Repair Diary #4 - Simpsons (Work in Progress)

    This site uses cookies. By continuing to browse this site, you are agreeing to our Cookie Policy.

    • Repair Diary #4 - Simpsons (Work in Progress)

      I purchased this clean Simpsons board inexpensively since it has sprite issues - the internal ROM test would pass successfully, but the title screen is obviously garbled,and during game play, the character sprites randomly glitch.



      I started diagnosing by first examining the board traces very carefully under the microscope - I did not spot any bad traces. Next, I grabbed the board schematic from a PDF of the manual, Googled the Konami custom processors, and marked them on the schematic.



      The schematic and part lookups showed that the Sound system (CPU, ROM, memory addressing logic,RAM) is located in the upper-left of the board, the CPU system is located in the lower-left of the board, the Background rendering system is located in the lower-right of the board, and the Sprite rendering system is located in the upper-right of the board. This board is experiencing a Sprite issue, so I next focused my diagnosis on the upper-right area of the board.


      I next checked for continuity on all of the lines coming out of the two Sprite rendering GPUs – I did not detect any continuity issues. I proceeded to get out the oscilloscope and start probing the ROMs, memory addressing logic chips, and RAM chips (referring to the datasheet on each chip model to make sure I was probing the right pins and interpreting the output correctly).

      It was here that Imade a discovery – the 32K static memory chip located at position 1J exhibited strange behavior on its data lines.


      When functioning normally, a memory chip should have 5V power on the Vcc pin, read high or show regular activity on the Enable pin, show activity on the Addressing lines, show activity on the Write Enable pin, and show activity on the data lines.
      Here is how the same model chip elsewhere on the board reads on the oscilloscope on the data lines:



      And here’s how thechip at 1J is reading on the data lines:


      I confirmed that all of the other pins on the memory chip were showing activity as expected. It appears that this memory chip is faulty. I have ordered a replacement chip – I’ll know soon if I diagnosed this issue properly.
      Images
      • IMG_3180.jpg

        421.66 kB, 1,104×621, viewed 85 times

      The post was edited 1 time, last by ShootTheCore ().

    • Hello,

      Yeah, indeed 053246/053247 are sprite generators ICs, in Konami games these chips always work in pairs.
      The memory IC @1J is a standard 6264 RAM, if you got any junk boards you will find it on them.That IC serves for the purpose of being so called "sprite list ram", but like in most konami games, at startup is being tested, so if the game passes its boot check, pehaps the ram is good, you should double check it.
      Heres what you can do for a start : with the board powered off, short pin 12+13 then power the board, repeat again for pin 13+14 (data bus +GND), see if you get a red BAD at startup @1J., the problem is likely not there.
      Repeat the same procedure for SPRITE maskrom @3N, 8N, 12N, 16L, if they are 42pin pinut = 27c800
      if they are 40pin = 27c4001, try to short DATA pins first, plz dont do this very often, as you may kill the graphic cpu. With these pins shorted , go to MASKROM test and repeat the test for every rom. Same, if you get a red BAD, the problem is likely not there.
      Your scope readings looks funny, thats not the way a digital signal is meant to look like... either you got faulty probe, or the scope is not working poperly.The ground wire of the probe is OK/good connection ?
      Have a good day, Dragos.
    • The scope readings look strange because it's a cheap hobbyist scope that can only track in the kHz range - you don't get the nice, square angles that you see on a proper mHz scope. But the cheap scope does let you see if a line is pulled high, low, talking, or floating at the very least. I prefer it to the logic probe, but a nicer oscilloscope is definitely on my To Buy list.

      I already desoldered the memory chip, so I won't be able to do your tests until the new chip arrives. If the new chip doesn't solve the problem, then I'll definitely apply your troubleshooting steps next. Thank you for the great advice!
    • Hello,

      You shouldnt have desoldered that ram IC.
      Anyway, clean the solder residue of the ram's pads, and power up the board.If the board's selftest reports a bad ram @1J , the ram most likely is ok.But keeping in mind that ram is a "evil" fujitsu, its worthy double check it.
      When you got some finances/ budget, i recommend you to buy a decent (read= chinese/poor man's programmer) that has ram & ttl test.Very useful. :thumbup:
      Regarding your scope, it will not help you, instead it will confuse you, because it gives false readings ?(
      Patience, perseverence and g***le are your friends in repairing arcade boards.

      Have a good day, Dragos.
    • Hello,

      Edit: i have powered up my Simpsons copy and i have ran some tests, the results are not what i have expected:
      1.At boot up the game does NOT test the ram @1J & @16G & @16F. Its the first time i have seen a konami board behaving like this.Perhaps they have borowed the "wisdom" from Capcom's CPS1 boot/self test that tells you "ALL GO", and the board is toasted...... :thumbdown:
      2.Shorting DATA pins @1J , prevents the game from booting ! , also shorting the ADDRESS pins, results in major sprite corruption, most likely the RAM is toasted, or its connections or, rarely the video processor.It would be nice if you can test the pulled ram, to see if indeed its toasted.

      Have a good night, Dragos.
    • Awesome! Thank you :)
      Btw if you happen to have an TL866A Programmer, this little jem is capable of testing many rams (2015/2018/6116/6264/62256/ecc ecc), mostly all TTL 74LS/HC family, even PAL/GAL.
      If you have some cash to "throw" i suggest you guys buying it, it's more than the usual eprom programmer (wich of course does pretty well tho)

      The post was edited 2 times, last by TheDeath ().

    • Hello,

      I have a "poor man's" programmer, named wellon vp390, it works quite well, actually it saved my ass quite a few times !
      Regarding the schematics of the Simpsons, the video section is not available, but the user can work on schematic of other game of that time period : lightning fighters, parodius da, surprise attack, cowboys of moo mesa etc, where the ICs are the same/applicable Once the user have repaired some boards of this type, he will understand the topology/layout of each section of the game/board. 8)

      Have a good evening, Dragos
    • good old or should say bad old fujitsu ram combined with being on the edge of the mobo where static electric discharge can happen by people pulling the boards in and out of cabs over the years.

      Best thing with the old fujitsu memory is to cut it out and pit in a socket vs de soldering. Much safer and easier in the longrun ;)
      You will get it running again right Im sure
    • I usually check these things on konami boards: Fujitsu TTL and Rams, 051550 Custom (master reset + clock) and priority chips. I usually left alone CPU's (even if some times could be bad, mostly 68000 12Mhz models, 16Mhz are way reliable) and tile/palette customs, those also are generally good (if not, lifted pins!).
      Also, for first, i do check priority proms that stores the programs itself (if they do, not all). I had MANY boards trying to troubleshoot (black screen) and i ended test proms with bad pins, bad program or even erased complitely. If they are socked, give them a shot and comprare them with mame sets.
      Last but not least, resistance between ground and pins, if there's not much resistance the chip could be ok but not good enought to work properly.. generally below 100ohm it's a signal for it
    • Hello ,

      TheDeath: the custom 051550 doesnt generate clock, the clock is generated by means of a Crystal oscillator, usually 24Mhz , then divided by 2 , then feeding the CPU.051550 generates reset + watchdog, and coin1 + 2 inputs.
      Starting circa. 1990, konami abandoned using priority Proms, instead they switched to a priority chip 053251

      Have a good morning, Dragos

    • I was wondering if you got this one?
      I repaired mine few weeks ago and i had similar problems with it, it ended up being a bad 27c160 equivalent prom, the upper one on the board. The mask rom check gave me "ok" but in the end it was not (miscompared on dump). I also reflowed all of the remaining proms underneath that one, they are very close to the edge and i found that solder prompt to fail easely.
      I also noticed that, while checking mine, this board boots up even without the 051550 custom. This is the master reset and i have no idea how that could happened without it o_O