Repair LogsArcade

10 Yard Fight - 1983 arcade board repair

IREM 10 Yard Fight(1983)

The board looks in good condition.

Powers on, very glitchy graphics, seems all sprite except ASCII text are gone. Sound is missing too, volume pot did make noise so amp part might be good(Later proved no sound in demo mode).

Diagnosis

Confirmed the word “ヤード” that the board is a Japanese 10-yard fight.

The B board has all EPROMs exposed their chips, although B board is usually covered by A board. This could very likely lead to ROM corruptions.

ROM pass

B board:

  • 5B, 5C, 5E, 5F, 5J, 5K, 6x 2764 SPRITE ROMs: PASSED

A board:

  • 3C, 3D 3E 3x 2764 CPU ROMs: ALL FAILED
  • 3M, 3N, 3P 3x 2764 TILE ROMs: ALL FAILED

First thing to do is to reprogram the 3x CPU ROMs.

The repair

Replaced all 3x CPU ROMs and 3x TILE ROMs. Still getting a similar glitch.

Probed the A board, found memory on 1K and 1F has no data activity. Less likely both are faulty. Keep messing around, the 7474N chip on 1N has a floaty pin. Piggyback a 74LS74 not quite the same but starts to display some graphics, especially title screen. That confirms the sprites are almost back.

A lot of poke around happens. First found the 74LS74 chip on centre board at location 6P is getting a really slow rising signal at pin 3. It gets the signal from top board’s 74LS367 at location 7B pin 9. Tried to lift those pins to no avail. Only found on schematics this pin is designed to connect to a capacitor. This is a by design slow rising signal. Red herring.

Then checked schematics D0B - D7B are all slow rising. For example data bus bit D0B connects to centre board’s 74LS245 at location 2E at pin 18, at the same time 74LS374 at location 2D at pin 2. Lift both pins only to find the signals are still the same. They are all connected to custom chips at 5-6K and 5-6M. So the chips are doing the low pass filtering.

Tested those chips, although having no accurate information about those chips, they are probably tile generators - taking bits from the tile ROMs and synthesising them into bitmap information and passing them down.

Probed the other chips around the tile generators and they seemed to be fine.

Try piggybacking one known good HM6116 RAM on the existing MSM2128 RAM, the game boots with some copyright texts but seems indexed to the wrong characters.

The field/tiles are mostly back with that. So at least one of the work RAM with the tile generators is bad. Confirmed at least the one at 4A is bad, that’s why the left part of the field is always filled with moving glitched pixel strips. The other one at 4C is probably marginal too.

RAMs need replacement.

Replaced 4A RAM, all the text and graphics became normal. I thought there will be more work finding out what shuffles the text, lucky.

Actually this board may have a set of ROM the Internet has yet not found, because I put back the original “corrupted” ROMs and it is still perfectly working.

Anyway, this unit is fixed.