|
Joined: May 2010
Posts: 1,051
Very Senior Member
|
Very Senior Member
Joined: May 2010
Posts: 1,051 |
D'oh- I thought the PPS-4/1 opcodes would be the same as the other PPS-4 chips. It looks like we need the PPS-4/1 Programmng Manual #29410N38 and/or the PPS-4/1 Operator's Manual for Universal Assemulator #29400N37 ([homer]MMMM assemulator[/homer]).
A quick patent search came up with source and object code for a telephone that uses the MM77: 4203011. Maybe the opcodes are the same as for the MM78, or maybe there's a patent with MM78 code.
|
|
|
|
Joined: May 2010
Posts: 1,051
Very Senior Member
|
Very Senior Member
Joined: May 2010
Posts: 1,051 |
Reset inits the L port to input, and makes SK output the CPU clock. The Entex Space Invader game sets the L port and SK to outputs, so when I let it run and then went into ROM dump mode (SO=3V, SI=gnd), it started dumping the ROM contents at whatever address the program counter had.
If the game didn't set the L port to all outputs, I would have had to do the forced instruction setup to do that. Because the game uses it as an output, I did lose the SK sync, but that was no big deal; I just looked for edges and sampled in the middle.
When I went from reset directly to ROM dump mode, I got what you got- a clock on SK, but nothing on L.
Since I was able to figure out where the ROM dump started, I didn't try to force instructions.
If your chip should be setting L to output, try tweaking the voltage on SO a bit. I had to adjust it depending on VCC I used.
|
|
|
|
Joined: Mar 2002
Posts: 1,370 Likes: 132
Very Senior Member
|
Very Senior Member
Joined: Mar 2002
Posts: 1,370 Likes: 132 |
The Speak and Read I got looks much more professional inside than the old Speak and Spells I have. The chips are CD2705B-N2L MES 8149, CD2801A-N2 ME 8144, CD2394A-NL MBU 8144 and CD2395A-NL MBU 8149. The CD28 and CD23s are 28-pin SDIP, and the CD27 is a cute 40-pin DIP that's the length of a 28-pin DIP.
I've got a backlog of chips that'll have to wait until work slows down a bit: Entex Baseball 3, Entex Space Invader with the TI chip, Kosmos Astro, and Light Fight should be here this week. CD2705 is the MCU kevtris dumped the CD23xx already (also rev.A, same as yours). CD2801A I don't know if anything's needed with that, lemme summon Lord Nightmare... Speak & Spell Compact, Touch & Tell, and TI Vocaid are assumed to be dumpable TMS1100.
|
|
|
|
Joined: Mar 2006
Posts: 1,080 Likes: 7
Very Senior Member
|
Very Senior Member
Joined: Mar 2006
Posts: 1,080 Likes: 7 |
CD2801A-N2 ME 8144 is the speech chip (variant of tms5100). I've never tested one, but plgdavid has, let me see... EDIT: see post below
LN
Last edited by Lord Nightmare; 03/04/15 11:33 PM.
"When life gives you zombies... *CHA-CHIK!* ...you make zombie-ade!"
|
|
|
|
Joined: Aug 2009
Posts: 192
Senior Member
|
Senior Member
Joined: Aug 2009
Posts: 192 |
//Matches "TMC0280NLP CD02801 W 8048 PHILLIPINES" (from a Language Translator) //Matches "CD02801A-NL MFU 8231 PHILLIPINES" (from a Speak & Math 80)
These share common ROM tables, different than TMS5100(A and not)
Last edited by plgDavid; 03/04/15 09:09 PM.
|
|
|
|
Joined: Mar 2006
Posts: 1,080 Likes: 7
Very Senior Member
|
Very Senior Member
Joined: Mar 2006
Posts: 1,080 Likes: 7 |
Yeah, its TMS5100 'patent' K coefficients and pitch table, but it uses the 'later' energy table from tms5110. The speak and math should use this same chip variant.
However, because of the way the tms51xx driver does variants selection (i.e. you must have a device of type tms5100 or tms5110a or etc and it must be declared as <required device ... > in the source code, it is a pain to have multiple drivers which use slightly different speech chips. There isn't any TMS_SET_VARIANT(*tms, variant) macro that can be used, each different speech chip has to have its own whole device, which is a giant mess, because the vsm rom module emulator has to point to the CORRECT device! It isn't possible to re-use the vsm definition from the 'speak and spell' as a root and have all other systems derive their config from that.
That's why I haven't touched tispeak.c to try to fix each device to use the appropriate speech chip yet.
The 'ME' on the chip sean has however implies it is revision E, while MFU would be revision F. There could be a change between those. Sean, do you want to try dumping the chip prom tables electronically (and testing the frame interpolation behavior?)
LN
"When life gives you zombies... *CHA-CHIK!* ...you make zombie-ade!"
|
|
|
|
Joined: Feb 2015
Posts: 160 Likes: 6
Senior Member
|
Senior Member
Joined: Feb 2015
Posts: 160 Likes: 6 |
hap: yep that looks like it! the LED displays and lamps are all in the same places and the size is right. I googled some but came up empty. Good now to give it a name.
|
|
|
|
Joined: Feb 2015
Posts: 160 Likes: 6
Senior Member
|
Senior Member
Joined: Feb 2015
Posts: 160 Likes: 6 |
re: the MM78. That patent with the code in it shows a polynomial program counter, but I clearly dumped 2048 bytes. The patent code's PC shows 64 byte banks or so it seems, which would be similar to the HD38800 and the TMS1000. But clearly the chip dumps all 2K!
Do they have a 5 bit binary counter that increments every bank, or what? Is the PC really just incremental? (the data sheet says the PC increments... but this is ambiguous I think, though implies binary IMO).
I suspect the 5th byte of my dumps is the 3C0h reset address, which is FFh on both dumps, and my data is inverted.
|
|
|
1 members (Duke),
171
guests, and
0
robots. |
Key:
Admin,
Global Mod,
Mod
|
|
Forums9
Topics9,331
Posts122,189
Members5,076
|
Most Online1,283 Dec 21st, 2022
|
|
These forums are sponsored by Superior Solitaire, an ad-free card game collection for macOS and iOS. Download it today!
|
|
|
|