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

The tms1xxx rewrite is uploaded finally, see:
https://github.com/mamedev/mame/blob/master/src/emu/cpu/tms0980/tms0980.c

And for the TI calculator driver, see:
https://github.com/mamedev/mame/blob/master/src/mess/drivers/ticalc1x.c
(ah I see I still have to do the default button mapping)

Thanks! I've been busy with work, but I got Amaze-A-Tron decapped and photographed. I'll composite the pictures and dump the ROM and post on my web page. I also bought Maniac, which is another Ralph Baer TMS-1000 game. And I picked up a couple of Speak and Spells. They should get here in a few days.

Ok, we finally have a decap of the TMC027x/TMS027x/CD270x die:
http://siliconpr0n.org/map/ti/cd2708n2l/top_metal_mit20x/

This is the die from the Speak and Math (French/Canadian version), CD2708N2L.

Note that the Speak and Math (USA) version uses an MCU labeled CD2704 so there probably is some as-of-yet unknown code change between this die and the USA one.


Comparing it to Sean Riddle's decap of the TMC0260 from the TI 5025 (see http://www.seanriddle.com/ti5025fulldie.jpg ), this die is definitely based on the tmc0260, but with more ROM, more RAM, and everything moved a bit farther apart from each other to make space for the extra routing lines etc. The segment PLA is larger with 48 terms instead of 32.

It uses 9-bit ROM words like the TMS0980.

The Segment PLA of the CD2708 probably does not match the speak and spell Segment PLA as it should have a few more math related characters in it.

The pinout of the CD2708 should match or nearly match the TMS0270 pinout, there may be a minor difference to allow driving a 9th digit of the VFD.

NOTE: the pinout chart of the TMS0270 I made was created before this decap was done and is likely to have errors and omissions! Do not hesitate to make corrections and/or stick the diagram in the MESS source code somewhere.


There is a patent on the Speak and Math which includes a code listing which should match (probably) the CD2704 USA version, US Patent 4,970,659 http://www.google.com/patents/US4970659. The source code is in the text area as a TABLE so it doesn't show up in images in Google patents, so best is to use pat2pdf.org instead http://www.pat2pdf.org/patents/pat4970659.pdf

As far as I know, the Speak and Math code has never been typed up like the Speak and Spell patent code has been.


There are two different sets of two VSM(TMS6100 or TMS6125) speech ROMs for the Speak and Math:

CD2392 and CD2393 are two 16KiB VSM ROMs used on the apparently older speak and math units.

CD2381 and CD2614 are a 16KiB and a 4KiB VSM ROM respectively used on the apparently newer speak and math units, despite the first VSM being a lower number than the other two from the 'apparently older' units. (There's probably a story here...)


LN

Nice! That decap should be useful (sean: hopefully you didn't purchase this specific revision =p)
I see they changed the way the TDO opcode works again. And it looks like RSTR opcode is back too. (I can't see RSTR on the instruction decode tho)

*edit*
0260 and 0270 have same amount of ROM and RAM as 0980.

...That's the wrong patent (it contains the speak and spell code), due to a now-corrected error in my notes.

The correct patent for Speak and Math is US Patent 4,946,391 http://www.google.com/patents/US4946391 http://www.pat2pdf.org/patents/pat4946391.pdf

The code starts at pdf page 23 and there's a lot of headers/comments/address map info which even the speak and spell source code lacks, which should be quite helpful.

A k-line to keyboard map/schematic is in there too (in ascii-art), as well as a symbol lookup for the Segment PLA.

LN

sean: can you get scans of the playfield, the control panel and the instructions for codename: sector?

There was a remake which had some nice (but not accurate except maybe the compass rose and some of the button art) artwork done up for it by CaptGlub AKA Captain Crybaby which used to be hosted at http://web.archive.org/web/20131208005953/http://stillmixtup.com/CNS.html which had some scans of the manual, playfield and the playfield reconstruction.

Unfortunately, none of those docs nor the executable seem to be up on archive.org and the site is down. I may see if I can ask CaptGlub if he can rehost the data if he still has it.

EDIT: Hasbro has a copy of the manual up, though it is a black and white scan, while I think the original was color:
http://www.hasbro.com/common/instruct/codenamesector.pdf
We still need a good scan of the playfield, preferably in color, and of the control panel (which due to its angle is probably best photographed rather than scanned)

LN

P.S. Do you have a copy of the manual and grid pad for Star Wars: Electronic Battle Command?

P.P.S: Are all the bits right in the Codename: sector rom dump? The game starts up after reset (and in teach mode) in MESS with Ship 1 at 35,24, which is wrong; the ship should start at 35,25 in both cases.

Sean already double-checked codename sector dump. If you want to check too, this is my method: compare the bits with the rom array acid pic. (for 0970, takes about an hour for me)

http://pastebin.com/mYXcFCTa

*edit* I will doublecheck it later myself, and stopthief too http://pastebin.com/69CGSKC3 the more eyes to doublecheck the better..

cnsector(raw) change offset 3e6: 0e to 0f

Next stop is stop thief :|

Sorry, I've been busy, and I've probably got another week of it.

Hap- thanks for the CN:S bit.

LN- I've got a used CN:S from ebay that is in OK shape. I'll scan the parts and manual. The SW:EBC is pretty rough (I go for the cheapest ones I see since I'm just going to ruin them), but I can scan it until something better comes along. No manual, though.

I got one of the S&S units- non-working, but good case. 4 chips on the MB and a "Vowel Power Word Module" plugged into the expansion port. 40 pin DIP TMC0271NCL NBS 7910, 28 pin DIP TMC0351NL NBU 7906, 28 pin DIP TMC0352NL DAU 7901, 28 pin SDIP TMC0281NL D 7906. Vowel Power has 28 pin DIP TMC0350NL CD2302 DAU 7914. I haven't desoldered them yet to see what's on the bottom.

No problem. In the meantime, I got all dumped games to working state in MESS. Except for the stopthief patent listing.
Speaking of stopthief, I'm pretty certain there's a bit under the bigger smudge (cf instead of 8f).