Bank Panic - Dead 1984 Arcade Board Repair
KONAMI Bank Panic(1984)
This 1981 Bank Panic from Konami arrived totally dead. Honestly I do not remember if the sync signal is present or not, but there’s no image or sound.
Diagnosis
First ran a pass of ROM check.
- ROM 1(27128 @ 7H): Bad
- ROM 2(27128 @ 7F): Good
- ROM 3(27128 @ 7E): Bad
- ROM 4(2764 @ 7D): Bad (only at 0x06D4 but still bad)
Out of 4 CPU ROMs we have 3 bad.
Other ROMs:
- ROM 5(2764): Good
- ROM 6(2764): Good
- ROM 7(2764): Bad
- ROM 8(2764): Bad (only at 0x017D and 0x067E)
- ROM 9(2764): Good
- ROM 10(2764): Good
- ROM 11(2764): Bad (only at 0x02D0) <- visible stuck pixel on texts
- ROM 12(2764): Good
Two FG ROMs are good. Out of 6 BG ROMs 3 are bad but only minor corruption issues. Probably still usable.
The repair
Replaced the corrupted CPU ROMs but still boot to the same screen. If only that easy.
The /NMI on Z80 CPU is stuck low. Upon check, found /IOREQ was only fired once on boot, and thus LS138 (6E) decode the output will only get fired once at pin 7, and this signal is used as CLK signal for LS273 (8H) to generate important signals like NMIEN, which then latches(through a LS74 (6D)) to become /NMI feed back to CPU. But /NMI never recovered from LOW meaning the access of memory probably failed. There are 74LS241, 74LS273, 74LS74, 74LS138 along the signal chain. Among them piggybacking 74LS273 at 8H and LS138 at 6E seems changes the symptom. Replace 8H and 6E first.
Replaced 8H and 6E sometimes can boot to a garbled screen.
Now turns out one of the memory transceiver LS241 at 8I is faulty on two bits. Replaced LS241, but didn’t have compatible chips, so did a botch with LS244 to be pin compatible. The botch can be easily replaced with a true LS241 chip. The only difference between LS244 and LS241 is an enable pin is flipped, and in this case it should be always enabled, so instead of connecting to VCC, it should be connected to GND. That’s all about the botch.
After that, the game boots and demos with sound, with some missing BG tiles and that can be caused by corrupted BG ROMs.
More test found that 74LS157 at 6J has an /G pin should be connected to GND but trace broken. Fixed the trace by wires.
More test found that all CPU data bus and address bus appears to be working, it’s difficult to answer why part of the graphics are missing. Turns out the whole BG tile is missing, the visible ones are actually all FG tiles. Arbiter 74LS127 at 6A has all BG index floating, so BG lookup table PROM at 5A isn’t working. Turns out the chip is alright, but pin 13, 14, 15 which should be connected to GND has broken traces. Fixed the trace by wires now fixed the BG tiles, mostly, despite the ROM corruption.
With new 2764s all corrupted ROMs are reprogrammed or replaced.
Sadly ROM 11 (2764) is still having a stuck bit, which result on the screen a stuck pixel in text. Erasing and replacing.
2764 has a high rate of failure. It took me several tries to find a good enough 2764 to both (1) Reprogrammed and verified correct (2) Installed and verified correct with cleaned socket. Finally ROM 11 is perfect without stuck pixels.
A fully working board to conclude the repair.