Bootlegs are worth a repair too especially when they are classic games like this copy of Slap Fight :

PCB was dead, no sign of activity.At a closer inspection, the main crystal which generates the master clock and the 74S04 inverter were missing from video board:

There are several bootlegs of Slap Fight and each one has a different layout and design but looking at online pictures mine was identical to original board.MAME source told me the exact value of the missing quartz :

/* basic machine hardware */
	MCFG_CPU_ADD("maincpu",Z80, XTAL_36MHz/6) // 6MHz
	 

I was not able to find a 36MHz quartz among my spares (it’s not a common value) so had to order it.I got it in few days:

With the correct clock the PCB sprang to life :

Game was playable but sprites and sound were missing at all:

Sprites data are stored in four 27256 located on CPU board.Data lines of these EPROMs are tied to the inputs of four 74LS166 8-bit shift registers:

Probing the components of this circuit I found that the shift load input (pin 15) in common of the four 74LS166 was stuck high.I traced it back to an output (pin 8) of a 74LS32 whose input pin 9 was stuck too.Tracing this back lead me to the cause of the issue:

 Patching the severed trace restored the sprites:

Now the lack of sound.”Listening” with my audio probe the output (pin 5) of the uPC1182 amplifier revealed that sound was present but for some reason it didn’t reach pin 10 partside of the JAMMA edge.A closer inspection revealed a missing electrolytic capacitor (used to filter the output of the amplifier)

I fitted a 1000 uF capacitor as suggested by the typical application circuit of the uPC1182 amplifier:

Another valuable bootleg fixed!

Picked up this faulty Knuckle Bash PCB on Ebay:

The graphics were totally wrong:

The board use a custom marked ‘GP9001’ @U13 which acts like a GPU generating all the parts of graphics (and video timing signals too)

Package is QFP 208 pin so it’s a common issue to find lifted pins of this kind of ICs after some time.Under a microscope it seemed OK :

But actually a lot pins were detached from pad when I probed them with a needle.Reflowing the IC restored all graphics but colors were wrong:

The palette is generated by two 2K x 8-bit static RAMs and all the 16 bits of color data are latched by two 74HCT273:

Probing the two RAMs revealed a stuck bit (D7, pin 17) on the one @U7 (comparison with an healthy signal on the left of the below picture)

This data bit is latched by pin 13 of the 74HCT273 @U9, piggybacking it restored correct colors:

Chip failed in some gates when tested out-of-circuit:

Job done.

Got a Rabio Lepus (japanese release of Rabbit Punch, a cute shoot ’em up released by V-System Co. in 1987) in lot of faulty PCBs.Hardware is made of a CPU and a VIDEO board:

Board had a graphical issue, sprites were splitted and misplaced:

 

Usually this kind of problem indicates an addressing fault so wrong address are generated hence wrong data are read/written.Sprites data are stored in six devices (four 27C256 OTP ROMs and two 28 pin 1Mbit MASK ROMs)

Reading them gave good dumps but the logic probe revealed that address line A7 was stuck high in all of them (address bus is shared)

Tracing this address line back lead me to an output (pin 19) of a 74LS273 @8A, the corresponding input (pin 18) was floating.All the other outputs were active (generating the other address lines for sprite ROMs) and inputs were toggling too.Each input is connected to a data line of a near 2k x 8-bit static RAM @10A so following this logical arrangement I figured out that pin 9 (data line D0) of the RAM should have tied to pin 18 of the 74LS273 but it was not so :

I restored the connection with a jumper wire and this fixed the issue:

As I expected, doing a visual inspection on solderside revealed a severed trace:

End of job.