2nd January 2026 A local friend asked me to look at a Sega Hang-On board. Well actually 2 1/2 boardsets. A local guy contacted him after seeing his Afterburner DX moving cab repair vid and asked if he would help restore his Hang-On ride-on cab that he picked up for 500 bucks. Unfortunately that guy didn’t do any research about the cab and knows very little about repairing and maintaining complex old electronics, basically he just wanted it because it looks ‘cool’. The ride-on cab does look cool but it’s very difficult to play because you have to swing the whole bike over from side to side. Even Sega knew that because they made the game a lot easier than the stand-up cab, with no sharp corners, no quick left-right corners and less obstacles on the side of the road. My local friend has a bit of a soft spot for helping people (and possibly him and the other guy also have a soft spot on the top of their heads too LOL!) and said he would try to fix it. He did all the other stuff and I fixed the PCBs. Rats had got inside and completely trashed everything including the PCBs. They were so bad that nothing on any of the boards was useful for spare parts. It didn’t help that they stupidly powered up the cab ‘hoping’ that it would just work LOL! In reality any chip that may have been good blew up instantly when power was applied. There’s a youtube vid about the repair here with the cabinet restoration. This is how the cab looked at the start and the boards that were in the cab as well as the good condition but dead or faulty boards that were purchased at a cost of another 500 bucks per boardset. Yeah it’s getting expensive already and nothing has been done yet LOL!

The last pic is the thing that makes this repair possible.... the tray out of the cabinet with the full (repaired) wiring harness, transformer, power supply and a complete boardset. I wired up some of it to a JAMMA fingerboard so I can plug it into my test rig including some pots for steering, gas and brake so the game can (eventually) be fully tested. The amp was also trashed and there’s a separate power amp sitting on the floor with a couple of Sony speakers from my 6+1 surround sound AV set-up plugged into it. The total number of items to repair is.... 3x CPU boards 2x Video boards 3x ROM boards 2x Sound boards These numbers don’t include the rusted rotten boards, they are only good for the bin.

This repair will be presented here a bit different to all the other repairs. This is mainly because the job took several weeks and I really don’t want to put up 1000 pics where I pull chips and show them failing in my IC tester. Plus there’s just way too much stuff going on with these boards. I did so much work that I can’t remember most of it, it’s just a blur or a nightmare that I’ve subconsciously blocked out of my memory LOL! Instead, I’ll try to find something interesting or noteworthy to talk about. These notes will be jumping around from board to board as there is no way to know exactly which board I was working on at that time. They all had multiple faults and often changing a bad chip had no effect on anything because there were many more bad chips in the same circuit. They were fixed and I moved onto the next one quickly. Later some of them failed again so were fixed again. Let’s start with a pic of the aftermath....

If you count the chips you’ll get a number starting with 1 and ending with two zeros. Yes that’s right, exactly 100 bad chips! These boards are almost exclusively using Fujitsu logic chips so the bulk of the repair is mainly just pulling every Fujitsu chip and replacing it because 99% of them are dead.

I started by pulling and testing all the ROMs and they all verified good against MAME ROM archives. The next step was to check the 68000 CPUs using one of my Amiga 500’s which has a socketed 68000 CPU. Two 68000’s were bad and were replaced with 12MHz 68000’s.

There was one board (shown kind of working in the youtube vid) that runs the title screen attract over and over. It has a credit auto-applied at power-on but start doesn’t work. At first I thought maybe the inputs were bad but it turned out to be a lot more convoluted. The first 68000 was working but the actual problem was the 2nd 68000 was doing nothing because the reset was low. The reset pin connects to a nearby 74LS02. To figure out any more I needed to check the schematic, which is unfortunately the usual poor quality 1-bit b/w low-resolution garbage scanned with a potato by some random loser. The 2nd CPU reset comes out of the previously mentioned 74LS02 on pin 4. The inputs are pins 5 and 6 which go to the edge of the page and the labels are totally unreadable on the schematic! I used my multimeter to trace all over the board and found that one connects to pin 39 of the 8255 at 4C labelled SRES.

So the 2nd CPU reset is controlled by the first 68000 CPU via one of the 8255 PIAs. The data bus was active and so were many other pins but a lot of the output ports were not active. Not being familiar with the 8255 it took some time before I eventually noticed the 8255 wasn’t enabled. Pin 6 Chip Select should be low and it was high so the chip was disabled and none of the output ports were active. The schematic shows the 8255 CS pin comes from a 74LS138 at 4D (IC31) pin 9 and pin 15 enables the other 8255. The inputs were active but pins 9 and 15 were only high. I pulled the chip and it failed.

After replacing the chip the board still didn’t have a reset on the 2nd 68000! Of course that 74LS02 was a Fujitsu and was dead too. After replacing the LS02 the board mostly worked but with missing background graphics...

This was likely a fault on the video board and caused by more bad Fujitsu logic chips. More on fixing this fault later.

Here’s one of the faults that developed later as one of the repaired boards was running on test for a while....

This is supposed to be the title screen. Since sprites are messed up (ignoring the bad colors) I checked the sprite RAM. It’s a strange and uncommon RAM which of course had failed....

Replacing that chip brought back the graphics but the colors were completely whacked....

The schematic shows the section on the lower right side of the video board has something to do with the colors so I probed the chips and many were dead (Fujitsu chips again of course). I piggybacked a few chips and the colors were vastly improved so I pulled and tested all of them. All failed and were replaced.

This is another strange fault that happened later....

There are lines through the sprites. The colors are bad too. The line-through sprites issue was caused by a bad 74LS244 near a bank of 2148 ceramic RAMs....

At one point one of the CPU boards just died for no reason. I probed the main 68000 and it was completely dead, no clock. The schematic shows the clock goes through an uncommon logic chip 74S112. These are also used on Sega System 32 in the 74F variety. I didn’t have a spare so I pressed my luck and took the chip from the rusted junk board and it tested good in my chip tester so it went into the board and that fixed that problem.

Here’s another strange color fault. This was initially caused by a dead 74LS32 logic chip. It can be seen in the color section of the schematic just below. It’s the chip at location 2C and controls the write-enable pin on both color RAMs. The colors were initially showing bad (1st pic below) but the fault got worse over time (2nd pic below). The LS32 failed high on pin 8 and one RAM was stuck in read mode so the color data was not being written to that color RAM. In the RAM test IC2 comes up bad which is the color RAM that holds the red data, hence why only blue and green shows....

The color RAM is just to the right of that logic chip. With the 74LS32 changed the RAM at IC2 now tests ok in the memory test screen. The colors were way better but something wasn’t right. I went into the test mode and checked the color test pattern....

There’s an additional in-game graphic glitch shown but ignore it... I’ll show fixing that later. RGB is present but the color gradient on green is wrong. The colors go through some kind of DAC circuit (resistor ladder or whatever) that is likely to be 4-bits or 5-bits per pixel or similar. I’ve seen this issue before on other boards and it means some ‘bits’ are missing in the color output circuit causing wrong colors on the color bars.

This is the color output section of the video board. The data comes in through 2x 74LS245, through 2x 2016 2kB SRAMs then through 2x 74HC273. Piggybacking the 74HC273 at 4B fixed the colors so that chip was replaced.

The missing background mountains issue mentioned above was caused by a bad 74LS253 on the video board at 5E or 5F (I don’t remember which one was bad but both were replaced). That chip connects to 2x 74LS374, some RAM (2016 & 2064) and then to the custom Sega graphics chips. The 253’s create part of the graphic tile ROMs address bus that eventually comes out of 2x 74LS244 chips at 7H and 8H.

Here’s another wacky fault on the very first video board I tested after (probably) replacing a bunch of bad Fujitsu chips. Also shown is exactly the same fault but on other video boards including the CPU board that mostly worked but had a credit at power-on...

Most of the graphics are missing. This turned out to be a very tricky fault to find. After going through and replacing all the Fujitsu chips the fault remained. It turned out to be a bad Sega custom 315-5025. The schematic shows how these chips connect and they are all in the graphics circuit connected to EPROMs via some other logic chips.

This chip is just a bunch of logic crammed into one chip. There is no real reason to do that other than maybe to save board space or as a protection device. Hang-On has several of these chips on the board and if any of those chips fail most of the graphics will not be shown. They seem to be isolated to the foreground and background graphics (tiles) because in one pic the Hang-On title screen sprites can be seen through the jailbars. I managed to find the bad chip by shorting out pins on the chips until some things appeared on screen. The outputs are pins 13 and 16 and this particular chip had a dead pin 16. So for example when pin 15 or 17 is shorted to 16 some things show on screen. Initially I pulled a chip off the junk board and piggybacked it but it made no difference. At this time I had a mostly working video board so I pulled one of the chips and put it into a socket then I pulled all the 315-5025 chips off the junk board and tested each one. Only ONE was good! I plugged that chip into the faulty board and it brought back all the graphics. The high failure rate of these chips suggests they were made by Fujitsu. These same chips are also used on Sega System 16A, Wonderboy, Chop Lifter and other Sega boards from that era. Of course they were no match for bootleggers who figured it out and made the same circuit from some common logic chips. Here’s a pic showing one of the little bootleg plug-in boards from a bootleg Wonderboy plugged into the Hang-On 315-5025 socket perched on a tower-of-sockets to get it to fit.

This board holds 3 separate 315-5025 circuits. The game showed correct graphics so the bootleg circuit is confirmed good. That means there are secrets that can be uncovered. Here’s the full schematic for the 315-5025 IC....

Note this schematic is a full reverse of the bootleg plug-in board which contains 3 of the custom chip circuits and is presented as-is. The circuits appear to have been optimised for either ease of routing or to minimize IC numbers, or perhaps both. As such the circuits are not completely separate. The 2nd sheet of the schematic contains a reduced version using 5x surface-mounted logic chips that is designed to replace 1x 315-5025 IC. I have no plans to make a 315-5025 reproduction or anything similar. Someone who knows what they are doing could easily create a PCB for one chip and use it to repair a Sega PCB that uses these chips. These same chips are also present on Wonderboy, Up’n Down, Chop Lifter and others (System 1 and System 2), Space Harrier, Enduro Racer and on System 16A where they are 20 pin chips. My guess is the 18 pin chips cost more to produce so Sega went with a common 20 pin DIP chip and because of that those 20 pin versions had a different IC number 315-5155 and the two pins on the far right are just grounded. You can see that on the Sega System 16A schematics. It’s almost exactly the same circuit that creates some ‘FFx’ and ‘SAx’ signals....

As I needed at least one more chip a junk Shinobi waiting to be repaired (owned by the same friend who was helping with the cab repair) was sacrificed and one of the 315-5155 chips was used. To do that just insert the chip with pin 1 in the socket and pin 10 on the right side hanging out of the socket and it works fine.

Here’s another random issue that I caught early on when pulling all the 74LS153 in the lower right color section of one of the video boards....

LOL! Check IC27. Looks like the previous loser thought Sega didn’t use enough solder to hold the chip in place so he added some more LOL! All those chips in that area were dead and were replaced anyway.

Here’s a look at the set-up while working on one of the video boards....

The top CPU board must be plugged in for the game to work but conveniently it can be flipped up to gain access to the video board. This is done by unplugging the sound board and control harness. The game will boot fine without the controls or sound board. To work on the ROM board do the same thing, just unplug and remove the sound board.

That small graphic glitch on the game screen previously mentioned was a result of bad RAM on the video board. Fortunately the board has a decent RAM check and some RAMs came up as bad. They were replaced and that fixed the in-game graphics glitches...

One video board had missing composite sync and the whole screen was a wiggly mess. This was earlier when this particular video board had bad colors. Sync is shown clearly on the schematic but it ended up being a tricky fault to fix....

HSync comes from two PALs (not sure why it comes from two but it does!) and vsync comes from another PAL. The signals were present on the 74LS86 inputs at 8S (IC115) but the output was dead. The LS86 was so badly weathered that there was no printing on top of the chip! Replacing the 74LS86 with the one shown in the pic below didn’t fully fix the composite sync.

It was kind of working but vertically unstable (I forgot to take a pic of it). The schematic above shows another two chips involved with vsync generation, 74LS669’s at 7K and 7L. However because the hsync and vsync signals were present on my logic probe it took a while to figure out what was wrong. I measured hsync and it was close to 15.7kHz but vsync was moving erratically and my frequency counter couldn’t get a lock on the frequency. I swapped the PAL at 7M but it didn’t change anything. Going back further there is a bus named V0..7 coming out of 2x 74LS669 logic chips. I probed the outputs of the chip at 7L and saw that pin 12 was not active (only high). I pulled the chip and it tested bad. After replacing the LS669 the syncs were fully restored.

This is another strange fault that came up later....

It’s not obvious from the pic but the red signs in the background were missing pieces and parts of them changed randomly. On the same schematic page above (Video BD 1/7) there are three more 74LS669 logic chips at 8R, 9R and 10R which have their inputs connected to the other 74LS669 logic chips previously mentioned. The outputs from 8R, 9R and 10R go to a couple of 74LS373 chips at 7N and 7P. The outputs from the 373’s go directly to the graphics chips. The 373 at 7N pins 2, 5 and 6 were dead. Replacing that chip brought back correct backgrounds in the specific area where the signs show. I didn’t figure that out at the time. This was just another dead Fujitsu chip that was replaced as part of the repair and that very tricky fault fixed itself by magic LOL!

As far as sound goes there were no major issues other than a bunch of dead Fujitsu logic chips and one bad TL074 op-amp. The Hang-On sound board is the same basic specification as used on After Burner but using all common off-the-shelf parts so it can be fixed no matter what goes wrong with it. Unfortunately the next generation board used by Afterburner, Super Hang-On and Outrun fully integrates the sound hardware into a single custom chip so when that fails it’s game over ;-)

The ROM boards were all ok as-is and didn’t need anything doing other than removing and cleaning all the ROM legs and putting them back.

That’s probably the end of the repair log. There were many more bad chips but I’m not going to go over everything. All I can say is if you encounter a board with lots of Fujitsu logic chips replace all of them and maybe the board will be fully working or at least in a better state, then it will make it a lot easier to fully fix it.

I’ll leave you with this final pic....

Be sure to put heatsinks on the custom graphics chips because they all run very hot and if you are very lucky the board will last many years. In this particular case there is no room for these higher heatsinks so I replaced the board spacers with longer spacers and longer screws to raise the top board higher.

28th December 2025 Here’s three more MAME layouts just completed. Update 19th Jan 2026: These are the final layouts that I will be doing.

Akai S3000/S3000i. The version in MAME has a couple of minor issues that I fixed later but has not been updated, and possibly won’t be updated now.

Akai CD3000i

Akai S3000XL

Of course this is just a whole lotta Akai cow-milking LOL! A surprising large number of users of these devices don’t know that most of these Akai samplers have identical or near identical capabilities to previous or later models but in a different looking box. Some of these share the exact same chips even down to the same analog output path and people still claim that ‘model x’ sounds better LOL!

Also, unfortunately most, if not all of the floppy disks that came from the factory with these units are not dumped and in great risk of being lost. If you have access to the original floppy disks that came with any Akai drum machine or sampler dump them now while the disks are (possibly) still good because in some more years the disks will die and the devices will be unusable.

Anyway, this completes the layouts for all 5 devices in the source file akai/s3000.cpp so it’s time to move onto the next target.

20th December 2025 Here’s the next MAME layout just completed....

Akai CD3000XL

16th December 2025 Here’s a couple of repairs I completed recently.

DoDonPachi The game works but the owner says there’s some audio distortion....

Personally I’m not sure I hear anything unusual LOL! The game is VERY busy with a lot of overlapping noises so it’s really hard to tell exactly what is not correct, if anything. I compared it with MAME and *maybe* there’s a bit of distortion but it’s very minor. If this was mine I would just leave it alone until it properly breaks. Actually, if this was mine I’d sell it for $$$$ since it’s just a simple joystick/button game and I would play the game in MAME where there are no board faults LOL! Anyway, this is the audio section on the PCB....

The audio is working so the digital side must be good and it could only be on the analog output side. There could be 6 things that can cause this problem, bad ROM, bad DAC, bad opamps (2 different chips), bad caps and bad main power AMP. The caps appear to have already been swapped as there are tell-tale signs of someone using a powered desoldering tool who only had half an idea how to use it correctly and burned the PCB in a few places LOL!

I pulled the caps and they tested good with in-spec capacitance and ESR so I put them back. If the ROMs were bad the sound would be a lot worse and if the main AMP was bad I think it would also be a lot worse. So it can only be the DAC or op amps. I swapped the DAC first then re-tested.....

That didn’t seem to make any change but it was already working fine IMO anyway. I just swapped the other two op-amps as well.....

I re-tested and *maybe* it’s better, maybe it’s the same. I ran the game next to a PC running the game in MAME and it sounds the same as far as I can see (or hear). There is some speech at the start ‘ATLUS’ and ‘PRESENTED BY CAVE’. On the S there’s a tiny bit of distortion but this is not a fault, that’s just very low sample-rate pcm samples. Maybe the guy thought that was a fault heheh! I have swapped every possible part that could cause any output sound issue so I’m calling it fixed.

Afterburner II Another Sega Afterburner II! The owner says the board was working and just sitting in storage and now doesn’t work at all. I tested it quickly and it only shows a black screen. BUT, if I set the DIPs to the deluxe moving cabinet the game shows testing the motors and then just shows a black screen, so it’s not completely dead.

The owner had done some preliminary checking/diagnostics and the board actually had the Afterburner test program ROMs installed (which I believe is not public/hoarded... more on that later LOL!). It comes up with this screen. Note this test program is running in MAME since it’s clearer to show it in emulation and I didn’t take a pic of the real thing at that time....

IC41 comes up as bad. That’s REALLY BAD because that’s a custom chip. The other ERR is not actually an issue. This error shows when the sub program CPU ROMs have not been removed. With those removed I now get this....

Now it shows ROAD LSB0 ERROR but on the next reboot it flags LSB1 as bad. Then goes back to LSB0 on next reboot. The ROAD LSB0/LSB1 is the RAM at IC38 and MSB0/MSB1 is the RAM at IC39, both just near the 2nd 68000. Funnily IC39 is a Toshiba TMM2063 (soldered in) which is known dog shit but it’s actually working hehe! IC38 is in a socket and with IC38 removed the LSB0/LSB1 error was different so that’s the one. I swapped it and that RAM now passes. I tested the old RAM (which is brand new) and strangely it passes all the tests in my chip tester but my tester can’t test stuff at the correct speed. It may be a slower RAM re-marked as a faster one (i.e. Ali-Express Chinese special LOL!) or just tired or something and the board needs faster RAMs or something. Anyway that RAM is in the bin now. So now I am stuck with this IC41 error. I pulled up the schematics and there are a bunch of logic chips connected to IC41. I pulled and tested all of them (many hours work) but they all tested ok. I also used the schematic to trace every pin on IC41 from the chip to the end point and everything is joined. This custom has SIX parallel pipeline 16-bit data buses (96 pins just for that) that all join to those same tested-good RAMs near the 2nd 68000 CPU and a few logic chips. I verified continuity on all the custom pins. There are some control signals too and most are missing as they come from a few logic chips but I’ve already pulled and tested all those chips. The next step is to swap IC41 but where do I get another chip? Well there is actually a 2nd chip on the board at IC108 and that one passes. The theory is to remove IC108 and transplant it to IC41, run the test and see what happens. So I did.....

Hmmm. Now IC41 passes and the test program gets stuck on IC108... which of course is missing. So it is confirmed IC41 was bad and the owner will need to source another chip from somewhere or that’s the end of this repair. So a few weeks passed and eventually another junk board became available. I took the chip off the board and put it onto IC41, powered on and got this....

Everything passes. Whoa! I put the game ROMs in and re-tested....

The game is working!! Unfortunately there are a couple of issues. One is minor and the other is major. The minor error is a small graphics glitch on the title screen graphics, especially on the AFTER BURNER logo, possibly just bad mask ROMs. The major one is there’s no sound and just some occasional screeches. Checking the schematic and probing the digital audio output on the custom PCM sound chip shows there’s nothing coming out. I checked some parts and the Z80 D0,D1,D2 are only low all the time. D3-D7 are active and so is the RAM CE/OE and WE. I pulled the ROMs and compared the CRC32’s to the MAME ROMset, all good. I pulled the Z80 RAM, tested good but since it’s a TMM2015 I swapped it with another chip. No change, no sound. Schematic shows D0-D2 goes to Z80, Z80 EPROM, Z80 RAM, YM2151 and custom sound chip. Now I could mess around for some hours or days swapping other stuff and hoping that it’s not the custom sound chip. But based on years of experience and seeing SO many dead Sega sound chips on Outrun, Super Hang-On and Afterburner I decided to stop f#%king around and I just went straight for the Achilles heel of the board, the custom sound chip. The previous one looks like it had been changed and was a bit of a rough soldering job, although looks like it was good enough, no shorts etc. First I removed it and tested the board. No Z80 activity at all. The schematics show the custom chip controls the Z80 and provides clock, reset, IRQ, BUSRQ, address and data bus, basically everything! I was hoping the YM2151 would be active and do something but no, the same clock and reset also go to the YM, and of course the Z80 has no clock or reset so it couldn’t control the YM anyway. So I swapped the PCM sound chip over and sound FULLY WORKS !!!!!!!!!!!! Hehehe!!!!! I did a quick controls test (all pots and buttons) and everything is working.

Now onto the minor issues. 1. In-game the horizon has a bit of strange and very minor flickering occasionally. 2. The title AFTER BURNER II graphic has some minor lines on the edges and minor flickering now and then.

These two issues are probably related and certainly the title screen graphic is sprites. I pulled, read and compared the CRC32 of all the graphics ROMs and they are good. I strongly suspect it’s one of the RAMs on the bottom right corner in the large bank of sprite RAMs. One RAM is in a socket and has been changed, all the others are untouched. I piggybacked all those RAMs and one affected the glitch so I pulled it but it tested good. I replaced it with a new 62256 RAM and it was way worse! Then I realised what’s going on... this is HM65256... it’s PSEUDO STATIC RAM. It’s the same bullshit used on Point Blank. I don’t have any in stock in the thin or even wide DIP versions I only have SMD types. I replaced the RAM I removed and put in a socket. I pulled the existing other wrong 62256 RAM in the socket and put in a CY7C199 and that almost fixed it but there is a distinct square cross-hatch pattern across the sprites now. It’s barely noticeable but I can see it hehe!

My guess is the glitch is caused by the RAM that has been replaced by someone previously in a narrow socket at position IC72. I looked on the junk parts board and there are tons of them of course, but all are the thicker wide type NEC D43832 which is the type printed on the PCB under the chip. It’s the same pseudo static RAM bullshit but wider. So I pulled the existing narrow socket and replaced with the wider DIP type socket and a NEC D43832 RAM. That fixed it!!

Ok so you’re still wondering about the Afterburner / Sega X-Board test program, right? I put it online HERE. Spread it far and wide. Happy repairing :-)

14th December 2025 A local friend asked me to install a Psikyo SH2 multi-game kit he recently picked up. His donor board is the less common donor board, a Strikers 1945-II which isn’t documented very well and is long-winded and convoluted in some areas where it doesn’t need to be and it’s also not in English so google translate butchers it to incoherent nonsense in some other areas. So here we go with the Guru version in 15 easy steps. Total install time is about 1 hour...

1. Remove the solder from all the holes on both expansion connector locations and solder in the connectors. Be sure you get them the correct way around by first plugging them into the top multi-board and sit it on top, hold it there and then solder a few pins then re-check alignment and then solder the remaining pins.

2. Remove the 74ACT139 logic chip at U21 and solder on 4x 15cm wires onto pads 4, 6, 9 and 10 as shown....

3. Lift pin 6 of the 74ACT138 at U22 and solder it to pin 5....

4. Lift pin 11 of the 74ACT32 at U39 (just leave it floating in the air)....

5. Solder pins 19 and 20 of the 74HCT244 at U45 together (no need to lift up the pin)....

6. Solder the supplied 74HCT244 in place at U44, clear the solder holes on RC2, RC3 and the EXT connector holes and solder in the 10 pin connector.....

7. On the supplied resistor packs bend up legs 3, 5, 7 and 9 and solder them in as shown....

8. On the bottom side solder the supplied surface-mounted resistors across pins 2-3, 4-5, 6-7 and 8-9 as shown....

9. Lift pin 12 of the ROM at U32, slide the audio fix PCB under it from the right side and solder the 12 legs to the audio fix PCB as shown....

(Hint to the designer: This would fit better and be easier to solder if the pads on the board were rounded like half a hole i.e. castellated....)

10. Near the SH2 CPU are some unpopulated 40 pin connector locations. Solder wires 15cm long into the hole at CN3 position 38 and CN5 hole positions 3, 4, 7 and 40....

11. (This is the most tricky part of this mod)... Lift the SH2 CPU pin 99 (which was tied to VCC) and solder a wire to it then solder that wire to CN5 pin 40 (yes the same place where a previous wire was soldered). Resist the temptation to clean up the area with IPA because you’ll most likely break the leg off. Just leave it alone! Alternatively you could *carefully* cut the trace joining the leg to VCC and solder the wire to the leg without having to lift it....

12. Remove all the surface-mounted mask ROMs (this is far quicker for me personally, around 90 seconds total, than pulling up pin 12 of each of the 8 maskROMs and tieing them high through a resistor to pin 23 of the same chip)....

13. Make sure all the loose wires near the CPU are pulled outwards towards the edge of the board and put the multi-board near the CPU with the GAL header on the multi-board close by then solder the loose wires into the holes like this.... CN3 pin 38 to GAL 6 CN5 pin 3 to GAL 7 CN5 pin 4 to GAL 8 CN5 pin 7 to GAL 9 U21 pin 4 to GAL 13 U21 pin 6 to GAL 14 U21 pin 10 to GAL 15 U21 pin 9 to GAL 16 CN5 pin 40 to GAL 19

14. Plug the multi-board into the expansion connectors and install some kind of stand-off so the board doesn’t sag. I used a PCB foot mounted upside-down and screwed in from the bottom which fits perfect hehe!

15. Power on and play the games :-)

If you’re thinking of getting one of these be sure you forget about eating food or paying bills for a few weeks ($$$$) LOL! Better to just set up a MAME PC and play the games for free minus any board faults ;-)

13th December 2025 A local friend asked me to look at a Point Blank gun I/O board. The issue is the gun solenoid does not trigger on the player 2 gun but works on the player 1 gun. I later found out the main board and gun board came from ebay so you already know what’s going to be discovered next ;-)

The issue is probably because of this mess....

I removed the chip and cowboy hack-job wires and it’s pretty bad underneath LOL!

There are 4 missing pads where some random cowboy loser with no brain tore up the chip without having the required skill level or the required equipment. He clearly knew he screwed it up and made some sort of half-assed attempt to fix it but it was never going to be working since the cowboy had already shown he wasn’t skilled enough to remove the chip correctly so there’s no way he could fix this problem correctly. So, let’s count the missing pads and the wires... hold on a sec, there are 4 missing pads but only 3 wires? Eh? That would make sense as this cowboy was not only missing the required repair skills but also missing basic arithmetic skills and couldn’t count above 3 LOL! I first checked the logic chip but it tested ok. This chip package is known as SOIC14 150mil, so the plastic package is 150mil wide or 0.150" or around 4mm. The PCB pads are designed to take a 200mil or 5mm wide chip. I have plenty of spares here so I’ll replace it with the correct chip. But first the missing pads and traces have to be reconstructed....

With that done I replaced the chip and soldered it in place. I later removed one of the wires and replaced it with a longer wire joining from the origin to the chip leg, hence why there is a tiny black wire there....

Another thing to check with this era of Namco boards is those white or brown muRata ceramic EMI filters. They are often cracked in half or not soldered to the PCB properly. In this case it was ok so I moved on.

I have several lightguns and one is already wired up to work with Point Blank so I plugged in the gun and tested the game. The solenoid did not fire but this is because I don’t have the 24V wired up. I generally don’t connect 24V to avoid any major blowing-up issues, and in reality the kickback is fun for about 5 minutes but then gets annoying and causes aiming issues due to distraction and the extra unwanted movements. As such I’ve never looked into wiring up the gun fully but in this case I needed to test it properly. I wired up the 24V according to the bare-bones wiring diagram in the Point Blank manual but still the gun solenoid was not triggered. The problem with the manual wiring diagram is it assumes you’ve got the full cabinet and have the metal box and filter board that came with the original cabinet. I don’t think I’ve ever seen a Point Blank filter board and over the years I’ve seen probably a hundred Point Blank PCBs. That means there are signals shown but it doesn’t show where they actually come from. I knew the large transistor chip was responsible for driving the solenoids so with the gun solenoid being triggered in test mode I probed the driver chip but none of the pins toggled. At first I thought the driver chip was bad but that would be a mistake to jump to that conclusion without fully understanding how it works. If you can’t understand how something works or at least know what it requires to make it work then you can’t fix it. I needed to look into this deeper and thus a 3 day research adventure entailed until I eventually figured out exactly how this thing works.

The details are all in this schematic I drew which shows the signals and how they are connected....

On connector J2 are the signals that go back to the solenoids. These are just a simple ‘logic low’. An easy way to test the function of the I/O PCB without 24V is to use a couple of LEDs tied through a 470 ohm resistor to 5V on one side and to those solenoid return signals on the other side....

In test mode the solenoids were checked on the GOUT screen and everything now works fine. The problem was simply the GOUT0 and GOUT1 signals on the main board were not connected to the I/O board, and of course the butchered 74LS14 was the actual problem with the gun I/O PCB. Why those ‘gout0/1’ signals don’t come through the I/O board edge connector is a mystery. All I know is with those signals hooked up to J4 the solenoids receive the return signals and fire so the board is fixed and can go back to the owner now.

12th December 2025 I’m working on something new for MAME layouts.

Dial Animation....

Very quick ‘proof-of-concept’ demo of working dial finger dimple animation. Unfortunately I don’t know enough to write a proper lay script to fully control the entire dial but this shows it’s possible with the existing bare-bones low-end hack-job cocky-snot MAME layout system... and you saw it here first ;-)

10th December 2025 Here’s the next MAME layout just completed....

Akai MPC2000XL

5th December 2025 Here’s three more MAME layouts just completed....

Korg microKORG (lots of TODO’s on this one because the emulation does nothing so this is essentially just a static pic)

Akai S2000 16-bit Sampler (rack unit). The layout in current MAME is an early work-in-progress and hasn’t been updated yet to this nice complete version. Update: The newer version was updated and is now in MAME.

IBM 5110

I believe this is the very first ‘proper’ keyboard layout in MAME and also the very first pseudo-3D full computer layout in MAME. Two firsts with one layout ;-) Very fitting as the 5100-series was the first commercially available portable all-in-one (keyboard + computer + monitor) computer :-)

Update 19 Jan, 2026: Scratch that. It was rejected after sitting in a PR and being totally ignored for 2 weeks. Not because of any technical reason or that it’s bad. No, in fact everyone liked how it looks and everyone thinks it’s cool. It was rejected because the TAB spacing was wrong. Yes really. WTF? Now I could fix it to some virtual ficticious ‘standard’ that some loser made up but I don’t jump through hoops-on-fire for anyone. That asshole at the top really needs his head checked. Until MAME is run by someone else anything cool simply gets rejected. It’s a set back for MAME progress and one of many in the past. Such a shame really as MAME could have been SO much better. So, there won’t be any more layouts from me for some time, maybe never again.

10th November 2025 Another interesting little board came in for repair. It’s a Donkey Kong Jr bootleg called ‘Donkey King Jr’ but the board is different to the one I dumped years ago. The other board was a single piece board whereas this one is a dual board connected with flat cables. It appears to be running but the screen is messed up...

While making up the JAMMA adapter I probed the edge connector with power connected and found the RGB pins fairly easily but I couldn’t find the sync pin (on my logic probe that pin sounded like squeezing a duck’s nuts lol!) so it looks like the sync is missing. The simplest way to start diagnosing this is to trace the sync pin (edge connector pinout was provided on an old piece of paper) to the first connected logic chip. It ended up going to pin 13 of this 74LS02 on the lower board. I need to check the datasheet....

The datasheet shows pin 13 is 4Y output. Probing that pin shows it’s completely dead, nothing! Pins 11 and 12 are inputs and are active so that looks like the culprit! I measured pins 11 and 12 using my frequency counter and they measured 59.something Hz and 15.6-something kHz... that’s VSync and HSync, so yeah, this chip is being used to merge the two syncs to make composite sync and it’s dead. I pulled the chip and it tested bad....

The display shows it has lows on pins 11 and 12 and outputs a low on pin 13 but a high was expected as per the truth table shown on the datasheet. I quickly piggybacked the chip with a good chip, powered on and it worked and at the same time the owner’s mind blew lol! After swapping the chip the game works great. The board was taken by the owner soon after that as he’s on a short tight leash so I fixed it while he waited.... his Mrs had phoned up just earlier and was badgering him to stop wasting time with this rubbish and get back home so she could tell him to do stuff lol! I wanted to dump the ROMs to make sure they were dumped but forgot so it’ll be coming back at some point to do that and if it’s not dumped it’ll be added to MAME. For now that’s the end of this repair.

Update: I dumped the board and it matches the dump of Donkey King Jr I did some years ago.

5th November 2025 An interesting little board came in for repair....

You wouldn’t know it by looking at it but this is a Galaga board. Well technically it’s Nebulous Bee but that’s just a Galaga bootleg with a different name. It’s kind of nice with both boards the same size and no silly little plug-in board emulating a custom chip like on the more common blue Galaga bootlegs. The game boots with strange colors (looks like one color is missing... more on that later) and shows an error. Sometimes the sync would also be messed up. If you’ve ever repaired any Galaga board this will look VERY familiar....

The code shown there is NOT the location of the bad part, it refers to the memory area. Galaga has several RAMs/ROMs named L and H with some numbers assigned to those memory areas. That’s just how the Namco engineers did it. L means low memory and H means high memory... as in upper and lower bits. The number is just an arbitrary assignment starting from 0 so you could potentially see RAM or ROM from 0L/0H to 3L/3H or even up to 6L/6H on games like Bosconian (which uses a very similar board). Most of that is documented on the net but it doesn’t really help in this case because this board is very different from any other Galaga I’ve seen so it’s going to have to be a guess as to which RAM is bad. But I’m not completely blindfolded. Those RAM errors are always on the video board so the problem has to be on the lower board. This board has a bank of 6x 2114 static RAMs. You could assume 0 is at one end and 3 is at the other end but that could easily be incorrect. The obvious place to start is piggybacking those RAMs and power-cycling to test it. Strangely nothing made any difference and the error persisted even with good RAMs in place. I probed the RAMs and one RAM had dead pins. In all the years repairing stuff I don’t think I’ve ever seen a RAM with dead pins with absolutely no output. I need to get the datasheet and see what those pins are.....

The dead pins were 12, 13 and 14 so they are 3 of the 4 data outputs. This RAM is 1kB x 4-bit which means it’s got only 4 data pins but obviously being used in pairs (low and high bits). Anyway, that RAM must be bad so I pulled it....

Yup! That one is goneski. I replaced it and expected the game to boot or at least progress to a different screen but no! It still said RAM 2L. I pulled all the other RAMs in that area but all tested good. I figured maybe one of the connected logic chips was bad so I put the RAMs back and re-checked it to make sure it was the same before checking the logic. At power-on now it works! Weird but I’ll take it, I’m moving on LOL!! I can’t comment on the other RAMs but in the pic below the far right RAM in the socket is 2L.

Now to the untrained eye it may look like a color is missing but that’s not actually the case here. If one was missing there would not be white. To get white in the RGB light spectrum you need all 3 colors together. The score was white so RGB is present. Plus when I probed the RGB outputs all 3 colors were active. The Galaga schematics can help to show the RGB signals a bit more clearly....

The large chip is a Namco custom chip that is documented as generating the starfield. The schematic doesn’t show any other video data path to the outputs so the 05XX must also process the game video data (coming into the 74LS259 pin 15.... classic hard to read schematic LOL) and split it up into 3 separate RGB paths which are fed into the output DAC. The chip below that is the color PROM which provides the color palette data. The PROM and custom chip outputs are mixed into the final RGB stream via a R2R ladder DAC. All the color PROM outputs are active and I had previously pulled and read the PROM and it matched MAME archives so I know it was good. The custom chip obviously isn’t present on a bootleg board but on the bootleg that magic is done by common logic chips which is why bootlegs are always fixable and original boards are very difficult or often impossible to fix when a custom chip is bad. Anyway, on this board all the logic looks correct. I’ll skip the part where I flailed around for an hour and got nowhere and jump straight to the fix. At one point while testing it I did actually see *almost* correct colors but then it disappeared. At that point I figured it was a bad connection somewhere so I checked the JAMMA adapter and found that the green wire was touching the sync wire. The harness supplied with this board is not the best and has some loose pins. That explains everything LOL!

It looks fine now, right? Well not really. Maybe it does if you’ve never seen a proper Galaga but here at Gurudumps Pty. Ltd. we go the extra mile because we know better heh! These bootlegs have some issues with wonky colours caused by incorrect resistors in the R2R DAC. As per the schematic page above, Galaga has a bunch of resistors tied to the color PROM. I had previously documented a repair where I swapped a few resistors that were wrong on a bootleg Galaga blue board but this board is different so that fix can’t be done 1:1. I traced the resistors from the PROM and it turned out that some of them were correct. I swapped a couple as per the recommendations but it didn’t really fix the issue. Basically green was very dark and white was pink-white. I shorted one of the resistors I had just swapped which is connected to pin 6 of the PROM (one of the green outputs) and I got this....

That’s maybe a bit too much green heh! This particular resistor is 100 ohms. The next most common lower value resistor is 47 ohms so I soldered it over the top and got this....

That’s much better. I removed the 100 ohm resistor, soldered in the 47 ohm resistor and that’s the end of this repair :-)

31st October 2025 A new MAME just hit the streets today (0.282) and it includes the working MPC3000 with my artwork :-) Following on with more artwork, I have just completed some for the MPC60 which is another drum machine made by Akai in 1987. Akai milked the hell out of this series of products and there are dozens of models, all doing basically the same thing with slowly increasing levels of better software and UI. Up to now it’s been non-working in MAME but R. Belmont just fixed one of the main issues stopping it from being usable and got it working better (but without sound as the sound stuff isn’t emulated yet) so now it has clickable artwork too. The improved emulation should be in the next version of MAME, not the one that was just released today. Just a reminder, the MPC3000 factory floppy disks are still undumped so if you have the real floppies or you know someone who does, please make an effort to get them backed up to images. Thanks!

23rd October 2025 Over the last week I’ve been working on some .lay artwork for MAME. This is a very special gizmo that has just been emulated in MAME by a bunch of MAME guys headed by R. Belmont and is now working. This is the legendary Akai MPC3000 drum sampler and sequencer machine from 1994. The unit has more keys than what is available on a regular PC keyboard and a developer request was put out for artwork to help with developing it and using it in MAME so I sprung into action and made some artwork. But this time instead of external artwork images like I did for Namco’s ‘Golly Ghost’ this was all done using XML code. This also includes all the fancy svg logos that I painstakingly re-created. Going this deep with artwork made only from code is a first for me and I think it turned out pretty nice in the end :-)

You can see a video of the emulation in action here. Ignore the off-center image and weird size, it’s a shitty Macbook with a shitty and non-standard screen LOL! On a Windows PC it displays full screen and centered :-) Well actually I was told it centers correctly even on a Macbook and recording with OBS was messing up the centering, but hey, that doesn’t stop me from kicking Apple in the nuts anyway LOL!

The MPC3000 is backwards-compatible with the MPC60 and that older software (which is also *very* difficult to find now in a _GOOD_ backup image) is being used for the drums in this vid. I think R. Belmont left it in the sun for too long and it went yellow LOL! No, actually that was an earlier version of the artwork I did. I decided to go with a ‘new look’ in my final version which is hopefully close to what it would have looked like when new in 1994. Note the official layout included in MAME won’t have the trademarked logos so the proper artwork will be available on this site in a bit. I’m not sure where to put it yet heh!

Now a request. In the past people who were using these musical devices were busy using these musical devices instead of properly backing up and securing the data that makes them work. Now 31 years later this has resulted in a lack of backups of the 4 floppy disks that shipped from the factory with the MPC3000. I don’t mean copies or like the poorly-scanned and completely unreadable MPC3000 schematics that random ebay sellers are selling or any one of the random MPC disks being sold on ebay by random sellers, I mean the original factory floppy disks shipped with the MPC3000. A high-res 600