MagicianLord77
Enthusiast
Hi guys, I got hold of an untested Konami TMNT PCB. Initial inspection was good, looks alright after cleaning with brush, no scratches with broken traces, notable details are just 2 socketed SRAMs (G8 and F23) and quite some soldering glob over each JAMMA pin, seems they were a bit rusted/corroded and somebody decided to melt an amount of solder wire all over each one, this causes the harness to get stuck when trying to connect it. So:
Booted it up, got greeted with the RAM ROM CHECK screen, everything OK but G8 reports BAD, then it reboots and does the same. Seems like an easy repair I thought, G8 is the socketed SRAM, a Toshiba TMM-2018AP-35 (6116 type), how about switching it with the other socketed SRAM, F23?. Nope, exact same issue. Just in case tested both srams in the Xgecu, both test fine.
OK, so this might be one of those dreaded Fujitsu logic ICs disturbing the access to the G8, checked the schematics to see how G8 is wired up:
Oof, it is 100% connected to a custom, the 051960!.
Grabbed the logic probe and performed some measurements on G8, it doesn’t have any activity on its address pins, and I guess it should have, since it’s not a shared bus. Also measured that with the IC removed from the socket, the data pins are also all in Hi-Z, not low or high.
I also considered the theory that something somewhere else might be tickling the 051960 in a bad way to operate like this, but that would be too much. After these measurements I’ve set my sight on replacing it, no matter how hard would it be.
After that I couldn’t find any more issues. Repair complete!
Cleaned up the solder from the JAMMA pins with a thick 3mm desoldering braid and a 3.00mm (“wide”) flathead solder tip to make the job easier, chose a rather high temp (420C), and got them all cleaned up, bringing them back a pleasant appearance. Cleaned old burnt up flux with flux cleaner. They look quite nice now, almost like if nothing ever happened. Also finished them up with a careful spraying of contact cleaner, that liqui-moly one is like oil in a spray, I didn’t enjoy it on IC sockets, would make a spill of oil around the board, but for larger kinds of connectors like the JAMMA edge its great. Harness fits like a glove!.
before and after (believe me they were thick):
before and after (believe me they were thick):
Took out all roms and dumped them with the Xgecu T48, all got identified ok using mame -romident command, so, no issues over here. Also sprayed the sockets with non oily contact cleaner.
Booted it up, got greeted with the RAM ROM CHECK screen, everything OK but G8 reports BAD, then it reboots and does the same. Seems like an easy repair I thought, G8 is the socketed SRAM, a Toshiba TMM-2018AP-35 (6116 type), how about switching it with the other socketed SRAM, F23?. Nope, exact same issue. Just in case tested both srams in the Xgecu, both test fine.
OK, so this might be one of those dreaded Fujitsu logic ICs disturbing the access to the G8, checked the schematics to see how G8 is wired up:
Oof, it is 100% connected to a custom, the 051960!.
Grabbed the logic probe and performed some measurements on G8, it doesn’t have any activity on its address pins, and I guess it should have, since it’s not a shared bus. Also measured that with the IC removed from the socket, the data pins are also all in Hi-Z, not low or high.
I also considered the theory that something somewhere else might be tickling the 051960 in a bad way to operate like this, but that would be too much. After these measurements I’ve set my sight on replacing it, no matter how hard would it be.
So, first step would be to get a replacement one, searched over Ebay for cheap Konami boards, and found at a very affordable price a Bottom of the Ninth PCB, that according to System16 (also checked on MAME sources), has a 051960, it’s a match!.
Now, about how to lift the customs off, I’ve seen different ways, I chose to go with my regular hot air station (an 852D but there are a lot of similar models), and got a QFP 28x28 nozzle for it so it can do the job. I have to admit I tried desoldering the custom with one of the regular round nozzles to avoid purchasing the QFP nozzle, it didn’t work at all
, learned that custom is too big for it.
Having got the nozzle, kapton tape’d the surroundings of the custom, and applied some flux on the pins (not sure if needed but just in case), began heating, took around 3.5-4 minutes for the custom to get loose at rather high airflow, because of the nozzle had to increase it. Picked it up with some antistatic tweezers.
Done the same on the other board, success! No bent pins fortunately. Cleaned flux residue of the donor custom IC with some flux cleaner. Also, using a more fine solder braid, cautiously cleaned remaining solder from the pads of where the donor will be placed, trying to achieve a flat landing surface.
Then, began preparing the area for soldering, covered the pads with “Mechanic Solder Paste 63/37”, I preferred to use the one that comes in a little bucket, picked it up and spread it over the pads with a toothpick. Then I had to place and align the custom, I think it would have been easier if I had one of those “smd vacuum pen pickup tool”, but I don’t so just went with the tweezers, tried to align it the best I could.
After alignment, started with the soldering! I don’t know if I should have used the regular round nozzle, but I went with the QFP one, it took a while and it got soldered up!, alignment was a bit crooked but good enough anyways. Couldn’t wait any longer and after letting it cool down for a while fired it up!.
Got an unpleasant glitchy screen I didn't get to take a photo of, but in between the glitches G8 was still bad.
Without allowing panic to take control, went to double check the pins, and began touching them up with the soldering iron, all of them.
Turned it on again and G8 reported OK, game started up flawlessly! (graphically). I guess there were some internal solder bridges that were fixed up with the soldering iron. Also, later heated it up again with the hot air station and aligned it more properly. Then cleaned the area up with flux cleaner, and also sprayed some alcohol below the custom, to try get rid of any residue from below it.
Now, about how to lift the customs off, I’ve seen different ways, I chose to go with my regular hot air station (an 852D but there are a lot of similar models), and got a QFP 28x28 nozzle for it so it can do the job. I have to admit I tried desoldering the custom with one of the regular round nozzles to avoid purchasing the QFP nozzle, it didn’t work at all
, learned that custom is too big for it.Having got the nozzle, kapton tape’d the surroundings of the custom, and applied some flux on the pins (not sure if needed but just in case), began heating, took around 3.5-4 minutes for the custom to get loose at rather high airflow, because of the nozzle had to increase it. Picked it up with some antistatic tweezers.
Done the same on the other board, success! No bent pins fortunately. Cleaned flux residue of the donor custom IC with some flux cleaner. Also, using a more fine solder braid, cautiously cleaned remaining solder from the pads of where the donor will be placed, trying to achieve a flat landing surface.
Then, began preparing the area for soldering, covered the pads with “Mechanic Solder Paste 63/37”, I preferred to use the one that comes in a little bucket, picked it up and spread it over the pads with a toothpick. Then I had to place and align the custom, I think it would have been easier if I had one of those “smd vacuum pen pickup tool”, but I don’t so just went with the tweezers, tried to align it the best I could.
After alignment, started with the soldering! I don’t know if I should have used the regular round nozzle, but I went with the QFP one, it took a while and it got soldered up!, alignment was a bit crooked but good enough anyways. Couldn’t wait any longer and after letting it cool down for a while fired it up!.
Got an unpleasant glitchy screen I didn't get to take a photo of, but in between the glitches G8 was still bad.
Without allowing panic to take control, went to double check the pins, and began touching them up with the soldering iron, all of them.
Turned it on again and G8 reported OK, game started up flawlessly! (graphically). I guess there were some internal solder bridges that were fixed up with the soldering iron. Also, later heated it up again with the hot air station and aligned it more properly. Then cleaned the area up with flux cleaner, and also sprayed some alcohol below the custom, to try get rid of any residue from below it.
Everything seemed fine until I connected the speaker, the intro song does not sound good, it’s like it is stuck on a loop of its first second of playing. I heard that the PCB has a specific circuit for the intro song, and it is really that way, the addresses of a ROM that contains the intro song are driven by 3 74LS393s. Began by checking them since the issue was looping sound, it must be something related to the addressing. Found that one of them, C7, was not outputting a signal to drive the following one. And it was a Fujitsu
Desoldered it, soldered a socket and installed a spare 74LS393. That was it!. Intro sound sounds complete and nice.
Desoldered it, soldered a socket and installed a spare 74LS393. That was it!. Intro sound sounds complete and nice.
In between testing, also another issue arise, audio would begin to generally mess up, have dropouts, after minutes of having the board on. Did some quick research over the internet on other TMNT repairs and found that many had a similar issue, and that either replacing the op-amps and/or some tantalum caps was the cure, so I went for it, first replaced the LM358 and LM324 op amps and it didn’t make a difference, then replaced the tantalum caps (C6, C73, C74 - 4.7uF 16v) with some electrolytics that I had at hand, and that seems to have nicely solved the issue, no more audio degradation .
.Could that be the last repair? I thought so!. Until I began having some fun with it, noticed some sounds were missing, mostly the ones when turtles hit stuff, but couldn’t tell from where they came from. So I used MAME to clearly pinpoint what was missing, by pressing the ~ tilde key, and changing the values of the volume sliders, found that muting the uPD7759 would make it sound the same as the PCB, so that’s what’s missing.
So did some research about the IC, it’s a chip that needs a clock, and upon receiving commands it plays sounds from a ROM wired to it. Logic probed it and did not found active signals on its data or address bus, weird, but could be on “standby” mode according to its datasheet?
Tried to look at it’s clock input, measured with a scope on the X1/X2 pins that there’s no clock, weird, I’ve heard crystals/resonators rarely break. The board has 2 640Khz resonators, one for the intro music circuit, and one for this circuit. The other one is blue, says “640J”, this one is green, hmm, what does it say? “H473K”, that’s a weird frequency, hold on a minute, how could I have missed that! That’s not a resonator! It’s just a film capacitor! I just thought green resonators could exist but no, they don’t seem to.
Why would anyone solder a capacitor in place of a resonator? (it’s also almost the same size), my guess is to have the uPD7759 in a “halted” state, if nothing’s on its place then might run unpredictably?. So I guess at some point the original resonator was kicked out (have read that’s a common issue), and replaced with that.
Let’s try to replace it with the proper part!. Went to Digikey/Mouser, found that 640Khz resonators aren’t a thing anymore, seem obsolete, bummer. Found packs of them on Ebay from Germany, but before getting into that, started searching on the junk PCB pile just in case, and found a pretty close resonator, a 655Khz one from a bootleg Golden Axe board!. It’s close but not a match, since it would be a clock for the CPU, I want to replace it with what it needs. But how about using the 640Khz one that came from the intro music circuit for this job, and putting the 655Khz one on it’s place, after all it’s just %2,34 faster, an almost imperceptible pitch shift on the song.
So, did that, now intro song sounds like half a note up or similar, not a big deal for the meantime, but when the time of playing back speech/sfx from the uPD7759 comes, it outputs weird noises in place of its sounds
, sometimes even gets in a loop and doesn’t stop playing.
I thought it could be an issue with it’s surrounding logic (F15 and B16, both LS273), or even it’s ROM (D18 1Mbit), even though they seemed fine with the logic probe, replaced them with sockets and tested them. All good.
Upon being unable to fix the issue I had to test replacing the uPD7759 IC itself, and fortunately the Golden Axe bootleg board had one, took it out. Installed socket, put it in the PCB, works perfectly, that was it!.
So did some research about the IC, it’s a chip that needs a clock, and upon receiving commands it plays sounds from a ROM wired to it. Logic probed it and did not found active signals on its data or address bus, weird, but could be on “standby” mode according to its datasheet?
Tried to look at it’s clock input, measured with a scope on the X1/X2 pins that there’s no clock, weird, I’ve heard crystals/resonators rarely break. The board has 2 640Khz resonators, one for the intro music circuit, and one for this circuit. The other one is blue, says “640J”, this one is green, hmm, what does it say? “H473K”, that’s a weird frequency, hold on a minute, how could I have missed that! That’s not a resonator! It’s just a film capacitor! I just thought green resonators could exist but no, they don’t seem to
.Why would anyone solder a capacitor in place of a resonator? (it’s also almost the same size), my guess is to have the uPD7759 in a “halted” state, if nothing’s on its place then might run unpredictably?. So I guess at some point the original resonator was kicked out (have read that’s a common issue), and replaced with that.
Let’s try to replace it with the proper part!. Went to Digikey/Mouser, found that 640Khz resonators aren’t a thing anymore, seem obsolete, bummer. Found packs of them on Ebay from Germany, but before getting into that, started searching on the junk PCB pile just in case, and found a pretty close resonator, a 655Khz one from a bootleg Golden Axe board!. It’s close but not a match, since it would be a clock for the CPU, I want to replace it with what it needs. But how about using the 640Khz one that came from the intro music circuit for this job, and putting the 655Khz one on it’s place, after all it’s just %2,34 faster, an almost imperceptible pitch shift on the song.
So, did that, now intro song sounds like half a note up or similar, not a big deal for the meantime, but when the time of playing back speech/sfx from the uPD7759 comes, it outputs weird noises in place of its sounds
, sometimes even gets in a loop and doesn’t stop playing.I thought it could be an issue with it’s surrounding logic (F15 and B16, both LS273), or even it’s ROM (D18 1Mbit), even though they seemed fine with the logic probe, replaced them with sockets and tested them. All good.
Upon being unable to fix the issue I had to test replacing the uPD7759 IC itself, and fortunately the Golden Axe bootleg board had one, took it out. Installed socket, put it in the PCB, works perfectly, that was it!.
After that I couldn’t find any more issues. Repair complete!
Last edited:
