As every arcade fan/collector probably knows, Konami makes great things but… sometimes weird.Technically speaking every their board (starting from old ones like Gyruss to latest ones ) has its own hardware layout which involves the use of  custom ASICs (Application Specific Integrated Circuit).These ICs are used to semplify the layout and prevent board bootleging.They can have most disparate shape and package and do various functions (sound, graphics, I/O and so on).In particular the Konami ones are really well made but often prone to failure.Some example of their customs :

 

 

The one I faced during this double repair log is the one marked ‘054573″ .It’s used on hardware of mid ’90s, it has SIL 15PIN package and it’s essentially a color DAC, it converts digital color signal into an analog one outputting it directly on JAMMA connector (so one IC is used for each RGB colors) :

Here is the IC ‘naked” , picture kindly provided by IronGiant:

It’s a quite fragile part, indeed while I was testing a Gaiapolis PCB, I switch it OFF/ON and I got this (picture on the right) :

Green color was suddenly missing.Once identifying the ‘054573’ DAC responsible of this color I compared with a logic probe its pins to  the ones of the other two DACs and they had exactly the same activity except for PIN2 which is , indeed, the output connected to PIN12 solderside of the JAMMA edge.Observed with an analog scope confirmed the output was silent compared to an healthy one of a good DAC:

Replaced it and the green color was restored :

The second PCB, a Metamorphic Force one which also use this kind of DAC, was perfectly working but someone made an awful job on the solderside, “piggybacking” a good ‘054573’ DAC on the faulty one (the green color one also here) and fixing it with hot glue:

Since the eye wants its part, I decided to revert this orrible hack restoring the original status and testing successfully the board :

Konami does great but weird things always, remember! 🙂

Maybe the name “Kyohkoh-Toppa” doesn’t say anything to most of arcade collectors/fans but if ,instead,I said “BreakThru” many of you will remember a game whose goal is to drive a dune buggy to “breakthru” the enemy lines of five different areas.So, for those who still have not understood, “Kyohkoh-Toppa” is the japanese version of BreakThru released  by Data East in February 1986 for the eastern market.

Here is the PCB:

Board booted fine but it had a graphic issue since sprites were missing some lines:

I noticed that problem went away if I flexed the board so there was some poor contact somewhere.In this kind of hardware all the graphics is generated in the bottom board so I reached it and found this:

Someone (certainly not manufacturer) replaced two TMM2018 300-MIL SRAMs with two 600-MIL equivalent ones adapting them in narrow sockets!For a better understanding of the package dimensions:

Not a neat job, for sure…Anyway, I could pinpoint sprites issue in the SRAM @11E (the one of the left picture) since problem was cleared when I pressed it.As you can see from picture they used some jumper wires on part side while patched some broken tracks on solder side :

So, I decided to remove this hack, check that all the RAM connections were fine and use a proper 300-MIL chip in a new socket :

In this way sprites were stably restored and board 100% fixed.

Got from a friend this The Battle-Road PCB :

Honestly I never heard of this game before.Anyway, it runs on IREM M62 hardware which is the same one of Kung-Fu Master and other games.

My friend said that sprites was completely missing and he was right but comparing what I got with MAME emulation  also text/characters colors was incorrect :

So time to investigate and start to study the sprites circuitry which is located on bottom board.Data from the six sprite ROMs are multiplexed by a custom maked “NANAO KNA6034201”, its outputs go to some TTLs (74LS157 and 74LS374) until they hit the RAMs I/O pins (four 2149 DRAMs) which were all stuck HIGH.Probing the custom revealed it did its job, all outputs were correctly toggling as well as piggybacking the four DRAMs didn’t change anything so fault was in the middle.When I probed the 74LS157 @H3 with my HP10529A logic comparator I got this:

So troubles on all its outputs that were confirmed from my logic proble which reported them as stuck HIGH.Obviously IC failed when tested out-of-circuit:

Fitted a new 74LS157 restored sprites but as I said early text/character colors were wrong compared to MAME emulation correct ones.Text colors BPROM was dumped fine so I went to dump the two ROMs containing this part of graphics and I found they were swapped into their respective sockets (on the left the wrong positioning, on the right the correct one) :

This fixed board completely.End of job.