Another board from the ‘portughese’ box : Tatsujin Oh, a great shoot ’em up released by Toaplan and known outside Japan as Truxton II.

Board was booting but sprites were mostly missing and other GFX glitches present too:

There was an obvious reason of this.The custom ASIC  ‘GP9001’ (which is the graphics controller of the system) was damaged : some pins were missing, other bridged, two traces going to it were ripped too.

Here’s a close-up under a microscope:

Given the extent of the damage the only possible solution was the replacement of the ASIC which is a very delicate operation (not for everyone…) because of  its package (208 pins in a QFP package with very fine pitch).For this purpose I’ve been sent by the owner a dead FixEight PCB as donor:

I removed the damaged part and cleaned the area preparing it for the trasplant:

The spare removed from the donor board :

After soldered it I took care of rebuilding the ripped traces using some AWG30 wire :

Testing was successful, board 100% fixed.

 

Received from Portugal a couple of faulty The Simpsons PCBs.The first one:

When powered the board up I experieced sprites issue, some parts were missing:

The MASK ROMs test reported three bad 8Mbit devices which store sprites data :

This was almost unlikely so further investigation was needed.All the sprites generation is accomplished by two ASICs : the ‘053246’ which generates addresses to the MASK ROMs and the ‘053247’ which reads their data.But, unlike other Konami PCBs with similar design, the data lines of the MASK ROMs are not directly tied to the ‘053247’ but some 74LS373 octal d-type latches are in the middle (two of them receive the whole 16-bits of each MASK ROM)

I opted for a ‘shotgun method and replaced all of them.The MASK ROMs check no longer complained and reported all devices as good:

Board completely fixed:

 

 

The second PCB:

The board was failing the POST reporting two bad devices @6G and 7G:

They are respectively the ROM and the RAM of the Z80 audio CPU circuit .

The 1Mbit ROM was socketed so I pulled it and found that one leg was missing:

I rebuilt the leg and dumped it in my programmer but I got inconsistent readings so I replaced the device.In this way the board booted into game but graphics were wrong.The sprites were missing as well as the backgrounds incomplete and with bad colors:

I ruled out the MASK ROMs by performing a check :

At this point there was not much to test, graphics generation is entirely accomplished by custom ASICs.There is one especially whose functions could be related to the issue :

The ‘053251’ is a priority encoder which takes on input different layers of graphics as well as some priority bits and outputs 11 bit of palette index plus two shadow bits (info taken from MAME source)

Probing the output pins with a scope revealed weak signals on most of them:

I removed the IC preparing then the area to accomodate the spare part:

I took it from a donor Premiere Soccer board:

This restored graphics.Board fully working again.

Received from Austria this Devastators PCB (manufactured by Konami in the distant 1988)

The board was constantly watchdogging sign that no valid code was executed :

Probing the HD63C09E main CPU revelead some address lines were dead:

I removed it :

Trying the CPU in another board confirmed it was really bad so I replaced it.The board booted up but it was failing the POST showing two bad devices which at fist glance I could not recognize due the complete lack of graphics :

Thanks to MAME I figured out they were the ones @C11 and C14 :

On PCB they are the two 2k x 8-bit static RAMs (Motorola MCM2016 used) which are part of the palette colors circuit

Probing the one @C14 revealed stuck data signals:

RAM chip failed the out-of-circuit testing:

Now the board booted into game but with severe graphics issue : the backgrounds had wrong colors and sprites were missing.Sound was absent too:

I dumped the four OTP MASK ROMs that store tiles data and found three bad ones :

But this made no real improvements.Board is almost fully populated with Fuitsu TTLs so I went to test them in-circuit with a logic comparator starting from this part of circuit which is involved in color generation  :

I found five bad mutiplexers with floating outputs:

This restored correct backgrounds:

I decided to troubleshoot for first the lack of sound.Found in rapid succession these bad ICs:

• YM2151 (with serial data output pin stuck high)
• sound RAM and ROM
• two 74LS74 @G5-F10

And a 74LS273 @B7 with some floating outputs:

Sound was back :

Now the lack of sprites.MASK ROMs check reported two bad devices @H4-K4 which are the ones  storing sprites data:

This particular PCB uses a bottom ROM board with sprites and audio samples ROMs :

But the top board can also host three 4Mbit devices for the same data so you can get rid of the ROM board :

This is what I opted instead of troubleshooting it since all the splitted ROMs were soldered in.I programmed three 27C400 EPROMs with MAME dumps:

The last issue I had to fix was releated to inputs, P1 UP was not working.Using schematics I quickly found and replaced a 74LS253 @E13 with a bad ouput :

Board finally 100% fixed and another battle won against Evil Konami and Fujitsu with this big booty: