[Origin Systems] We produce state-of-the-art interactive movies, simulations, and computer adventures. Life's an adventure. Play it hard. ---------------------------------------------------------------------------- [Previous] [Next] [Hitlist] [Get Thread] [Author Profile] [Post] [Reply] ---------------------------------------------------------------------------- Article 17 of 91 Subject: More board level fault diagnosis... From: swanp@btlip22.bt.co.uk (Paul Swan) Date: 1996/05/07 Message-Id: <4mmsrf$aq@pheidippides.axion.bt.co.uk> Article Segment 6 of 6 (Get Previous Segment) (Get All 6 Segments) SYMPTOM (20/04/96) ------- (1) The background was corrupted during game play (2) The sound volume was very low. SOLUTION -------- (1) Flexing and tapping the top board made the problem worse and intermittent. Changing the ribbon cables had no effect - the fault was definately on the top board. The nature of the corruption was that large graphic blocks contained the wrong peices of backgroung in some columns, and this was carried down the screen as the screen scrolled down. This suggested the the background RAM was either not being loaded correctly or the output of the RAM that was used as the character (ROM) address was being corrupted. Re-soledring the top left ribbon connector, the video RAM, the ROM address latches above the RAM and four or five miscellaneous TTL chips directly below the RAM fixed the game. (2) Visual inspection and trace of the speaker pin back to the amp revealed a leaking 1000uF capacitor with a leg partially pulled out. Replacing it fixed the sound. __________ BOARD NO: 0047 GAME: Ghosts'n Goblins MANUFACTURER: Capcom DATE: 1985 HARNESS: CAPCOM-S SYMPTOM (20/04/96) ------- The background colour was corrupted. SOLUTION -------- Flexing and tapping the boards made the problem worse and intermittent. Changing the top left ribbon cable fixed the game. __________ BOARD NO: 0089 GAME: Karate Champ MANUFACTURER: Data East DATE: 1984 HARNESS: DECO-S SYMPTOM (13/04/96) ------- ROMS were missing, but were replaced and verified as correct. The game generally failed to initialize, but would occaisionally boot and run correctly. SOLUTION -------- Flexing and tapping the board had no effect. Changing the inter-board ribbon cable had no effect. With no physical intermittence evident and no other fault evident when the game was up and running it was time to look at the power-on-reset. Tracing the reset lines from both processors revealed that the Z80 closest to the back left corner of the board was the master, with the reset for the sound Z80 fed from a latch controlled by the master. Tracing the master Z80 reset back through an LS04 up until it terminated in a custom SIL module in the front left corner didn't reveal anything suspicious. To veryfy that the power-on-reset was the problem, a pulse injector was used to pulse the reset line on the master processor. This started the game correctly every time. Disconnecting the switched mode supply and re-connecting it (using a dummy load to maintain the supply voltage) always started the game where as switching the supply off and on didn't. In the end, the board was declared good and the fault put down to the switched mode supply taking longer to stabalize the voltage than the game's reset period could accomodate, since the gane was probably designed with a linear supply in mind with quicker stabalization at power on. (or, my supply is slow to stabalize - take your pick). ======================================== END ======================================== [Previous] [Next] [Hitlist] [Get Thread] [Author Profile] [Post] [Reply] ---------------------------------------------------------------------------- [ Home ] - ] [ Search ] - ] [ Contacts ] - ] [ Help ] ----------------------------------------------------------------------------