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

I got a Funtronics Jacks game. Opened it up, and the CPU die (COP410L) is bonded directly to the PCB with no packaging, like the Channel F carts. So I won't have to decap it, but it'll be awkward to photograph since my XY platform is only about 1" square. I guess I'll use a little double-sided tape to hold it in place. I'm not sure if the pins needed to dump it are bonded out. If not, it'll be tough for me to electronically dump it.

I also got a 1978 TI Little Professor, which is the one where they added the On/Off and Level buttons. The chip is marked TMC1993. I decapped it and the die is marked 1990C-c3C. It looks a lot like the 0970, but it has a different pinout; all the LED connections are on one side and all the keypad connections on the other. There's no oscillator input, and the keypad rows and LED digits are on different outputs.

the 2801 ROM tables have already been dumped. Of course it could be different from variant to variant. I could run the tedious PROMOUT logic on your unsoldered chip, though I do have a SnR here. Also dumped two new SnR carts. Only two left to dump now.

1990? Interesting =) I wonder what else they changed from TMS970. Is the OFF button done with discrete logic, or on-die?

Like the TI-30, there's nothing but the chip, LED bar, battery clip and keyboard, so the power circuitry is on-die, like the 0980. 9 pins of the chip go to the 4x5 keyboard matrix, 16 pins go to the display, 2 pins are power and one pin is a no-connect.

Here's a really poor die shot: www.seanriddle.com/lp78.jpg

It looks like the middle of the die was just copied-and-pasted from the 0970.

From Datamath.org, the TI-1000 also uses the 1990. Dataman and TI-45 use the 1980, which I think is just the VFD version of the 0980.

TI really put a lot of effort into making their high volume stuff like that 1 chip. I remember as a kid opening up some TI stuff and quickly being disappointed at how there was literally almost nothing inside of them except the chip/display/keypad like you said.

re the COP410 bonded out, shouldn't be too difficult to determine if it's dumpable or not, just find pin 1 and count bond wires? I've had to do that for NES pirate cart mappers before to figure out the pinout.

sean: did you get that MM77 tested out?

Looks like the HD38800B part dumps the same as the A. the problem was stupid. my socket pins got bent a bit by my HD38820 adapter so they weren't making good contact on the chip. Measuring with the meter showed a few opens so once I bent the pins in slightly it worked. Unfortunately no one makes 42 pin ZIF sockets. They make 48 pinners so I will have to use one of those.

Its just that the cd2801 sean has is a revision E part, the one you dumped is revision F, so there might be a interpolation behavior or lpc/pitch/energy table difference? (probably a minor one if any)

LN

Yeah, the Kosmos Astro astrological calculator is the exact opposite- it's crammed full of components. But it was for a different market.

I just haven't had time to get the Jacks PCB under the scope yet to see what pins are bonded out. I'm not sure if they would bond the pins used for test mode- I guess they might.

I've just got an MM76EL right now. The game runs it with a 9V battery, so I tried dumping it at 9V and got nothing. With the test pin at VSS it looked like it was running normally, but at VDD there was nothing at all, even it I reset it. The chip takes 6.5V to 11V, so I tried 7.5, but that didn't work either. I was probing the RIO port with my Logic16, which isn't really supposed to be used over 6V, although I've used it at 9V before. I didn't want to risk it at 11V, but I gave it a quick try and actually got something. So I'm going to add a level shifter to help make sure I don't burn it up.

The data I got is weird- on power up I got about 1/4 second of pulses on RIO1-RIO8. Zooming in, it's actually 18 groups of pulses. All 8 signals go high for 3.3mS, then there's about 0.5mS of pulses. The 18 groups are similar but not the same. So I'm wondering if I'm in some other test mode, not ROM dump mode. The data sheet says test mode "tests ROM and allows testing of the RAM and instruction logic".

I'm letting most of the pins float, including PIx. Were you driving any other pins when you got your data?

Sean: on Entex Baseball 3, is it possible without the MCU to measure the output frequency of the RC circuit? (with and without the PRO switch enabled)

sean: I saw some weirdities on the data output of the MM78 when it was near the edges of the test voltage range (i.e. 11.5 and 12.5) where it'd seem to oscillate between test mode and operation.


re: cop 411 dumping. I had a hard time trying to get this chip to dump. turns out the damn docs are WRONG! they tell you to use 04h on the LEI instruction (33 04) but then the signals are showing it as 33 44. turns out it has to be 33 64 or else it won't work.

the docs for the chip say the second byte of LEI is 6x, so I changed it and sure enough now data is coming out. I found you can drive the RC input with a regular clock and it works fine, the even mention that this is possible.

The charge dump transistor is very weak sauce so a regular PIC IO had plenty of oomph to overcome it. I checked on the scope and it looks good, no funny business.

The timing on the data into the chip is kinda stupid- you have to input data 0-200ns after the rising edge of SK so what I did was let CLKIN go high, then waited about 15-17 nops, and then output the new data so it occured at 100ns after the rising edge.

This seems to work fine and I got a dump out that repeats every 512 bytes. I am not sure it started in the proper place though, so I was going to poke around with it some more. I ran out of time last night but did finally get data to come out.

Today I received a tandy "cosmic 1000 fire away' and wouldn't you know it, the damn thing has a bastard TI part in it. drat.

Also got a ms pac-man tabletop that I can dump (HD38820) and the frogger tabletop with the mitsubishi micro I will try dumping