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

Yeah, the change to SETR and removal of RSTR may have something to do with it. I still haven't read enough to understand that change yet. I don't have any working TMS-0970s left, but I have a couple of TMS-0980s, which likely output the same. I could do some logic captures if you think it might help.

I uploaded the TI-1270 post-acid pic.

I looked at the Stop Thief output PLA and found 0-9, C, d, F, G, t and blank. But the game also displays - and r. Do you know how? The TMS1x00 output PLA used the 4-bit accumulator and the status flag to index into a 20-byte table of outputs, but I only see 16 outputs on the Stop Thief die.

It doesn't work the same way as TMS1000. Instead of outputting all segments at once, it updates digits per-segment. This way, it's possible to make a digit partially blank.

It's not that simple. There are 2 PLAs involved, the segment PLA and output PLA. The chip checks for a coincidence between the two to determine whether or not a digit should be updated.

For details, see TI-30 patent description of fig.13

OK, so you could tell it to display 8, but only segments e and g, and that would display r?

Stop Thief's segment decoder has different patterns for several digits- 6 and 9 don't have the "tail" segment (a for 6 and d for 9), and 1 is scooted to the left, using segments e and f instead of b and c. It also has a C (adef), d (abcdg), F (aefg), G (acdef), and t (defg). The game also displays - and r which aren't in the segment decoder.

I uploaded ROM dumps for TI Programmer and Business Analyst I, and updated the pinout files and output PLA info. I need to clean the output PLA of TI Programmer to transcribe it, although I think we know enough to reconstruct it if need be. Along with 0-9, it displays A b C d E and F.

"OK, so you could tell it to display 8, but only segments e and g, and that would display r?"
Yes I believe so. Though I don't have anything with TMS0980 emulated yet to confirm it (I only have TI-30 hooked up, but it gets stuck in a loop very quickly).

BTW I forgot to ask before: could you also post your script (or description with pseudo code) to convert a raw dump to tms1000-order dump?

Here's the pseudocode. It's just a few loops around the code I posted before:

Ah, right.
Anyway, buttons (1/x), (x^2), and 3,6,9 don't work right on my ti1270 emulation. I checked the bits on the die pic under the smudge for faults. I found 1 bad bit! But seems to have no effect at all on the emulation.

ti1270raw.bin offset $174, change $11 to $51.

Status update of tms0980.c rewrite:
- microvsn, merlin, simon: work as good as before
- tisr16: boots, can input numbers, but lots of bugs and can't do calculations. If it's an emulation problem it's strange, since others with tms1xxx work well. I see black lines on the die pic, is it possible your rom extract algorithm gets confused there?

- wizatron: works
- comp4: works
- cnsector: at first glance, it works. Still need to flesh out the MESS driver.
- ti1270: boots, can input numbers and some calculations, but some buttons don't work (see above, it's as if they function as RCL and STO)

- ti30, tiprog, tibusan1, stopthie(also the patent one): execute nonsense code and get stuck in a loop quick. Either CPU emu problem or the word/bit order is different.

D'oh- sorry about that bit. I just checked the full die again and I didn't notice any other errors.

That's a weird collection of buttons not to work- I'd expect a full row or column, not a mix of buttons on 3 different columns and 4 different rows.

Cool!
The SR16 die got stuck on the metal frame and cracked, which are the lines you see. I thought everything was readable, but I'll double-check.
Maybe the Stop Thief patent code is in LFSR order, not execution order. I reordered the raw 0980 bits to match it, so that would explain why they all wouldn't work.