Jump to content
Due to a large amount of spamers, accounts will now have to be approved by the Admins so please be patient. ×
IGNORED

Wonderboy III Monsters Lair Repair Log


Recommended Posts

Had @Slypinball's Sega Wonderboy 3 boiard set on the bench this week, the fault when he sent it to me was that there was no video, but he said initially it did give him an all-blue image, but eventually even the blue gave up.

 

IMG_7592.jpg

 

Except when it arrived here it was totally dead, when connected to my cabinet via a System 16 to JAMMA adapter it did nothing. Am always slightly concerned with any board that has a suicide system in it, if the CR2032 battery in the Hitachi FD1094 is original, then it would be nearly 30 years old. Clearly they will all die one day, and many have already run out of puff. I did a quick dump of the two program ROMs at A5 and A7, these are encrypted with the same key that was loaded into the FD1094 at the factory and if they don't match the board will do nothing. So when one came back as "unknown" and one came back as Tough Turf (another Sega System 16 game) I assumed that was why the board was dead, that the ROM board did not match the FD1094 and that one ROM was probably dead.

 

Turns out the info in the WinROMIdent database file I am using is wrong for these checksums, I can understand it not knowing all games (it knows about Wonderboy III sets 1, 2 and 3, but not set 4 - which is this game) but having totally wrong data is another thing. Clearly it is wrong as I dumped a graphics ROM for Wonderboy III and it came back as being both for Wonderboy III and Tough Turf - a double dragon style beat em up. Anyway- reseating all the ROMs and the encrypted CPU finally got me back to the starting position of a game that plays blind.

 

toughturf.png

 

Working on a board with a suicide system is risky, especially as it could pack up at any moment, so I swapped in a decrypted Golden Axe ROM board and a 10Mhz 68000 CPU to work with. It played blind too, so at least the fault I was chasing was back. The first port of call was under the ROM board where the RGB signals present at the edge connector first appear.

 

IMG_7594.jpg

 

The three resistors top right (R25/26/27) are where the looooong tracks that carry the RGB signals to the edge connector start.

 

IMG_7595.jpg

 

Really odd to see the 16 pin black and 8 pin blue IC sockets used back-to-back to make the 24 pin socket for the 2018s, they are original from the factory so Sega either ran out of the right sockets or was trying to use up some old stock.

 

IMG_7597.jpg

 

The most obvious thing in this area was that someone had been here before and had replaced the LS244 at J7 and had been playing around with the 74ls125 at J6. The 74LS244 is not the original one as the date code differs from all the other 244s on the board. I couldn't tell if the 74LS125 was replaced, removed, or if it had just been resoldered. Eitherway it was a bit of a mess, flux everywhere, far too much solder on most of the pins, and hardly enough on others - you can see daylight through the middle of the lower row.

 

The 5th rule of board repair is to always be suspicious of other people's work, especially if it doesn't look they knew what they were doing. On a board of this era the chances that they damaged tracks, tore out plate through holes and up sections of the board is high. What is also surprising is that it is very common in these cases to find they fitted a faulty part installed. Either because they killed it during a torture session getting it off another board, or just roasted the crap out of it when they installed it. I suppose equally possible is they bought a new part, but put the old one back in when they chucked the board in the bin.

 

In this case it was just board damage, the 74LS244 and 74LS125 tested as fine in my EPROM burner, although pin 1 fell off the 125 during desoldering so I think it had taken some abuse.

 

IMG_7600.jpg

 

The plate through hole for pin 2 has been torn out of the board, taking the track that goes to the via under the "S" with it, and breaking the connection of both to the track that heads down to the bottom of the shot. Pin 20 has had a good roast and taken some abuse in the struggle to get it free of the board.

 

IMG_7602.jpg

 

At the other end of the LS244 the ground pin has had much the same treatment as pin 20. The 5v and ground pins connect to a large network of fairly wide copper traces inside the board, these act as very effective heatsinks so without the right tools it can be a battle to get these joints to let go, without burning the board, or tearing the guts out with it. Pin 11 (second from the right on top row) has no hole plating remaining and the link to the track heading off to the right towards B15 on the ROM board connector is no longer connected.

 

Not pretty, but not too bad, just a job for some hookup wire on the underside once the socket was refitted.

 

IMG_7603.jpg

 

There was no track damage for the 74LS125 so I fitted a socket and put it back on the board. Normally I don't fit sockets for TTL chips, but in this case I wanted to be able to swap the LS125 if there were any glitches, considering it could have been cooked.

 

IMG_7604.jpg

 

With all that fixed I powered up the board and got the same fault - still no video, which is not really surprising. The video signal is made of three channels, Red, Green and Blue signals. If the board was only showing a blue image before it finally dies then I guess the original fixer was chasing the missing Red and Green signals.

 

The most logical place to start seemed to be at the final active stage of the video output, basically at the palette SRAMs which sit right next to the 74LS125 at J9 and J10.

 

IMG_7613.jpg

 

Both seemed to be doing very little on the scope. The address lines and the control lines were all doing what they should, and the activity inbound looked normal...

 

IMG_7614.jpg

 

...i.e. decent deflection between logic low and logic high, with clean transitions.

 

The data bus outputs were a different story, all low and twitching, putting out nothing at all.

 

IMG_7616.jpg

 

The fact the board had lost red and green initially, and finally blue does make sense, these boards have 4 bits colour depth per colour, giving 12 bit colour in total. This is striped across the 16 bit space provided by the two 8 bit RAMs.

 

schematic.png

 

Therefore one of the 8 bit SRAMs handles the 4 bits for red and 4 bits for green, and it presumably died first. The other chip is only providing the 4 bits for blue, with 4 unused bits, and it seemed to have survived a bit longer before packing up. Cause of death for all is probably static handling damage.

 

One trick when met with apparently dead RAM is to somehow inject a signal into one of the data output pins to observe the effect on the fault. In this case it was easy to short the address line A2 onto data line D0. The data is clearly going to be meaningless in the contect (ie address line signals are not colour data) but it can give you a clue.

 

First up was a red and green clue, yellow being a product of those two in RGB land...

 

IMG_7608.jpg

 

...and a blue clue was available on the other chip.

 

IMG_7611.jpg

 

Time to whip them off the board, cleanly this time.

 

IMG_7617.jpg

 

Both thoroughly failed off board tests, showing out of spec pins.

 

BadRAM1.png

 

BadRAM2.png

 

Thankfully I had refilled my stock of 24 pin 0.3" sockets, the last repair I did used up the last option for salvage from scrap and Jaycar no longer stock them. I think I bought enough to last a while...

 

IMG_7618.jpg

 

...especially as all of those bundles are double-sided. It would have looked "interesting" when that parcel came through the customs scanner.

 

IMG_7619.jpg

 

New chips installed.

 

IMG_7620.jpg

 

Confident that that was it I swapped in the FD1094, the Wonderboy III ROM board and flicked the switch.

 

IMG_7624.jpg

 

IMG_7628.jpg

 

Fixed!

Edited by Womble
Link to comment
Share on other sites

Well played as always.

 

Thankfully I had refilled my stock of 24 pin 0.3" sockets, the last repair I did used up the last option for salvage from scrap and Jaycar no longer stock them. I think I bought enough to last a while...

It all depend of your work load. I restocked 50 of 28 pin sockets (narrow 0.3") few months ago in prevision of a massive repair job (almost 100 boards) and they're almost all gone now...

But I'm having a break from repairs at the moment, too much time consuming.

Link to comment
Share on other sites

Well played as always.

It all depend of your work load. I restocked 50 of 28 pin sockets (narrow 0.3") few months ago in prevision of a massive repair job (almost 100 boards) and they're almost all gone now...

But I'm having a break from repairs at the moment, too much time consuming.

 

Very true, sadly I don't need and 28 pin sockets as I can't find a source for 0.3" 6264 SRAMs and am virtually out of them on scrap. They went from being cheap and plentiful to rare and exxy pretty damn fast.

 

Whereabouts in the Waikato are you based? I lived in Hamilton for a couple of years.

Link to comment
Share on other sites

I can't find a source for 0.3" 6264 SRAMs and am virtually out of them on scrap. They went from being cheap and plentiful to rare and exxy pretty damn fast.

Just use 62256 instead and if pin 1 isn't connected to anything on the board tie it to pin 2. This way is much neater than running a wire from pin 1 to Vcc or Gnd.

 

Whereabouts in the Waikato are you based? I lived in Hamilton for a couple of years.

Hamilton!

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...