Saturday, 25 July 2020

Pitfall II : The Lost Caverns repair log

Received some days ago for repair a Pitfall II PCB (on Sega System 1 hardware)
Board was in very good condition :

But on power up it gave only a solid black screen and nothing more hence deeper troubleshooting was needed.Probing the main CPU/RAM/ROM circuitry revealed that the lowest address line (A0) was stuck high on the three program ROMs and nearby RAM :

ROMs and RAM are addressed by the main CPU (a custom Z80) not directly but thru some 74LS244.I traced the stuck address line back to an output of the 74LS244 @IC124 which, according to my multimeter,  was shorted to +5V :

I was almost sure the TTL chip was bad (it was from Fujitsu too) hence I removed it but it turned out to be good.So there was something else shorting the address line to VCC preventing the game to boot.I flipped the board over and traced on solder side the shorted address line back from the ROMs/RAM.While I was following the traces I came across this :

A close-up under the microscope :

A VCC pin was bended on the nearby trace of the address line peeling-off the solder mask and causing the short, unbelievable!I promptly straightened the pin and board sprang to life :

I played some games, no further issues were present, board was 100% working.Another successful repair accomplished.

Sunday, 19 July 2020

Sega '315-5242' reproduction

The '315-5242' is a custom IC used on some Sega arcade systems :

  • System C/C2
  • System 18/24/32
  • System X
  • System Y

It's basically a video DAC connected to the color RAMs, which provides analog outputs for driving a monitor.It accomplishes also some special functions like making the video output grayscale instead of color or enabling shadow/highlight effect.It comes under the form of a ceramic module covered by black paint or epoxy :

The 'core' of the '315-5242' is a surface mounted IC (PQFP44 package) located on top of the module, the 'M71064' from OKI manufacturer :

No info or reference was available for this fully digital IC until some time ago when 'furrtek' decapped it and traced schematics from the silicon die revealing the internal structure and figuring out the functions :

Most of the work was done so the step to a real replacement was the easy and logical consequence.And here we are!I designed a proper board based on a +5V CPLD :

Testing on all boards I could try was successful.

Columns II (System C2)

Michael Jackson's MoonWalker (System 18)

Spider-Man (System32)

Thanks again to 'furrtek', without his brilliant work none of this would have been possible.

Saturday, 18 July 2020

Toaplan 'HK-1000' reproduction final revision

As many of you may know I made some time ago a reprodution of the 'HK-1000', a custom IC used on some PCBs manufactured by Toaplan :

Toaplan ‘HK-1000’ reproduction

The replacement was welcomed by arcade enthusiasts and collectors who used it to repair faulty boards but actually it was not a 1:1 replica of the original chip because the handling function of the coin counters/lockouts was missing since I intentionally omitted it, not considering it of primary importance. 

For the uninitiated a coin counter is mechanical style counter that increments every time a coin is inserted, here's picture of a +12VDC powered one:

Lockouts are, instead, a mechanism to prevent users from wasting an inserted coin when the cab is off or not properly operational.

Lately I've been ask more than once to implement in my 'HK-1000' reproduction the circuit needed to add these missing functions.And here we are!!This will be the final revision of the replacement, all functions of  the original chip have been implemented in it.Outwardly it has not changed from the previous ones :

But from now on it will be finally possible to properly use a coin meter in your arcade machine when you play those particular Toaplan boards using my reproduction :  : 

Wednesday, 1 July 2020

Data East 'VSC30' reproduction

Some time ago I had for repair an original Bogey Manor PCB by Technos (but running on Data East hardware).I was able to fix the board but I had to replace the custom IC marked 'VSC30 with a same one taken from a donor board, you can read the whole story in my repair log :

Bogey Manor repair log

After some time I sent the board back to the owner I  thought it would have been a good thing to reproduce this custom IC as well.The part come in a DIP40 .600 mil package :

Its pinout can be found in schematics of different Data east games like Express Raider, Last Mission and Gondomania :

It seems the part is involved in tilemap/background generation performing some kind of clock dividing.With this info and looking at how the same functions were achieved on previous hardware I came to a couple of designs using simple TTL gates ( I even used DIP ones because of low number needed)

I used as testing platform my original Express Raider PCB which carries one 'VSC30' on bottom VIDEO board :

The replacement worked great :

As far as I know the 'VSC30' is present on these PCBs :

  • Bogey Manor
  • Breywood
  • Express Raider
  • Gondomania
  • Last Mission
  • Shoot Out
  • The Real Ghostbusters

And perhaps some others running on similar Data East hardware of same era (see 'dec8.ccp' driver in MAME soruce) so feel free to send your findings.Also, it could be that few boards, like Bogey Manor, make use of extended features of this custom IC so, please, contact me if you are willing to test my replacement on your board.

Monday, 29 June 2020

Toaplan 'GXL-01' reproduction

My reproduction work continues with a new chapter.This time I took into account the custom IC used on some Toaplan/Taito PCBs.Silkscreening under it says 'GXL-01' :

But you can find it with different labels.On Wardner/Pyros it's 'WT1' :

'T.T.1' on Twin Cobra/Kyuukyoku Tiger :

'L-01' on Sky Shark/Hi Sho Zame :

The IC is a DIP42 .600 mil package and it accomplishes graphical functions generating the backgrounds.First of all I identified the pinout and external connections with the board.Then I thought about different designs of the equivalent circuit also studying how same functions were achieved in previous hardware.Lastly I came to a replacement board with same dimensions of original IC :

The "smoke test" was successful, the replacement worked on all boards I could try.

On Wardner/Pyros :

On Twin Cobra/Kyuukyoku Tiger :

Tuesday, 23 June 2020

Toki repair log

Received some days ago a bunch of faulty boards for repair.There was an original Toki in very good condition :

The board booted into game and was playbale with sound too but sprites were wrong :

As always I started my troubleshooting with a visual inspection and noticed that a couple of 74LS273 TTLs were replaced and sockets installed :

The work had not been done properly because some traces were broken and then patched with jumper wire on solder side :

I opted for redoing the job starting to remove the previously installed sockets :

This revealed some pads/traces were destroyed as well as some rivets ripped off from the holes.After checked all connections and rebuilt the traces I reinstalled the ICs on round machine sockets :

This lead to some improvement but the sprites had still wrong colors :

I focused on the reworked area :

Probing around I found that shorting some pins of the 'SIS6091' surface mounted custom ICs restored correct colors.I also figured out that these pins are in common between all four 'SIS6091' (probably a share bus) but I found that one pin was not.I carefully inspected the solder side of the board at the affected area and after some time I noticed a strange break in one trace.The microscope and a check with multimeter confimed that trace was really broken (it seems to me from factory more than accidentally)

I promty patched the trace, this restored corret sprites :

I played some games with no other issue found so I could successfully close this repair.

Monday, 8 June 2020

Namco '07XX' reproduction

Here's another reproduction of a custom IC I made recently.The part in question is used on Namco PCBs and comes in a DIP28 600 mil package with stamped number of  '07XX' (but the 3rd and 4th digits can vary because they actually refers to the lot/date production hence the first two digits really matters) 

Technically speaking it's a clock divider, specifically it that generates the horizontal and vertical video timing signals divided from the master clock.I made my own implementation with simple thru-hole TTL gates :

I succesfully tested my design on a Galaga PCB :

This open the way to a future implementation of this custom IC onto a CPLD or FPGA (already done by someone else though).

Saturday, 6 June 2020

Salamander repair log

Some days ago I received for repair an original Salamander PCB.The two stack set was in good shape in both CPU and VIDEO board :

The board was not booting and kept resetting meaning watchdog was active  :

The video board was marked by owner as good hence I focused on the CPU board./HALT line of 68000 main CPU was stuck low meaning it was stopped :

Board was heavily populated with Fujitsu TTLs and we are all aware about poor reliability of these parts.They usually fail showing floating outptus and this is what I experienced while checking them, my probe was not detecting any logic level on some (all 74LS244, a bad lot perhaps?)

A good rule when you encounter Fujitsu TTLs would be to replace them all because they sooner or later fail.In this case I started to pull out the suspicious ones, three of them were really bad :

After replaced them the board finally got past the POST and entered in game but while testing I noticed that a music track and some sound FXs were missing :

Checking with an audio probe revealed music and sound FXs were fully generated but then they didn't reach the inputs of the amplifer.I scratched my head to figure out this and after some time I remembered that this board has stereo sound and, like in other Konami games, you must close a jumper (the 'CN2' MONO or the 'CN3' STEREO one) in order to hear both channels :

This is what I did giving full glory back to this great game!


You can read more about Konami stereo sound in this post by my mate Yves:

Konami stereo sound 

Monday, 1 June 2020

Namco '6xA1' reproduction

Another reproduction of a custom IC, or better, this is a 3-in-1 replacement of the Namco MCUs marked '60A1', '63A1' and '64A1'.

They are mostly the same part with identical package (DIP40) and pinout but they are not interchangeable.

  • '60A1' found on Pac-Land, Dragon Buster, Sky Kid (all revisions except early one), Sky Kid Deluxe, Alien Sector, Baraduke, Metrocross,  Rolling Thunder and all other System86 games.

  • '63A1' found on first revision of Sky Kid.

  • '64A1' found on all System 1 games (Splatterhouse, Galaga88, Pac-Mania, etc)

Needless to say the original part is very prone to failure (especially the '64A1')

My replacement covers all three MCUs, you can simply select the desidered one via a dipswitch:

Here's testing of the replacement on different boards :

'60A1' on Rolling Thunder :

'63A1' on Sky Kid (first revision) 

'64A1' on Splatterhouse :

A special thank to my friend 'uchopon' for helping me to achieve this result.

Tuesday, 19 May 2020

Konami '502' reproduction

Another custom IC successfully reproduced.It's the one found on Konami PCBs and labeled as '502' in schematics but with scratched-off part name like this manufacturer is used to do.It's a 28 pin 600mil IC :

You can refer to this useful spreadsheet (credits to ‘mattosborn’ on KLOV forums) to check which boards use it :

Konami custom ICs

I made a first replacement using simple TTL gates:

I tested it with on a Track & Fied PCB, it worked fine :

So, happy with result, I ported the design into my own 28 pin CPLD board :

Here's the final successful test after some optimization work  :

Wednesday, 1 April 2020

Demon's World repair log #2

Lately I'm going through all my PCBs I bought as faulty to see what is worth a repair.A Demon's World came up :

Board was in good state and marked with a "OBJECT RAM ERROR" label :

But actually I got only a black screen when I powered it up./RESET and /HALT lines of 68000 main CPU were stuck LOW, no proper transition to HIGH state .

The /RESET circuitry is a typical one built around the 'PST518' voltage detector IC plus all parts needed by its typical application circuit (resistors and a capacitor) :

Like I did for my previous Demon's World repair I went to check in-circuit the ESR of the 10uf 16V electrolytic capacitor (its role is important because when the power supply voltage drops below a certain threshold, it suddenly discharges generating thus the required reset signal).It measured 9.2 Ohm, way beyond its typical value :

The capacitor was clearly bad hence I replaced it with a tantalum one :

Finally I got displayed the 'OBJECT RAM ERROR' reported by the sticker on the PCB :

I experienced this issue many times on other Toaplan PCBs so I knew where to look at.The sprites RAMs are two 6116 (2k x 8-bit) devices located just below the custom ASIC 'FCU-2' which is the sprite generator that addresses the RAMs and processes their data :

 A closer inspection revealed a not perfect soldering of the 'FCU-2' ASIC :

I made a full reflow of it, soldering was now bright as it should be, sign of a good electrical conductivity  :

On next power up I was greeted by this :

Board was fixed and game fully working.Another successful repair accomplished.Let's move now to the next faulty PCB in the pile!