Lucky Moron PCB Repair - #3 Strikers 1945 II

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

    • Lucky Moron PCB Repair - #3 Strikers 1945 II

      Lucky Moron PCB Repair
      Episode 3: Strikers 1945 II
      (or: Measure Once, Cut Anyway)

      Even though I know about The Glitch now (awesome stuff @Hatsune Mike!), I still suck at Dodonpachi and can't even 20cc it, so it was time to put that board away and load up to some good old-fashioned Psikyo shooter goodness. Drop me some P's! Upon plugging in my pcb though, I was greeted with a horror of horrors. An affront to the senses. A crippling, game-breaking error. This travesty should be obvious to even the most novice of arcade hobbyists:



      That's right, you guessed it, you savant!
      Those are clearly not perfect circles.

      Dear reader you should know, I've had the uncanny ability to draw absolute-perfect circles, freehand (checked with a BJCP-Certified Eccentricity Interferometer to the 6th decimal). Here's an overlap I whipped up, so you can clearly tell the shapes masquerading as "circles" are indeed ovular, and that I should probably adjust my monitor slightly to solve the problem. HOWEVER, I HAD JUST CALIBRATED MY MONITOR WITH DODONPACHI so obviously there is a RAM fault plaguing this pcb.



      As you know I hate RAM issues, and I love procrastination, so let's check all the other basics first. My 5v power line is hovering around 7.2v, so first I turn that down. Whoops! Still no change. Video issues like the above could easily be due to aging or leaking capacitors. Check out this one, just waiting to bulge someday! Who ever heard of Nichicon anyway!?



      Pro tip -- if you lick a capacitor and taste acid, you should visit a hospital. I've heard that all black capacitors are tragedies waiting to happen, so we'd better recap everything! As anticipated this (very rudimentary) troubleshooting step didn't fix the abhorrent distortion. Also my tongue burns a little from all the licking. My money is still on the RAM.

      Still, I need to run through my basic troubleshooting checklist, so it's time to double-check the ROM data. Admittedly it's a bit of a long shot, but it could be an issue with bit-rot, or a factory error (I've seen TMNT2 data on Quiz Gakumon pcb's for example!). The program ROMs are socketed, but every other ROM is surface mounted, and compared to the time it would take to look through the MAME driver it's faster (and way less reading) to just use a heat gun and remove all of them at once.

      BEGONE!!!



      Let's boot it up to see how we're doing.



      Perfect!!! Those pesky ovals are gone. Maybe it was a ROM issue, perhaps the circle sprites were bit-rotted. Assuming I'm right, which I probably am, let's replace all the ROMs on the board. As we have taken all the original chips off, we should archive them for safekeeping. I always store my original ROMs in vacuum sealed cryo-stasis and replace them with EPROMs that are both a) newer and b) less old so that if/when I ever sell a pcb in the future I don't have as many hours on the original data chips.

      Unfortunately (if you like easy jobs) or fortunately (if you like external wiring) there are no direct replacement ROMs available for these SMD's, so we'll just enlist some MX29L3211, which run at 3.3v. It's never a good idea to run these on 5v power or the internet will completely mob your ass, so we'll affix a voltage regulator to a nearby capacitor to create a proper voltage line.

      OK, ROM data ROM data... let's write some ROMs. Looking up the ROM data from my MAME romset I realize what was most likely the issue all along:



      There were 8 missing ROMs! Must have fallen off during shipping; I did buy this off of KLOV, and it was shipped inside a two flat rate USPS boxes duct-taped together and filled with sand for padding. Wouldn't you know it? There's even an extra program ROM that was supposed to be on here! Surface mount U16 - the file is 8mb but half the data is empty, so AM29F400B should work for that. Looking at that datasheet, it's a 5v chip we're gonna need to lift-up and tie RESET and write enable (WE#) pins HIGH to set it to Read. (and yes, the below red circle is also hand-drawn-perfection).




      After writing the rest of the data to the MX29L3211chips, we similarly need to lift-up the Vcc and the write enable pins on each, and send those both 3.3v from a voltage regulator. It's also a good idea to both add a decoupling cap to each ROM and also clean up your flux, neither of which I did because I'm lazy. While I'm wiring everything up I notice a few missing ICs on the board too. There's a missing header (could this thing have 3/4 player mode??!), simple logic 74HC244, and some resistor arrays. I find those chips on a sacrificial non-working Psikyo PS4v2 pcb (Taisen Hot Gimmick 17 Electric Boogaloo) that comes in handy from time to time. I also add the little smd cap thingy (it looks very important) by just stealing one from a nearby empty ROM site.



      Sure it's a little extra effort, and the PCB looks awful afterwards, but we're saving the original ROMs after all. What is original hardware for if not to sit on a shelf while we play MAME! Here's the finished product, ready to test. AMB eat your heart out!



      Returning to the hypothesis central to our scientific methodology, if the oval nonsense was previously caused by bad ROM data, it should be all gone!
      Let's power it up!!!!!
      SHOW ME DAT PHAT ARC!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!



      Fuck me, they're worse ovals AND it's running horizontal now.

      It's in moments like these that I wish I were more diligent like @twistedsymphony, who painstakingly researches every project beforehand and pauses after each and every step of every mod to photograph, note-take, and evaluate any differences between expectation (i.e. his custom-compiled MAME build) and reality. Then if he sees something awry (as if that ever happens to him lol!), he only ever has to take one step backwards. Additionally, when he's done, he's already written a manual for someone (and I mean someone else, not me -- I just wing it!) to follow.

      I on the other hand, basically make every hypothetical adjustment all at-once, and forget all details shortly thereafter. So before my mind gets foggy, I restore the original ROMs from cryo-statis, rip off all my other mods, and generally reinstall everything the way it started. A slightly misshapen circle is one thing I can learn to abide, but I can't have a TATE game running in YOKO, oh no.

      Then I get to thinking (for once). Maybe this Strikers board is actually fine! I do recall that my Dodonpachi was acting a little bit odd... I mean, I keep dying over and over and I'm usually pretty good at shmups, or at least Giga Wing (2 player of course). Maybe it's my DDP board that actually needs some replacement caps, or a ROM teardown job, or an oscillator is shoddy and the circle_draw function relies on it's precision!

      Logic dictates (and here's my real Sherlock moment...) that if I had originally calibrated my monitor to my Dodonpachi, but Dodonpachi was compromised, then all I really need to do is calibrate my monitor to Strikers and it'll look fine. So, with one hand, I unlock my cabinet's security door, open it, and resolutely reach towards my monitor's remote board. Ever so slightly I finesse the one singular HSIZE potentiometer by one-quarter-degree, clockwise.



      Circles; bitch.

      Then, breathing a heavy sigh of OCD-sated relief, I add a Dodonpachi repair note to my projects list:
      "Probable eccentricity error, starting by removing all the customs"
      But that's the next log! For now...



      Game Fixed. was fine the whole time.
      cabs: sega blast city x2 | taito vewlix c + egret ii
      links: klov/vaps games list | custom fight sticks
      working on: playing more games instead of working
    • xtrasmiley wrote:

      Nicely done. I'm going to follow your lead here on my Haunted Castle PCB. It's OK on the circles front, but the damn game is too hard and sort of sucks. I'll start with the caps then then move to the RAM...
      Honestly if the game is too hard I think just swapping out the CPU for a slower one helps a lot. Sort of makes it “dumber” you know? Might want to do that first.
      cabs: sega blast city x2 | taito vewlix c + egret ii
      links: klov/vaps games list | custom fight sticks
      working on: playing more games instead of working
    • I'm confused now... What was not the issue and what did you not fix?
      Looking for:
      - faulty Space Harrier or Enduro Racer
      - Super Nintendo CPU (3 needed)
      - Sega trackball I/O board 837-8685 (will be reproduced and gerbers made freely available)



      "I'd 1cc games but I have a real life."
      01010011 01000101 01000011 01010010 01000101 01010100 00100000 01001101 01000101 01001110 01010101 00100000 01000111 01010101 01011001
    • ekorz wrote:

      xtrasmiley wrote:

      Nicely done. I'm going to follow your lead here on my Haunted Castle PCB. It's OK on the circles front, but the damn game is too hard and sort of sucks. I'll start with the caps then then move to the RAM...
      Honestly if the game is too hard I think just swapping out the CPU for a slower one helps a lot. Sort of makes it “dumber” you know? Might want to do that first.
      Yeah, I'll desolder the CPU and swap it with the one in my TurboDuo, since that Castlevani is so much easier, thanks for the tip!