Pang repair log

Going through some PCBs I've bought recently in a lot.

Here's on the bench an original Pang PCB :

 

At closer inspection I noticed the suicide battery was recently replaced and still with charge so I powered the board up.The board booted up and played fine but colors were wrong, screen was blueish (this denotes troubles with red ) :

Upon a quick look at the board I located the palette RAMs, two Sony CK5814 (quite unreliable part in my experience) :

Probing the one @8C revealed all the eight data pins stuck high (while the address lines were toggling)

Sure enough I removed the chip:

And tested it in my Retro Chip Tester (great device, highly recommended!), it failed :

 I fitted a round machined socket and a good RAM chip :

This restored correct colors.Another board 100% fixed.

 

Spider-Man (conversion on Sega System32 hardware) repair log

I got for repair a clean Spider-Man PCB set, actually a conversion on a donor Sega System 32 ROM board (motherboard was from Rad Rally) :

The board booted up and played fine but sound samples (FXs, speeches, etc...) were totally missing :

Looking at hardware the samples are handled by a surface mounted IC marked 'ASSP 5C105' :

The chip is actually a 'RF5C68A'in disguise, a PCM sound generator IC manufactured by Ricoh and used in Fujitsu's FM Towns computer series, along with Sega's System 18 /32 arcade system (sometimes you can find it marked as 'ASSP 5C68A' and Sega '315-5476').Datasheet is available so I looked at  the pinout :

Pin 36 'SDLH' is the serial data output connected to pin 8 of the near Sanyo LC7881 DAC (for digital to analog conversion), probing this pin I found it stuck low all the time :

This was a clear sign the 'ASSP 5C105' was internally fault.Sure enough I removed it with my hot air station :

 

I took a spare from a dead System32 motherboad and installed it :

 This restored all sound samples and fixed board completely.Job done.

Capcom '86S105' reproduction

Thanks to the always brilliant work of 'furrtek' another custom IC has been reproduced and preserved.

It's the Capcom '86S105, the sprites generator used on many PCBs like :

• 1943
• Ashita Tenki ni Naare
• Black Tiger
• Block Block
• Capcom Baseball
• Capcom World
• Dokaben 2
  
• Pang
  
• Poker Ladies
• Quiz Sangokushi
  
• Quiz Tonosama no Yabou
• Super Pang
• Side Arms 

This custom IC comes in a PLCC84 package (with part number sometimes scratched off) and it's a very prone to failure part :

I made different revisions of this reproduction.The first one was based on a bootleg replacement but it was not perfect due to wrong design used by bootlegers :

 The second one was a FPGA prototype based on 'furrtek' work :

The third revision of my reproduction was a single part interfaced to PCB footprint by the use of castellated half holes design :

Here's successful testing on a Pang PCB :

The fourth and latest one , always FPGA and 'furtek' work based, is a two parts reproduction with a top board that plugs on onto a bottom board (soldered on game PCB always by castellated half holes) :

Successful testing on a 1943 : The Battle of Midway PCB :

This latest revision worked fine on all other PCBs I could try (including Pang, Super Pang, Block Block, Side Arms and Poker Ladies) except Black Tiger where glitches are present :

This problem apart, it's nice to have a working replacement of this unreliable custom IC.A huge thanks again goest to 'furrtek' for his invaluable work.

Robocop 2 repair log

 Got for repair this original Robocop 2 PCB by Data East :

Board booted up and plays but sprites were completely absent :

Sprites are generated by the surface mounted custom ASIC marked '52' :

Schematics for this board are not available but I was able to find the pinout of this custom IC on the ones of Desert Assault.There are some input timing signals to the '52' custom IC, when I went to probe them I found pin 98 (labeled 'VCNT' on schematics) stuck high :

 

 I traced the pin back and found on the path a gouged trace on solder side :

I patched the trace with some wire wrapping wire and sprites came back :

Job done, another repair successful accomplished.

1942 repair log

Got for repair an original Capcom 1942.

It's a two PCB stack made of a CPU board

and a VIDEO board

When I applied power the board booted but I could not coin up

Looking at schematics the coin input from the edge connector is routed to pin 14 of a 74LS367 @A6 on CPU board

 

 

Checking it with a logic probe revealed a floated signal (usually inputs are in high state pulled-up by resistors)

Using a multimeter I measured few Ohms of resistance between pin 14 of the 74LS367 and GROUND (whereas usually it should in order of MegaOhms)

I pulled the IC :

 It passed the out-of-circuit testing on the Retro Chip Tester :

 But it failed the one of my ChipMax2 programmer :

I fitted a socket and a good IC :

 This fixed the issue, the game registered again coin input.No other problem was found therefore board 100% working again!

'ZILOG' - a MSX2+/turbo R FDD emulator

'ZILOG' is an internal floppy disk drive emulator I designed for the MSX2+ and turbo R computers ( equipped with built-in FDD).You can read more details about these machines here :

MSX2+

MSX turbo R

Like for 'The Thing' (the FM Towns Marty FDD emulator), 'ZILOG' (which I named like this in honor of the Z80 CPU that equips these home computers) replaces the internal mechanical floppy drive with a board able to emulate the floppy interface reading disk images stored on a USB flash drive :

'ZILOG' emulator is powered by the great Flash Floppy firmware and can be controlled by a small external rotary/OLED board :

 

 

 

Testing of the emulator on my MSX2+ Panasonic FS-A1WSX :

Mitsubishi 'MN50005XTA' and 'M60002-0118P' reproduction

The Mitsubishi 'MN50005XTA' and 'M60002-0118P' are two custom chips found on Tecmo arcade PCBs of mid/late '80 like Rygar/Argos no Senshi, Silk Worm, Gemini Wing, Solomon's Key and few other boards.

The Mitsubishi 'MN50005XTA' is a DIP28 .600mil plastic part :

Whereas the 'M60002-0118P' is DIP42 .600mil (always with plastic case)

I had these two custom ICs in my reproduction TO-DO list since some time, it was only matter to find some bootleg boards to study how the function of these custom ICs were reverse-engineered.

As soon as I found them I extracted the equivalent circuit and then designed the replacements with simple TTL gates :

'MN50005XTA' reproduction :

'M60002-0118P' reproduction :

Both replacements worked at first try when tested on my boards (Rygar and Silk Worm)

 

 

 

 

 

 

 

Since the designs have been validated a CPLD version of the TTL reproductions will follow later.

Altered Beast repair log

Got for repair an original Altered Beast PCB (on Sega System 16B hardware)

Board was in decent condition :

 It was marked with 'GFX issue" and, indeed, at first power up I could see it :

There were missing horizontal lines all over the screen.The fault(s) was located on motherboard since the ROM board was successfully tested with a known good motherboard.

At first glance this issue appeared to be related to backgrounds graphics, indeed someone previously replaced in vain the '315-5197' PGA custom tiles generator and a couple of RAMs connected to it.

I ran a memory test which reported all RAMs as good.But on motherboard there are other memory chips not addressed by main CPU therefore these are not covered by the test, in particular these RAMs are tied to the '315-5196' PGA custom sprites generator.Here is a RAMs chart for reference :

The sprite RAMs ICs were all Sony CXK5814 (2K x 8-bit), I know from my experience that these are quite prone to failure.I started to probe the two @G2 and G3, the signals on most of data pins looked like unhealthy compared to a good one (on the right of picture below)

I pulled both chips :

Actually only the one @G2 failed the out-of-circuit test (the other was good)

I installed a good RAM chip on socket :

This fixed the issue and board completely.Another successful repair.

Irem 'KNA65005-17' reproduction

Another Irem custom IC has been defeated and successfully reproduced.

After the 'KNA70H015(11)' and 'KNA70H016(12)'  

Irem 'KNA70H015(11)' and 'KNA70H016(12)' reproduction

I took into account the 'KNA65005-17', a DIP48 part found on M72 and later hardware like M81, M82, M84 and M85 (although with scratched-off part name)

The functions of this custom IC were reversed by analyzing and comparing the signals between original and bootleg hardware.As always I ended up with a prototype made with TTLs gates:

After some tweaking, it worked fine on all PCBs I could try.

Here's testing on R-Type (M72 hardware) and Dragon Breed (M82 conversion) 

 

Since the design has been validated a CPLD version of this TTL reproduction will follow later.And a full re-engineering of the entire Irem PCB is more possible now.