear PC is this: http://tsk-tsk.net/net/temp/linear_boxing.bin
pages 0-31 are shifted, not 32-63?

*edit* The game boots now, but glitches everywhere. I'm sure the prg and melody roms are correct, I assume the problem is the MCU emulation. At first I was getting no LCD graphics at all, it turns out MAME struggled with linefeeds(0x0A) in your SVG element titles. Here's the fixed one: http://tsk-tsk.net/net/temp/gnw_bx301.svg

I had never linearized SM5 code before, so I got the LSFR code out of increment_pc(). I guess I did it wrong, but I'm confused that my mistake resulted in shifting the wrong half of the file....

I'm not sure what went wrong with the SVG. I used Inkscape under Windows, and didn't hand-edit the file. I did paste in the titles from a spreadsheet that I had made with the segment info, so maybe that added in extra linefeeds.

With SM5* in dump/testmode, is it possible to send sequential opcodes? There are a few opcodes that I'm not sure about, like SBM and LB. If possible, these opcodes can be reverse engineered.

Sure; if my timing works correctly, I can send any opcode. But we'd have to figure out how to detect the results. The S outputs are tied up in test mode, but I guess we could conditionally set/clear segments as indicators.

I think you probably saw Paul Robson's blog: http://watchdev.blogspot.com/ The posts have some of the ideas that he got from the datasheets and patent 4,504,062. The links are dead, but I had downloaded the files if you think they would be useful.

I posted the visual ROM dump for Donkey Kong II JR55: www.seanriddle.com/gameandwatch.html I don't have the LCDs for that one.

Yeah, I've downloaded Paul Robson's SM510 related files too a while ago. His emulator looks unfinished.

I double checked JR55, looks correct
What is the die label for this one?

The package labeling is

JR-55
53YC
Nintendo

and the die label is CMS54C / KMS560. There's also another 560 on the die.

I'm pretty sure the "brochure" mentioned in patent 4,504,062 is part of the patent wrapper ($55), but it doesn't look like it has any more info than the 1990 data book. I did some other patent searches, but didn't find much else. A Google Patents search for "Sharp SM" found SM4, SM530 and SM551, and "Sharp microcomputer" found a suntan indicator watch with an SM530.

Does anyone know what game is in patent US4895376A?