Bomb Jack Repair Logs

From Arcade Otaku Wiki
Revision as of 00:34, 12 September 2012 by Davewellington (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Repairer: VectorGlow
Forum Thread: Bombjack PCB Repair

Symptom: Three horizontal lines of white pixels towards the bottom of the display (bottom board)
Diagnosis: Thought it was probably a video RAM problem but decided to see if the fault changed as the chips warmed up. Sure enough some of the pixels faded away so applied freezer spray to the suspects on the lower board. Sure enough the faded pixels came back good and strong when freezer spray was applied to the 6148 at location 4B. Tried piggy-backing a 2148 (I didn’t have any 6148’s) on top of that and that ‘fixed’ the fault, so removed the old 6148, popped in a socket and a new 2148
Cure: Replaced faulty 6148 RAM at location 4B

Symptom: All moving sprites are ‘shredded’ into vertical bars and data in one ‘bar’ is produced in the rest (bottom board)
Cure: Pin 10 stuck low on the LS283 at location 5E – replaced

Symptom: Horizontal lines (about half a character high) throughout the background graphics
Cure: Bad output on pin 15 of the LS194 at 7K – replaced


Repairer: Paul Swan

BOARD NO: 0072
GAME: Bomb Jack
MANUFACTURER: Tehken
DATE: 1984
HARNESS: TEHKEN-Sa

SYMPTOM (27/04/96)



No video. Game and sound seem OK.

SOLUTION



A logic probe indicated that R, G and B were always low but the Sync was running. The colour signals came from a couple of resistor networks and two LS174's. A frequency meter indicated a 6Mhz pixel clock on the CLOCK pin and a 57Hz clock on the CLEAR pin of both devices. The inputs to the LS174's came from three 2114 RAM devices (to give a programmable colour palate, one 2114 per colour giving four bits per colour), and the inputs were shared with the CPU via a couple of LS245 buffers. The RAM data pins were pulsing, as was the buffer control lines and CPU data bus on the other side of the buffers. The address and control lines on each of the 2114 RAM's were pulsing. Using a pulse injector on the 2114 data pins affected the video and sometimes caused moving colour bars to appear on the monitor, that cleared shortly after. This indicated that the LS174's must working, and at least some of the 2114 RAM locations. Setting up a logic analyser on one of the 2114 RAMS revealed the the CPU was writing valid data to valid locations in the RAM, but most of the address lines were always low when the data was read out during read cycles, resulting in the read out data being all-zeros for all three colours. The address lines were fed by a bank of LS157 two input multiplexors. The A-side inputs were probably fed from the CPU address bus and were pulsing. The B-side inputs of some of the multiplexors were always low. The mux control lines were pulsing. The B-side inputs were in turn fed from am LS273. The corresponding inputs to this device were always low, and it's clock and control line were pulsing. At this point I noticed some damage to some traces near by. Three traces were damaged and two were severed. Connecting the two disjoint traces together had no effect on the fault (ho hum). Continuing tracing back indicated that the LS273 latch bank was being fed from a bank of LS153 four-input multiplexors. These probably acted as the video mixer, selecting pixels from the three (presumably) video planes (one input was unused). Examining two devices in detail indicated the all the inputs, outputs and selects were low. All these inputs were fed from various sources. The SELECT inputs were fed from a LS148 priority encoder, the inputs to which were all low. Some inputs were fed from a couple of LS174's and some from an LS27. It looked as if an entire section of the video circuitry wasn't running. Pin 10 of one of the LS153's had a resistance of 15 Ohms to ground, and this was selected as a fault (even it it wasn't THE fault). The suspect net also went into an LS27 at pin 9 (an input). Another input (pin 10) of the LS27 had a resistance of 70 Ohms to ground. Tracing the second faulty net revealed that both of the faulty nets were the two that had been severed. Checking the third of the damaged traces revealed that it had a resistance of 6 Ohms to ground. Whatever had damaged the three traces originally must have had a high voltage on it such that it had distroyed some devices attached to the nets (operator probably caught the board on a live monitor chassis whilst removing it or something). All three damaged traces went off card onto the CPU board and disconnecting the ribbon cable cleared all three resistances on the video board. The faults were on the CPU card. Two traces originated from an LS157 and the third went into an LS04 near the ribbon connector and an LS08 at the other end of the board. Removing the LS157 and the LS04 cleared the resistances and replacing them fixed the game completely.

Repair Logs converted to wiki format by Brad from Aussie Arcade.