 We've been here before! Last time I wrote about repairs to a Teenage Mutant Ninja Turtles (TMNT, Data East, 1991), I was discussing a restoration I performed for a customer. This machine was a little different. I bought it after doing some repairs for its previous owner. They decided to move and didn't want to move the game with them, so they sold it to me instead. It's a weird experience to start repairing a game for somebody and then finish repairing it for yourself. Suddenly you're not constrained by time frames or repair budgets, and you can really take the time to make sure the game is playing as well as it can be. You can also take the time to make any improvements or modifications that you want. And with all of the extra features I planned to install, I needed a lot more time than most people can justify in a regular service call. But what "extra features" am I referring to? Most pinball manufacturers back in the day would produce a small number of test or prototype machines before starting full production on a new game. Data East was no different. A lot of the time, features would be removed from the games before they entered production in order to save costs. This meant that prototype games often had more features, and playfield mechanisms than the more common production games. Sometimes, the features cut were quite minor and there's no discernible difference between the prototype and production games at all. However, there are a lot of differences between the prototype and production versions of Teenage Mutant Ninja Turtles. So, I decided to convert it to the prototype version by installing all of the mechanisms and features that were removed from the production machine, including a vertical upkicker, ramp, rotating beacon topper, several lamps including two feature lamps on the top of the backbox, topper plastic, autolauncher start button, and the prototype game code to run everything. That's quite a few new features, and it was a long journey to get them all working as they should! No power to game
This is one of those problems which stumps you and causes you to think of all kinds of silly reasons for what is going on, but actually turns out to be stupidly simple.
When I first tried to turn the game on (before I bought it, in fact), nothing would happen. No general illumination lamps would turn on. No LEDs on the MPU board would light up, either. Voltage test points on the power supply all read 0 volts. I tried a few basic things including checking the main power fuse, reseating connectors (including transformer outputs), making sure the power cord was properly plugged in, but all of these things seemed OK. I started checking some of the transformer outputs out of curiosity and noticed that some of them were measuring far below what they should have been. Hmm—first clue! I figured there was either a problem with the transformer, or the power coming into the game. Transformer problems are unlikely, so I started investigating the second hypothesis. I checked the power module in the cabinet and everything seemed OK there. The active, earth and neutral wires were all securely connected. 
Interior of the power module including line filter, main line fuse and volume potentiometer.
Maybe it wasn't an issue with the game, but with the power supply at this location? Nope. There was a steady 240 volts at the wall. Following the power through, I checked the voltages at the active and neutral pins of the power cord as they entered the power module and attached to the line filter. Aha! This was measuring around 20 VAC. Way too low! I started tracing the power cord back through the cabinet and outside the game. Power cords on this era of games can often simply be pulled in or out of the main cabinet as there is no strain relief or securing device to lock them in place. Just outside the cabinet, I found a weird-looking kink in the cord which appeared to have been caused by some kind of mechanical impact. Some of the insulation on the conductors underneath the outer insulation layer was visible. Clearly this cord had had a hard life and someone had either smashed it or run over it with a vehicle. Either way, it was suspect and had to be replaced. There are a couple of options when replacing these mains cables. The turnkey solution is to use a mains plug with bare wires on the other end (Jaycar). However, these might not be available as readily or in in the desired lengths. Instead, it's easier to grab an extension cable of the desired length, cut off the female receptacle, and then strip the wires as required. You can get extension cables anywhere and everywhere, so pack a couple in your toolbox. You'll need them a fair bit, as the power cord is often the most used and abused part of any game! Ensure that, whichever cable you buy, it is correctly rated for mains voltage and current. I replaced the cord and tested the voltages at the line filter again. All good! It appeared that we were now getting power to the transformer and further up the chain to the backbox boards. 
New line cord installed.
Power supply issues
It's not a Data East repair without some power supply issues. Even after replacing the power cord and getting some voltage readings at the test points on the power supply board, we weren't out of the woods yet. I was still not getting the game to boot up. I had fixed the power cord issues, so we were getting good voltages into and out of the transformer. Next port of call was the power supply board, where the transformer voltages are routed to (top left of backbox). After checking that the fuse clips were not broken, I saw evidence of capacitor failure below the heatsink.
Evidence of leaking capacitors on the power supply.
The culprit in this case was C2. This capacitor commonly leaks, interrupting the 5 volt circuit path, which the CPU board needs to boot the game. The problem is well documented on Pinwiki. I tackled this with the same approach as tackling battery alkaline damage. Sand the area with sandpaper until the bare copper is revealed, treat with vinegar and rinse with isopropanol, then apply a conformal coating before repopulating the board components. I listed the necessary board components in a previous repair blog. Apart from typical resistors, capacitors and diodes, IC1 (regulator) is getting a little harder to source. Luckily it is still available locally (PSPA). 
Board view after removal of components and conformal coating the affected area.
I dutifully replaced all of the affected components, reinstalled the board, and flipped the power switch. Still nothing. What?! I double-checked all of my work, ensured all of the correct components were installed, but I couldn't find a problem. I started to look further along the trace that connects pin 3 of the LM723 regulator at IC1 to the test points at the top of the board. This trace runs below C2, alongside the legs of the bridge rectifier, and progresses vertically up the board. I saw some potential evidence of leaking electrolyte further up this trace. I figured there may be a problem further up the trace, so I decided it would be best to treat the electrolyte leakage more extensively. Before I did this however, I decided to test the bridge rectifier at DB1, as this can also contribute to game booting issues. To my surprise, I found the issue here. One of the diodes within the bridge rectifier had failed. This was preventing power from getting to the CPU board. I got the Dremel tool out and cut the lug-style bridge from the board. I hate these types of bridge rectifiers! Luckily, it removed quite cleanly, so I proceeded with sanding of traces and abatement of alkaline damage until all of it was gone.
Then, I covered the exposed traces with solder and installed a new bridge rectifier with thinner leads (RTBB, PSPA).
Once that was all done, I reinstalled the board again and turned the game on. This time, it worked! This was a bit of a weird board with two problems; one much more obvious than the other. At the end of this I can't confirm that the leaking capacitor at C2 was contributing to the game's inability to boot, as the bridge rectifier appeared to be the main cause of that. However, the alkaline damage still needed to be abated, so I'm not annoyed at having to do the extra work. The board has been totally refurbished at this point, so it should be good for a few more years! 
Fully rebuilt 5-volt and 12-volt circuits.
Dead display
Teenage Mutant Ninja Turtles is one of the five games Data East released with the smaller format 128x16 dot matrix display. These displays bridged the gap between alphanumeric displays used in the very first Version 3 MPUs and the later, more common 128x32 format display which became the industry standard in the early 1990s.
My display worked without issue for a month or so when I first bought the game and sorted out the initial issues with power and the MPU board. Then, it went dead. I turned the game on one day, it booted up as normal, and the display turned on. A second later, the display went dark. I could still start a game and enter the diagnostic menu, but without seeing what was on the screen, I was flying blind. The first step was to check the voltages on the power supply board. The display requires five main voltages to function: -110v, -98v, 68v, 12v and 5v. The -110, -98 and 68 voltages form the "high voltage" circuit and are used to run the plasma display. The 12v and 5v circuits are used throughout the game, mainly for component-level logic, but do not really have anything to do with igniting the gas in the display glass. Of course, you'll need the 12v and 5v circuits for the game to function properly, but the 68 and higher voltages are the most important when it comes to seeing something on the screen. I tested my voltages at CN5 on the power supply board, which is the connector that supplies the display panel. All voltages were within spec except for the 68v circuit (pin 4), which reading less than 20v. A missing 68v circuit would definitely result in nothing showing on the display, so I had found the source of the problem. Pinwiki has a great section on diagnosing issues with these displays. One of those sections deals specifically with failures in the 68v section. It also includes a very handy table listing all components in the 68v circuit which I referenced a lot while doing this repair. But I still had to find the problem component, so it was time to get the board out of the backbox and see what I could find. 
Schematic of the 68 volt circuit on the power supply board.
When I opened the backbox initially, I could smell the telltale smell of burning componentry. So, I figured I might be able to see which component had spilled it guts with the naked eye. It took a little bit of searching both with my eyes and nose until I found that TR3, the MJE340 transistor that drives the 68v circuit, had suffered some kind of failure and burned at the base of pins 1 and 2 (emitter and collector, respectively). Some kind of short had developed causing the two pins to arc. Yep, this will definitely cause you to lose your display! 
The other components in the circuit appeared to be OK, so I took this to be an isolated incident and simply replaced the transistor. You can replace with an MJE340 like the original (Jaycar, PSPA, RS Components) or you can use a heavier duty MJE15030 transistor instead (RTBB, PSPA, RS Components). If using the MJE15030 transistor, keep in mind that the pinout is different from the original MJE340! The pins are actually reversed, such that the base and emitter have swapped positions, so you need to install the MJE15030 backwards on the board. The heatsink will also need to be turned upside-down so it fits on the board.
I happened to have some MJE340s spare, so I decided to use these. I cleaned up the board with alcohol, soldered in a new transistor, reinstalled the board, and turned the game back on. Still no display! And I detected the same burning smell that I smelled when the display first died. Not a good sign! When I opened the backbox, I saw that TR3 had, once again, shat itself. Hmm. So, I took the board out again, removed TR3, and again inspected the rest of the components. Pinwiki has some good advice in this regard and actually suggests replacing components in this circuit in a specific order. First, all resistors in the circuit should be checked and replaced if necessary. Next, the 3.9v and 68v zener diodes at D6 and D9 should be changed, followed by the transistors at TR3 and TR1. A failed zener diode can cause a perfectly good transistor to fail, so replacing the transistors first is pointless because a failed zener diode may cause them to blow up continuously. Perhaps that was what was happening here? The zeners both seemed to test OK in one direction when I tested them with my meter, but a zener diode will allow current to flow backwards once a certain voltage is applied. I didn't have a power supply which could give me an output of 68v, so I decided to replace both the 3.9v and 68v zeners just in case one of them was the problem. So, I replaced them both, as well as a couple of resistors which tested slightly out of spec, and reinstalled the board. As soon as I turned the game on—same problem! TR3 burned out again. What the hell?! Clearly, this was not a coincidence. I had replaced most of the components in the 68v circuit and TR3 still blew up. So, it was time to remove all of the components in the circuit and have a closer look at them all. This is where it helps to have a strong magnifying glass or a loupe to find small, physical issues on boards. Significantly, all of the components except for TR3 appeared to test OK once removed from the board. Two of the pins of TR3 has shorted internally (the emitter and the collector). This seemed odd to me. When I looked at the through holes where TR3 was mounted, I saw charring between the two through holes that the emitter and collector sat in. Out of interest, I grabbed my meter and placed it across the two through holes in resistance mode. Surprisingly, I got a reading! It measured a few hundred ohms, and on the diode setting, the meter also returned a positive reading. This was not good. These two pins should have no continuity at all between them, and placing meter probes across them should return an open loop (OL) result. There should definitely not be a resistance as small as a few hundred ohms between them. So, this was the source of the problem. The failure of the original transistor had created an arc between the two through holes, which had carbonised the fibreglass between them. This created a tracking fault through which current could run, effectively shorting the collector and emitter of the transistor. Any component installed in these through holes was therefore doomed from the start! So, what to do? I tested this theory by scrubbing the area vigorously with alcohol and a cotton bud. During several minutes of scrubbing, I tested the resistance between the trough holes at regular intervals, and saw that it slowly increased every time I cleaned the area. So, by removing the carbon residue I was slowly but surely removing the conductive path. Eventually this no longer had any effect, and my meter was showing there was about 5M ohms resistance between the two traces. I wanted to make sure the bridge between the two holes was totally severed, so I cut a line between the two holes carefully with a scalpel and dug into the fibreglass. This helped remove the carbonisation which was embedded deeper in the board. Normally I hate having to butcher a board in ways such as this, but in this case there wasn't really another option. Eventually, resistance between the two holes rose high enough and the meter stopped returning a reading between them in diode mode. That was good enough for me and showed that there was no longer a conductive bridge connecting them. 
Severing the carbon tracks between pins 1 and 2 of TR3.
So, I sheepishly reinstalled all of the components including a new MJE340 transistor at TR3. Then, I turned the game on. Success! The display turned on and was nice and bright. I was now getting a full 68v from the power supply. So this one turned out to be quite a sneaky problem which took the sacrifice of a couple of transistors to figure it out. This kind of issue is likely only one that you'll encounter in high voltage circuits such as those which driver plasma displays. Only at higher voltages will you get an arc substantial enough to carbonise enough material to actually bridge two traces together. In the same vein, if this was the -98v or -110v circuit, I expect the damage to the board would have been much worse. In those circuits, TR4 would have been the victim rather than TR3. So, if you ever encounter a blown high voltage transistor on these Data East power supplies, don't neglect to check for damaged traces. 
Fully rebuilt 68v circuit.
Knocker and flashers not working
During coil and flasher test, I was getting no response from the knocker in the cabinet as well as some flashers. The affected flashers were the ones beneath each of the turtle inserts in the centre of the playfield above the flippers. Now, recall in Data East games that coils and flashers are "multiplexed" by the relay on the playfield power board (PPB). That is, when the relay is in the "A" position, a coil will fire when the driving transistor is grounded. However, when the relay is in the "B" position, the same driving transistor can be used for activate other devices. Most of the time, A side devices are coils, while B side devices are flashers. Page 28 of the game manual describes the theory of operation. In this case, I had solenoids 8A and 8B not working, which suggested a problem with the driving circuitry on the CPU board rather than a problem on the playfield. The excerpt from the manual below tells us where to look.
Schematic for the devices connected to solenoid drive 8.
The very start of the power train is on the CPU board at transistor Q39, so let's look there first. 
Solenoid driving transistors on the CPU board.
I didn't have to look for very long to find the problem! Q39, the transistor that drives solenoid 8, was missing. Completely gone. I took the board out to try and figure out what happened, but my best guess is that somebody removed this transistor because they needed a spare somewhere else and the knocker coil was a relatively harmless coil to deactivate. It looked to have been removed carefully, too, which was lucky. Replacement was easy; these transistors were originally TIP122s however these can be replaced with the more common TIP102s (element14), which are a little beefier. That fixed the problem, and the knocker and associated flashers started working again. 
New transistor installed at Q39.
Another thing I learned during this repair was that there is a mistake in the manual. The knocker is NOT driven by Q3 on the PPB board as indicated on the playfield coil diagram. Q1, Q2, and Q3 are actually not used at all on TMNT. Instead, the knocker is driven by Q4. I chased my tail for a little while trying to figure out why Q3 was not connected to the knocker circuit, until I realised that the wiring at connector J8 on the PPB was not connected to Q3 at all, but Q4 was. The wiring colours in the manual are still correct, however. 
J8 connector wiring on the PPB board showing only the Q4 and Q5 circuits are wired up.
Minor cabinet repairs
The upper leg bolt at the rear left of the cabinet could not be screwed in all the way without forcing the bolt. Turns out, the leg bracket was bent out of shape from years of overtightening. Plus, the threads were chewed out as well. I opted to replace it with a new style leg bracket (PSPA, RTBB). Pro tip: there are new style leg brackets to suit Data East games specifically, but they are a little more expensive than the Williams style ones, and the bolt holes are the same width apart anyway!
The old brackets sit in a nice groove in the timber, however the new brackets are too large for this. Instead, they screw into the sides of the cabinet. I placed the bracket in the correct spot and tightened the nuts on the other side. This kept it steady while I pre-drilled the holes for the mounting screws, and then screwed them in. I routed the ground braid behind the bracket, too and checked there was good continuity between it and the leg bolt.
The last cabinet repair was a minor one to the backbox insert panel. The screw that holds the locking clasp had stripped out the hole in the fibreboard panel. I filled it with some toothpicks (cut to size), some PVA wood glue, and inserted the original screw. It held tight, and dried nicely. The clasp moved backwards and forwards easily, and the screw stayed put. Easy fix!
Lockdown bar and leg refurbishment
The lockdown bar was a little worn on either side where peoples' hands would sit while playing the game. Painted lockdown bars like this don't stand much of a chance against constant touching and attack from skin secretions. So, let's patch it up.
I followed the exact same process here as I did in a very recent TMNT restoration. I scraped any flaking paint off with a scalpel, and then sanded the bare metal areas to remove any traces of damaged paint and provide a good surface for painting. The rust had progressed a little way under the paint mask, but not too far, so I didn't need to remove a huge amount of paint. I used the same paint products as in the blog post I linked to above. A coat of primer, followed by two coats of satin black and two coats of satin clearcoat had it looking quite nice, with the damaged areas covered up nicely. I also removed the beer seal (Bunnings) on the underside of the bar as this had disintegrated. Goof Off (Bunnings) and a scalpel came to the rescue this time as well, as the adhesive was very hard to remove.
The legs were a similar affair. They had chipped and scuffed in several places, so after a thorough clean, I gave them a couple of coats of satin black paint followed by a satin clearcoat. They came up well, and the satin finish imitates the powder coat look and feel well.
Installation of new playfield plastics
The playfield plastic directly above the orange standup targets in the centre of the playfield is almost always broken on Teenage Mutant Ninja Turtles playfields. The ball smacks into the targets, deflects upwards, and breaks the plastic directly above. There is a metal ball deflector to prevent this, but somehow, it never seems to work. It reminds me of the drop targets on The Simpsons (Data East, 1990), which are located in the same position on the playfield and are always broken from ball impacts. My game was no different, but somebody had replaced the plastic with a piece of red polycarbonate. They were obviously skilled with a jigsaw, too, because the edges were nicely cut and smooth.
Replacement plastic made out of red polycarbonate.
As I am a bit of a purist, I started looking at replacement options. You can buy a small set of commonly broken plastics including this one (Ministry of Pinball). Alternatively, you can buy a whole plastic set (RTBB). The single plastic pops up on eBay from time to time as well. I considered getting a complete set because most of my plastics had faded to orange from an original red, so they looked a bit crappy.
Ultimately, I decided to buy an entirely new plastic set. The reproduction set is reasonably good, with nice deep colours which appear to be a close match to the originals. The key lines on the graphics are a little thick, which takes away from some of the detail in the artwork, particularly in the turtles' faces. Some annoying defects in the original plastics have also carried over, such as cutting out text and graphics on plastics on the left and right sides of the playfield. It would have been nice if these were fixed! Alignment on a couple of the plastics is also a little off. The TURTLES plastic that sits on a mounting bracket at the top centre of the playfield is particularly noticeable for this, as the lettering is far down on the plastic. Thankfully, the plastic piece I was missing was in this set and looked quite good. I scanned it in so I could print it again in future if necessary. 
This plastic is bitchin'!
Installing the new plastic set required moving the flasher domes from their original plastics to the new ones. These are riveted onto the plastic, so you need to drill out the rear edges of the rivet, then tap it out with a punch. The original flashers had cardboard tubes inserted over the top of them to direct light upwards into the dome. I didn't bother with these, as the LED flashers are bright enough that they light the dome up easily (one of the LEDs points straight up).
New slingshot plastic installed with protective washer.
For some reason, I had a word of trouble getting the double ramp assembly to sit on top of the plastics to the left of the pop bumper area. I ended up having to remove a hex nut from one of the posts to give the ramp enough clearance to screw down. I'm not sure why the fitment was not quite the same as when I took it apart, but removing a hex nut is no big deal.
The new plastic set is a huge improvement over the washed out original plastics, which were orange as opposed to red! Although some of the images aren't as sharp or focused as the original prints, I would happily recommend the CPR reproduction plastic set. 
New plastic set fully installed!
Installation of prototype Vertical Upkicker
The production version of Teenage Mutant Ninja Turtles has a number of differences to the prototype. You can read about some of these differences on this Pinside thread. I hate it when manufacturers remove features to cut costs, and Turtles had a couple of cool features cut out of it. Namely, the upper left captive ball area had a vertical up-kicker and wireform ramp removed. I thought I would try my hand at reinstalling these features back into my game. First, I tackled the vertical up-kicker in the captive ball area at the top left of the playfield. This area is occupied by a general illumination lamp and a switch on production games. I removed these first.
Hole at the top of the captive ball area, occupied by a lamp and switch.
Luckily, the hole is the perfect size for the ball to fall into. All we need to do is install an assembly to hold the ball once it falls in, then eject it. Initially, a tossed around a couple of different ideas on how to implement this. I thought about using a standard saucer eject assembly (part no. 500-5051-11). However, this is difficult because you need a lot of horizontal space for the mechanism and there isn't enough space in this corner of the playfield. As a result, we have to go for a full vertical upkicker (part no. 500-5326-00) which occupies less mounting space on the playfield. Note that while Turtles does have a VUK in the sewer assembly, this is a "super VUK" designed to receive balls from a subway ramp, whereas we want a standard VUK for this area as the ball simply falls into the hole from above the playfield and doesn't fall below it. I had a Williams style upkicker mounting bracket in my parts box, but unfortunately the Williams brackets have alignment pins on the topside of them, designed to slot into small holes on the playfield underside. I ground these off so that the bracket would fit flush with the playfield. That's the bracket done.
I installed a VUK rebuild kit (RTBB), which included a plunger and cup and all the other necessary bits to turn this into a functional upkicker (sleeve, spring). I put in a spare coil, opting for a 23-800 (RTBB, PSPA) as this is the coil that is typically used on Data East upkicker assemblies. 
Upkicker mechanism ready to go.
Next problem was installing a switch to register that a ball had fallen into the hole. This is easy from a wiring perspective, as I simply used the wires already attached to the captive ball switch (switch no. 49). The best switch for this purpose would be a microswitch mounted directly to the VUK bracket. However, my VUK bracket did not have mounting holes. Instead, I used a fork actuator switch (PSPA). There is just enough space to install this switch adjacent to the "Save April" standup target and a playfield hole used to route wiring. Be careful that the fork actuator does not touch the upkicker bracket, either. This will send coil voltage through your switch matrix if the coil power wire comes loose and contacts the bracket. Not good! At this point I could screw everything into the playfield and check fitment. My upkicker mounting bracket was very, very close to touching the lug of the general illumination lamp just above it, so move the lamp lug or the upkicker to ensure the two aren't touching. Otherwise, make sure there is enough clearance behind the upkicker bracket so that it does not hit the playfield slide bracket. 
Vertical upkicker and switch assemblies installed.
The switch wiring was easy as it just needed to be moved from one switch to another, but the coil is not so simple. There are no connectors or wires hanging around this area of the playfield which you can simply hook up to the coil. I needed to find out how this coil was connected. Some reference images from Pinside were extremely helpful here. Pinsider Pantor80 posted the following image which revealed that the wire colours that supplied the VUK coil were brown, and violet-green. 
Photo of the original prototype VUK assembly.
Apart from the coil wiring, the photo shows the VUK mounting bracket faces outwards towards the cabinet wall, whereas mine faces inwards towards the centre of the playfield. As I used a Williams bracket designed to have the coil lugs facing outwards, I could only mount it this way. Not a big deal as it is easy to remove the bracket completely if the coil needs servicing, and the screws for the coil plunger bracket are still accessible with a small screwdriver bit. Now that we know the wire colours, we can figure out how the coil should be connected. There are two ways to identify this. First, the in-game diagnostic menu. If we enter the diagnostic menu and scroll through to coil test, we can select individual coils and the menu will display the wire colours for that coil. The hard part here is that the VUK coil was "removed" from the menu. There is no coil labelled as such. However, there are three coils identified as "not used", one of which no doubt represents the original VUK coil. So, we can match up the wire colours to find out what position the coil occupies in the playfield wiring diagram. Coil "5A" displayed the same wire colours as in the original prototype photo. 
Coil test displaying information for the removed VUK coil.
Now, the second method. Check the playfield coil and flash lamp wiring diagram in the manual (page 46). Note that the coil designation 5A is missing at the top of the page where all of the other coils are displayed. All of the wire colours leading to the coils are displayed here. The power wire (solid colour) at the top going to most of the coils is identified as ORG (orange), but this is incorrect. It is actually brown when you check the wire colour connected to the coils. So we know we have to connect the power lug of the VUK coil to the brown wire. The nearest coil we can piggyback from that is also connected to the brown wire is the manhole VUK coil halfway up the playfield. This coil is connected to the brown wire via an IDC connector. So I replaced the IDC connector with a Molex type connector, and crimped the original brown wire as well as a new wire into the same terminal. I then connected this new wire to the new VUK coil, thereby giving it power. I used a green wire instead of sticking to brown, so I lose some points from the purists there, but green was all I had at the time.
Now to the drive wire on the other coil lug. We know the coil is driven by solenoid drive 5 (5A). The solenoid drives are listed at the right of page 46 of the manual, and solenoid drive 5 is carried by the grey-green wire, driven by Q42 on the CPU board. Following this wire into connector J1 on the PPB, it passes through a couple of diodes and resistors, and then gets sent to connector J2. This is where all of the coil drive wires connect to the PPB. Solenoid drive 5 ends up at pin 5 of J2, which has no wire connected to the other side of it (no surprise there, they removed it!). So J2-5 is where we have to connect our coil drive wire. 
Playfield coil wiring diagram with VUK coil added.
One important thing to note about connector J2 is that the pin numbers printed on the board are the reverse of those in the manual (i.e. pin 1 on the board corresponds to pin 9 in the manual)! There are lots of errors on these pages of the manual, which is frustrating because they are so important. I used a violet wire to connect to the drive wire lug of the coil (observe correct diode orientation - the power wire should be at banded end of the diode). I zip-tied the wire to the wiring loom and loom support brackets on the playfield, and slipped it into the plastic conduit leading up into the backbox. Then I simply replaced the IDC connector on J2 of the PPB with a Molex type connector, and crimped new terminal pins onto each of the wires.
All done! Testing the coil was simple enough in coil test, and it all appeared to work OK. Below is a slow motion video I shot of the upkicker working for the first time, shooting the ball into an improvised ramp. Installation of new insert lamps
One other oddity at the top left of the playfield are the inserts in the captive ball area (labelled from 10K at the bottom to 50K at the top). If you look at them under the playfield, you'll realise that there are lamps under each of them except for the one at the very top (50K). I have no idea why this insert has no lamp associated with it, but it seems like a waste of an insert! So, I added a lamp socket (RTBB) and connected the socket in parallel with the lamp socket for the adjacent insert. This means that both inserts light up at the same time. Unfortunately there is no lamp matrix position that is coded to light up the 50K lamp individually, so this is the best that can be done.
Note that there is a mistake in the manual on the lamp matrix chart. The chart depicts the lamps in the captive ball lane as lamps 50, 51, and 52. It depicts lamp 51 as being doubled-up in the two middle positions. This is not the case; only one lamp lights up when lamp 51 is active. 
Excerpt from playfield lamp matrix chart.
Installation of Prototype ramp assembly
The prototype game had a wireform ramp at the rear left of the playfield. When the captive ball entered the VUK, it would be kicked onto the ramp, and then dropped back down at the bottom of the captive ball area. It's a cool way to make the captive ball shot more interesting. Below is a photo from Pinsider DugFreez of what the ramp originally looked like.
Original wireform ramp in the captive ball area.
"This would be way cool to have on my game," I thought. However, I can't weld, and I know that if I tried to bend some steel rods to make the wireform myself, it would look dodgy and shitty. So, I reached out to David Morrell of Avid Creations who is making some really cool wireform ramps for various games. I sent him a couple of photos and asked if he would help create this wireform from scratch. He offered to help, but the cost was prohibitive for a one-off without a template to copy. I started to think about ways I could fabricate the ramp simply and cheaply, mainly to test the concept and see if it would work. Then, perhaps, I could get a nice one made once I was happy with the shape and fitment. So, I headed to Bunnings to figure out what I could use. I strolled into the plumbing section and found the perfect solution: PVC pipe and couplings! These things are cheap, readily available, and there are plenty of sizes and couplings for prototyping. I put a few lengths of PVC together with couplings. For the initial test, I simply cut a hole into the coupling at the base of the assembly so the ball could move through. Then, I turned the game on, left it in attract mode and placed a ball into the upkicker. The video below is in slow motion, but shows the result of the prototype test. Success! The ball was ejected from the upkicker and successfully made it out of the ramp exit. Now all I needed to do was mount the assembly to the playfield. Normally, this assembly would be a wireform with mounting points that could secure it to other plastics, or the timber rail on the side of the playfield. However, I wanted to keep this assembly completely independent of the rest of the playfield, to facilitate easy installation and removal (I anticipated I would have to fiddle around with this a lot). I found the simplest method of mounting it to the playfield was via a bolt and nut which secured the ramp to the timber back panel. A plastic spacer between the ramp and the back panel kept it centred above the upkicker hole. I also cut away most of the plastic pipe at the front (entrance) of the ramp, to give a wide hole for the ball to enter the upkicker. I experimented a bit with how to construct the ramp, but determined that a single vertical pipe section above the upkicker with a 90-degree coupling that serves the ball onto a long section of horizontal piping was the most reliable shape for the ramp. I tried some 45-degree couplings and other sizes of piping, but they didn't work as well. Unlike the prototype ramp, I initially wanted to have the ball dropped onto the pizza spinner from the ramp, rather than serving the ball back into the captive ball lane. However, the pizza doesn't spin for long enough for this to work; the ball only gets to the spinner after it has stopped. Instead, I cut the ramp short and simply had it drop the ball onto the green wireform, which delivers it to the left inlane. This works well from a flow perspective (right flipper, to upkicker, to left flipper). Unfortunately, there is no way to position the ramp and leave the Raphael figurine in place, so I moved him to another spot on the playfield. Now I had the mechanical design of the assembly figured out, I moved onto cosmetics. I decided to lean into the use of PVC piping and thought I could turn the piping into a sewer, which fits in perfectly with the theme! I had some leftover yellow spray packs from another project, so I primed the assembly with plastic primer (Bunnings) with light sands prior to each coat. Then, I coated it with the yellow spray paint (Bunnings). To really sell the idea of this being a "sewer", we need some toxic sludge, don't we? Just like the sewer depicted on the playfield. However, to see the sludge inside the sewer, I had to cut away the top half of the horizontal ramp section to allow the player to see into the piping. This also has the benefit of allowing the player to track the ball as it moves through the ramp, instead of being surprised by it when it exits. For the toxic sludge, I used some Createx 5404 Fluorescent Green paint, and brushed it on in thick, heavy layers on the inside of the piping. Then I allowed it to drip into the base of the pipe, to give the appearance of a deeper layer of sludge at the bottom of the pipe. Then, some drips on the exterior of the pipe to make it look extra dirty and toxic. A few coats of clearcoat sealed everything in at the end.
I have a stack of old playfield plastics left over from plastic set replacements and playfield part-outs. I happened to have a plastic from Junkyard (Williams, 1996) in my stack, which depicts an alligator tail sticking out of a sewer grate. The alligator isn't quite a turtle, but it's close enough! I attached the plastic to the end of the ramp to hide the otherwise plain plastic pipe. 
Plastic installed over the end of the ramp.
I think I did a pretty good job in the end. Below is a video showing how the upkicker and ramp work during gameplay. Note the removal of the captive ball and posts in front of the upkicker lane, to allow the ball in play to access the upkicker hole. Installation of prototype game code
Of course, before we can get any prototype features working, we need to have the correct revision of the game's code. Prototype features were all removed from the production version of the game ROM, so you can't just install the prototype mechanisms and hope that they work. There are two ROMs that need to be changed; the CPU game ROM and the display ROM on the display panel. After some searching I found them on VP Universe. For the display ROM, you'll have to remove the display panel from the backbox insert panel. Simply undo the two connectors attached to it, and pop the display board off the plastic standoffs on the front side of the panel. Then remove the ROM with an IC removal tool (Jaycar). It is the only chip in a socket on the board. The CPU has two game ROMs on it at positions B5 and C5. These can be removed in the same way.
Rear side of the display panel. The ROM is at left.
I decided to reuse the display ROM, so I erased the old image, and burned a copy of the prototype code onto it. However, for the CPU ROMs, I decided to combine the two ROM images into a single chip. Most later Data East games used a single game ROM rather than two, and it makes it much easier to swap ROMs and swap CPU boards out when they are all set up to take a single ROM only. Modifying the CPU board to take a single ROM is described on Pinwiki. But all you need to do is remove jumper J5 and install jumper J4. 
Removal of jumper J4 and installation of jumper J5.
I stuffed around for a little while here because I burned the prototype ROMs incorrectly. The above link to Pinwiki outlines the command line prompts used to copy the ROM files and this procedure works well. However, I forgot that TMNT used a 128 Kb ROM at B5, and a 256 Kb ROM at C5. I burned the combined ROMs onto a larger 512 Kb EPROM, but the CPU would not boot up. I then realised that the 128 Kb ROM image would need to be copied twice in order to properly fill up the space on the new 512 Kb ROM (128 + 128 + 256 = 512 Kb). Once I did this, we were good to go! To make life easier for you, you can download the combined ROM file below and burn it directly to a 512 Kb EPROM such as an M27C512. 
If all goes well, you should be able to boot up the game and see the following splash screens displaying the prototype version numbers. Interestingly, the CPU ROM version numbers have "VUK" in their name.
Installation of prototype topper beacon light
Perhaps the most noticeable new feature of the prototype machine is the green rotating beacon light that was supposed to be on top of the backbox. I think interactive toppers are way cool, so it was a real shame that they cut this from the production game. To date, I have not seen a photo of an original beacon. All of the photos of prototype games seem to depict the plastic that goes around the beacon lamp, but not the beacon itself. Damn. So at this point, we don't have many clues as to what this beacon should look like, or how it is powered. I've only ever seen the beacon in one of Todd Tuckey's factory tour videos, and even then it was obscured by a blanket.
Lethal Weapon 3 (Data East, 1992) was made only a year after Teenage Mutant Ninja Turtles and it also featured a beacon on the topper. I assumed that this beacon was designed very similarly to the one that was meant to be on Turtles. We can tell a few things about the beacon from the Lethal Weapon 3 manual. 
Parts diagram for the rotating beacon light on Lethal Weapon 3.
I had a look to see if I could buy the entire Lethal Weapon 3 assembly, or parts thereof, to retrofit into Turtles. No such luck; it is not available anywhere and most Lethal Weapon 3 owners have replaced their missing beacons with more readily available replacements. So, I set about finding a replacement as well. A lot of Lethal Weapon 3 and Rescue 911 (Gottlieb, 1991) owners replace their broken or missing topper beacons with ones made by Wolo Manufacturing. These are drop-in replacements that can be wired up to the same connector as the original beacon. These are a great option, however, the Turtles beacon is meant to be green and Wolo does not sell beacons of this colour. Off I went to Aliexpress, where I found rotating beacon lamps were readily available. Green, red, blue and amber seem to be the predominant colours for sale. A few styles are around, but most of the lamps look the same. I grabbed a 12-volt DC version (without a buzzer) as this is the easiest voltage to tap into in the backbox (AliExpress). Of course, I also grabbed a green LED (AliExpress) to replace the incandescent globe that comes with the beacon, as the incandescent is not very bright. 
New beacon lamp and bright green LED to go with it.
The base for this beacon light has a number of mounting holes, so I was able to screw it directly into the top of the backbox. I also started looking at how I could connect it to power. Referring to the Lethal Weapon 3 manual, I found that the beacon (Mars light) was connected to power at CN4-6/7 on the power supply board. On TMNT, this is an 18-volt power supply pin meant for lamps. Even better is to connect it to one of the 12 volt circuits, as this is the supply voltage the beacon needs. Conveniently, one of these is accessible via CN6-3 on the power supply, which is normally vacant. Splice a new wire in and connect it to one of the wires for the beacon light. Half way done! 
Purple power wire for the beacon lamp spliced into CN6-3 on the power supply.
As I was adding this beacon to the regulated 12 volt circuit, I needed to consider whether the operation of the lamp and motor would put too much load on the fuse and blow it. I did a test with an incandescent lamp compared to an LED, and found that the LED drew significantly less current (0.09 compared to 0.54 amps). 0.09 amps is negligible, so there aren't any problems with adding the beacon to this circuit.
On Lethal Weapon 3, the beacon is then driven by transistor Q25 on the CPU via CN12-7. While this seems simple enough, Q25 drives flash lamps on the playfield on Teenage Mutant Ninja Turtles, so it was already "occupied". I couldn't find a vacant driving transistor anywhere that I could reasonably assume would have been used to activate the light. So while power was easy, I wouldn't have access to any driving circuitry to actually ground the light and turn it on. 
Wiring diagram for the beacon light from Lethal Weapon 3.
I began searching for another way to activate the topper light. I considered what powered devices on the playfield I could potentially hook into to turn the light on during certain parts of the game. The perfect playfield mechanism for this is the pizza spinner. It spins constantly during multiball, and occasionally during normal gameplay when you hit a ramp or activate a feature. These would be perfect situations for the topper light to activate as well. Based on the wiring diagram, there are a couple of voltages and currents associated with the pizza spinner. 
Wiring diagram for the pizza spinner motor.
The motor is activated via a relay board under the playfield. Q13 on the CPU board grounds the relay, which itself is supplied by 32 VDC from the power supply board. This relay then switches on the motor, which is supplied 28 VAC from BR2 (left side of the backbox).
So I decided to use the same Q13 driving transistor to drive the beacon light. I hooked up the beacon light to Q13 by splicing it into the CN19 connector, pin 9 (blue-black wire, which supplies the relay). As soon as I did this, the beacon and the spinner both activated and stayed on. I realised that I was actually grounding the relay that controls the spinner motor via the beacon light. The solution to this was to install a diode, which blocked current travelling from the relay through the beacon motor. Install this diode on the wire connecting CN19-9 to the beacon, with the banded side of the diode facing the CPU board, and the non-banded side facing the beacon. 
Beacon connected to CN19-9 (shared with blue/black wire) with a blocking diode.
Out of interest, I looked at the solenoid driving transistors on the CPU board again. Surely one of these had to be programmed to activate the beacon light, right? In particular, I noted that there were three coils listed as "not used" in the test menu. One of these, 5A (Q42), was for the upkicker mechanism (see section above). But that left coils 3A (Q44) and 6A (Q41) as vacant.
So I hooked the lamp up to both of these driving transistors and watched what happened during a game. Nothing! I'm not sure what the original purpose of these transistors was during development, but they don't appear to be related to the beacon light, or anything else on the playfield at least in the prototype version of the game code. Either way, the beacon is now fully wired up and goes off whenever the spinner goes off during gameplay. It's awesome! It looks even better with the plastic surround and feature lamps installed; check the next section for that! Installation of prototype topper lamps
As well as the beacon light on top of the backbox, the prototype games also had two lamps on either side of the beacon. These illuminated two inserts, which were marked "50K Awesome Scoring" and "100K Super Awesome Scoring". As you might have guessed, these lamps lit up during the Awesome Scoring and Super Awesome Scoring modes, respectively. A photo by Pinsider Tsskinne shows how the prototype topper looks, with both lamps being mounted to a larger plastic that was designed to surround the rotating beacon light.
Prototype topper plastic.
Fortunately, the photo also gives us a clue as to how the lamps are wired up. We can see that the wire colours going to the lamp on the left are YEL-VIO and RED-GRN. Whereas the wire colours going to the lamp on the right are YEL-VIO and RED-BLU. Now, initially I suspected the lamps on the topper to be flashers. But that is not the case, as these wire colours are for standard feature lamps. Flashers, on the other hand, would have an orange power wire. With the wire colours in mind, now we can look at the lamp matrix to see how these lamps interfaced with the CPU board. YEL-VIO are lamp matrix column wires (column 7). Following this column down, we find two "Not Used" lamps in the column. Hmm. If we look along the row for the wire colours associated with these "Not Used" lamps, we find that they are RED-GRN and RED-BLU. Now that's a a little suspicious! It turns out that these "Not Used" lamps were indeed originally supposed to be the Awesome Scoring and Super Awesome Scoring lamps! Note: the 50K Awesome Scoring is the lamp on the left (RED-BLU) and the 100K Super Awesome Scoring is the lamp on the right (RED-GRN). 
I rigged up a lamp socket with a diode on one end, and connected it to CN7-8 and CN6-6/CN6-7 via test leads. I activated each of the scoring modes, and the lamps started to flash as expected. They also flashed in sync with the rest of the controlled lamps during attract mode and other game features. Cool! That confirms that the lamps are still programmed in the game's code, but they have just been "deleted" from the lamp matrix. At this point I simply wired the lamps up. The YEL-VIO was the column drive wire for both lamps, so I ran it from CN7-8 to one lamp, and then piggybacked the other lamp onto it. To make this a clean install at the CPU board, I soldered my new wire to the end of the YEL-VIO wire already in CN7-8. I then punched the original YEL-VIO wire back into the connector. For the red row wires, I opted to wire these wires to controlled lamp sockets that were already on those matrix rows. This saved having to mess around with connectors on the CPU board. Luckily, there are two lamps within easy reach on the speaker panel which are on the RED-GRN and RED-BLU rows. So, I piggybacked wires onto those lamp lugs and ran them up to the top of the backbox, along with the new wire attached to the YEL-VIO column. The plastic grille covering the vent holes at the top rear of the backbox can be pulled up a little to push the wires through to the exterior of the cabinet.
Choosing how to mount the Awesome Scoring feature lamps was a little more difficult. On the prototype games, it appears the lamps were shrouded by a small metal box at the rear of the topper plastic. I don't know why this metal shroud was necessary, as it wouldn't make any difference to the amount of light distributed through the lamp covers. It was probably more a safety feature to protect the lamp wiring. I opted to ignore it, and mount the lamp sockets directly to the rear of the topper plastic with nuts and bolts (see next section for details on the plastic). Each of the lamps was covered with a red light dome (RTBB). As discussed above, the wiring to each lamp consisted of the common column drive wire, a row return wire, and a diode in series with the row return. I used some spare bayonet lamp sockets for the lamps, with mounting brackets that allowed me to use the same bolt holes that secured the domes to the topper, allowing me to minimise the number of holes I had to drill. I added some connectors for the beacon lamp and the awesome scoring lamps to finish it off. 
Topper lamp wiring.
Fabrication of Prototype Topper Plastic
As the image in the section above (Installation of Prototype Topper Lamps) shows, there is supposed to be a large plastic on top of the backbox that sits on top of the rotating beacon light, and houses the Awesome and Super Awesome scoring lamps. The production version topper (RTBB) is similar to this, but has been cut short so that only the topmost section of the prototype graphics are shown.
For those with more money to splash around, buying a topper for the Stern version of Teenage Mutant Ninja Turtles (Stern, 2020) might be an option (Marco). However, having installed the missing rotating beacon light and the feature lamps, I decided to go the whole hog and recreate the original prototype plastic as well. This involved a bit of graphic design work. As a first step, I scanned in my reproduction topper plastic and used it as a base to create the rest of the topper.
Designing the rest of the topper was a task for Fiona, as she is a whiz in Photoshop. As the bottom half of the topper is mainly simple colours and text, it actually wasn't too difficult to draw the missing elements in including the holes for the lamps, and the text above and below the lamps. It took some time to find an appropriate font for the text, as I did not want to use the original fonts used on the prototype topper because they looked so terrible and uninspired. There are a few replica fonts available online which match the original TMNT style, and I settled on one aptly named "Turtles" (FontBolt). We also adjusted the dimensions of the hole in the centre of the plastic to make sure the new beacon light would fit in the space under the topper "arch". The original topper design also has some weird colouration going on around Donatello's mouth. He seems to have a bottom lip - or something - that isn't coloured. So, we fixed that up so his mouth looked normal again. However, I still thought it was missing something. It seemed weird to me that there were only two turtles featured on the topper. Surely they should have all four, right? I looked for a while to find some reference TMNT images online that I could use as a template, but it appears the artwork on this machine is completely custom drawn, so there wasn't a suitable reference which matched the art style. So, I simply copied Donatello and Michelangelo's heads, and changed the colours of their bandanas to create Leonardo (blue) and Raphael (red). For Raphael, I also changed the shape of his bandana so it looked a bit different from the others (we can't have too much repetition now, can we? The next step was to have the design printed onto a piece of plastic. I decided to utilise the services of a local sign shop for this (Stylz Signs). The end result was very good, and is reverse printed vinyl onto 3mm polycarbonate. It looks quite similar to an original, silkscreened pinball plastic. Cost was also quite reasonable (~$70). 
Final topper plastic after printing.
I opted not to use the original mounting brackets (RTBB) for the topper as my topper was much bigger and heavier, and needed more support. So, I used a larger angle bracket and secured it with two nuts and bolts on each side. This type of bracket was good because I could adjust the topper up and down until it was sitting at just the right height for the player to view all of the text on the topper. 
Brackets securing the topper to the top of the backbox.
I think the final result is pretty impressive, and looks reasonably "factory". 
Complete topper assembly with beacon light, plastic surround, and feature lamps.
Replacement of manual plunger with launch button
I wrote extensively about some of the issues with the ball plunger and turboboost kicker assembly in a previous blog post.
I decided to do something about it. I decided to do something extreme. I decided to remove the ball plunger entirely. I know that sounds extreme, but hear me out. The main problem with this system is that the interface between the plunger rod and the kicker arm is contentious. The kicker arm needs enough space to have its largest range of motion so it has enough power to kick the balls out of the shooter lane. On the other hand, the shooter rod needs to be long enough to make good contact with the rear of the kicker arm so that manually plunged balls will have enough power to make it to the top of the skill shot ramp. This is a hard balance to achieve, and can involve a lot of fiddling around with both assemblies to make things work. But if you think about it from a game mechanics perspective, what is the point of having a manual plunger anyway? There is no skill shot to achieve with a manual plunge on Teenage Mutant Ninja Turtles. The only skill shot involves hitting the ramps once the ball reaches the right flipper. Whether the ball is manually or automatically kicked onto the playfield makes no difference. So, on that basis, I decided to experiment a bit with removing the plunger and having the turboboost kicker as the primary means of serving the ball onto the playfield. This relies heavily on one aspect of the latest (version 1.01) game code, where the game will detect when manual plunges fail to send the ball up the skill shot ramp. Once this happens the turboboost kicker will activate to get the ball into play. This is triggered by the ball moving away from the shooter lane switch, and then falling back onto it, twice (i.e. two failed manual plunges). Without recoding the game's ROM, we can trick the game into thinking that two manual plunges have failed, which will then trigger the turboboost kicker. At this point, the turboboost kicker is functioning like a standard autoplunger. To do this, I wired a launch button switch in series with the shooter lane switch. 
The shooter lane switch, connected to GRN-RED and WHT-BLU wires. WHT-BLU has been disconnected in this photo.
You can splice this new switch into the column drive wire (GRN-RED) or the return row wire (WHT-BLU), it doesn't really matter. The important thing to note is that we want the new switch to be normally closed instead of normally open. So, make sure you solder the wires to the normally closed (NC) terminal of the switch. We don't need a diode here either, as this new switch is not really going to be a part of the switch matrix. It is just going to make/break the connection to the shooter lane switch. 
Wiring diagram for the new launch button switch.
The logic is that when a ball is sitting in the shooter lane (shooter lane switch closed) and we press the launch button (normally closed, then open when pressed), the game will think the ball has left the shooter lane as the launch button will manually "open" the shooter lane switch. Do this twice, and the game will think a manual plunge has failed twice, and will automatically launch the ball. Most importantly, the shooter lane switch will still function as intended during multiball, ball save, etc. because our new switch is normally closed. There is some minor fabrication to be done here. Since I wanted to use a standard launch button assembly (RTBB, PSPA), I had to cut a small piece of sheet metal that would replace the mounting plate for the shooter rod. I then cut (poorly) a hole in the plate so I could fit the body of the launch button through. It worked pretty well, and the button is just big enough to cover the hole in the cabinet completely.
An important thing to note is that this button will not work with the prototype code. Only the production version of the game code (1.01) has the "failed launch" logic which will kick the ball out of the shooter lane after two failed plunges. The prototype code (A 0.7) will not do anything in this situation. I played several games to test the new launch button and I think it works well. However, it is annoying that you need to press the button twice to actually launch the ball. I'm sure there is a small amount of componentry that could be installed to register two switch openings instead of one. However, I wasn't keen on taking the mod that far. This is a potential solution for operators with games on location who do not want to bother with maintenance of the shooter rod. This mod does require that the turboboost kicker has to be working well. Potentially, upgrading the mechanism to the new, Stern style of autoplunger would also be a good mod to do, too. I decided not to go that far with this one, but might try it at some point in the future. I had to do one last thing to make this autolauncher as reliable as possible. Occasionally, the autoplunger would strike the plastic apron on the left side of the shooter lane. This would cause it to be weak, and resulted in balls getting stuck in the shooter lane. There was a small amount of slop in the assembly which allowed the armature to lean a little to the left. The solution was simply to shave some of the plastic material off, so that the autoplunger armature could move freely. Be careful not to cut too much away, as there still needs to be enough material to stop a ball falling past the apron. 
Autoplunger armature leaning too far left.
Flipper rebuilding
The flippers hadn't been serviced in a long time. However, they still had plenty of power to make all of the necessary shots in the game, and the correct coils were installed. There were just a few things to address. First, the flipper bats. One was cracked in multiple places, so I replaced both bats and shafts with some spares I had on-hand. These ones also had the DE logos on them (RTBB, PSPA), so bonus points to me for accuracy!
Next, let's take a look under the playfield. The flipper mechanisms were... odd. Data East games of this era typically have a barrel/compression spring over the plunger, which returns the flipper to the rest position. These flipper assemblies had an extra bracket bolted to the coil stops, with an extension spring that attached them to the flipper cranks. This is brilliant. It basically upgrades the flipper assembly to the newer style that Data East used where the compression spring on the flipper plunger was replaced with an external extension spring, which was more "snappy" and less prone to breaking. I actually decided to keep using this current system. But, I moved the mounting position of the extension spring so that it was closer to the flipper shaft. This was because the springs were under quite a bit of tension when at rest, which means the flippers were weaker than they should be because they had to overcome the high tension of the spring. I installed 10-364 (RTBB) extension springs, and drilled a hole into the mounting plate for a metal screw to hold the spring. This puts less tension on the spring when it is at rest, and gives the flipper its original power. Williams flipper return springs are about 45mm long when at rest, so I tried to replicate this distance on the Data East assembly. This aligns with the halfway point of the flipper coil, coincidentally. The plunger links and bushings were well worn, so I replaced these as well. I opted to replace these with equivalent Williams parts as I did not have the correct Data East parts on hand, and they worked very well.
I also installed some connectors so the flipper mechanisms could be removed from the game and repaired on the bench. 
Connector installed on right flipper assembly.
One part I did have to replace were the coil stops. These were very worn. However, new Data East coil stops are about $10 each from most local pinball parts suppliers, which I wasn't keen on spending. They are not interchangeable with Williams coil stops, as the bolt holes are larger. So, I decided to refurbish the coil stops by drilling out the worn cores/slugs, and replacing them with Homepin coil stop replacements (part no. HCS-01). Homepin doesn't seem to make these anymore, but I have a stack of them from when they were first released. A very useful product! I drilled the rear of the coil stops with a 9.5mm drill bit until I was flush with the coil stop bracket itself, and the core was beginning to get loose in the hole. I drove it through with a hammer and punch and it popped right out. The new coil stop was simple to install and was the correct length when finished. Much cheaper than buying a whole new bracket (about $10 per coil stop).
One other part of the flipper system needed attention. The right flipper switch was flaky. I tested it with a meter and found that, unless it was fully depressed, it would not register a proper switch closure. In some cases there would still be relatively high resistance across the switch terminals (up to 50 ohms), or the switch would close and open in rapid succession. These scenarios would result is failure to flip or a rapidly machine-gunning flipper. Not what you want in the middle of a game, so the switch needed to be replaced. Any large style microswitch will work for this application, as long as the actuator can be bent into the correct orientation.
Pizza spinner refurbishment
The spinner in the middle of the playfield is awesome fun. Unfortunately, however, it usually looks like crap. The colours on the decal fade over time and as they get dirty, leaving it looking dull. Luckily, replacements are available (RTBB). The original decal is stuck down with a strong adhesive, so I used a razor blade to pry it off. Once clean, the new decal went on and looks much better!
The spinner was also making a little noise when turning. Not a lot, but it was enough to make me remove the spinner from the motor shaft and inspect the gearbox. The gearbox on this assembly has an open case. I didn't see much grease left on the gears. Otherwise the gearbox looked fine, so I cleaned off as much of the old grease as I could and applied some new grease (Nulon). The motor spun more quietly after that, so the lack of grease must have been the problem. Surprisingly, these motors are still available to buy if they need to be completely replaced (Marco).
Installation of Additional Turtle Figurines
Back in 2019 I attended the Collector Con Toy & Hobby Fair. I took a couple of pinball machines along with me, which was great for exposing the general public to pinball. As it was a toy and collectables show, I also found a set of three original Playmates-branded ninja turtles figurines; the exact same as the ones found in the pinball machine! Unfortunately, Raphael was missing from the set, but I still had Michelangelo, Donatello and Leonardo.
Luckily, I kept them, because now I had a machine I can actually use them on! I don't normally like to cover playfield with figurines and mods, but I think these are a worthy exception because they are original pinball parts! I had already moved Raphael because his original position above the captive ball was no longer tenable due to the new upkicker and ramp assembly I had installed. He now lives at the top right corner of the playfield, above the skill shot ramp. There is a clear piece of plastic covering the ramp which is a perfect mounting spot; just drill two holes in the plastic and screw his feet in from the underside. 
Raphael's new home above the skill shot ramp.
It made sense to put Michelangelo above the amber standup targets, because those were the Michelangelo targets, right? But it seemed weird to have two Michelangelos so close together (the original one is on the lower pop bumper). Instead, I put the second Michelangelo at the bottom left of the playfield above the left outlane. I had to squish him down a bit to get him to slide under the glass, but he fits. I screwed his feet into the timber siderail on the playfield. 
Michelangelo above the left outlane.
I put Dontatello above the amber targets. There is a metal snubber plate that sits above these targets which is a perfect mounting point. I drilled a hole in the plate and put a screw up through it to screw into Donatello's foot. His left hand can be stretched backwards to grab the green wireform ramp for stability. 
Donatello mounted to the metal plate above the orange targets.
Finally, I put Leonardo above the entrance to the yellow ramp. I removed the hex nuts from the posts at the ramp entrance, and drilled holes in Leonardo's feet. Then, I could simply push his feet onto the post threads. There wasn't enough thread to reinstall any nuts, but this doesn't matter as his feet grip onto the threads so tightly. 
Leonardo's new position above the ramp entrace.
Coin door and coin mechanisms
Single slot coin door (after removal from the cabinet).
The game came with a single slot coin door. Now, single slot coin doors are usually an indicator of an "export" game. Single-slot doors typically come with a single electronic coin mechanism (usually a Coin Controls or Sentinel branded unit) which is programmed to take multiple coin denominations in the currency of the country the game is destined for. Australia was one such country, so I was expecting to see an electronic coin mech on the rear side of the door. However, this was not the case. Instead, I found a single mechanical coin mechanism on the rear of the door. It was a 20c mechanism, which matched the label on the coin slot. I had done some switch tests earlier and realised that coins were not registering properly in switch test, too. 
Rear side of the coin door with mechanical coin mech installed.
First, I worked on figuring out why coin drops were not being registered properly. I discovered that when a coin was dropped through the mechanism, it did actually accept or reject the coin as expected, but the wrong switches were being triggered when the coin switch was activated. I was getting a coin switch occasionally, but other times it would register a playfield switch. Weird! I had a look at the connector which the coin switch was connected to. 
Coin door connector.
The problem here was the coin switch was connected to the wrong pins in the connector. The connector is a seven-pin connector, with the following pinout:
Annoyingly, the manual does not show switch column 1, so the green-brown wire doesn't actually appear here. But it goes to the coin door to service the coin switches. The coin mechanism, as I found it, was connected to pin 2 (ground) and pin 6 (switch return). So, when the switch was closed, it was grounding the switches on that row, which isn't right. Instead, the wire in pin 2 should have been connected to pin 4, which properly completed the switch matrix. At this point it started registering coin drops properly. Now, the way the coin mechanism was connected was a bit weird. Individual pins were inserted into an IDC connector. This is strange because the game would have shipped with an electronic coin mech, which plugs into an interface board. This interface board sits on the coin door, and is connected to the switch matrix by seven-pin connector that the coin mech was currently directly connected to. My game was missing the interface board, which is not necessary if you are using mechanical coin mechs, as they interface directly with the switch matrix without the need for an interface board. 
Electronic coin mech interface board as normally found on Data East coin doors.
Now, having a mechanical coin mech with a single slot coin door is a bit silly, because the game can only take one denomination of coin. This coin mech was a 20 cent mech, which was also ridiculous. When was the last time someone paid 20 cents for a game of pinball? So, I had two options: install an electronic coin mech and an interface board, as was originally intended, or install two mechanical coin mechs on a two-slot door. I had a spare two-slot door, so I decided to go for the second option. The spare door was a Williams WPC era door, so it would fit into a Data East cabinet. The door also had two coin mech brackets installed. One held an electronic mech, and other held a $1 coin mech. I removed the electronic mech, and installed a spare $2 coin mech that I had. Now, I had a means of accepting $1 and $2 coins. I wasn't worried about 20 cent coins, otherwise I could have opted for a three-slot door! 
Test-fitting the two-slot coin door.
The wiring harness was fairly simple to swap from the old door to the new one. I had to undo a couple of zip ties to lengthen the harness a little, as the slam tilt switch is further away than on the Williams door. The other modification I had to do was wire in the second lamp socket, as my original wiring harness only accommodated a single lamp. This just needed a couple of crimp terminals and some wire, and it was good to go. Otherwise, no modification of the door itself was needed. The hardware mounted to the existing holes on the door, and I had enough space with the wiring loom to move things around to where they would fit. 
Rear view of the two-slot door after transfer of all components to it.
There was only one thing left to do: print off some new coin slot entry labels. You can download a template from Inkochnito's website and edit it as you see fit. I printed off some new $1 and $2 labels to match the coin mechs on the door, and put them in. Even on regular paper, these labels look OK. After some testing I realised that the $2 coin mechanism was not properly registering coins inserted into it, even though these coins were passing through the mechanism successfully and were not being rejected by it. I double-checked the switch, and confirmed that it was OK. So why was it not registering? I took some slow motion footage of coins passing through the mechanism and realised that, due to the small size of the $2 coins and the relatively large "accept" slot that they pass through to trigger the switch actuator, the coins were sometimes falling too far forward. Therefore, they were missing the switch actuator, or not pushing it down far enough. This actuator cannot be adjusted very much, as there is a metal cutout in which it sits. To counteract this, I installed a long bolt through this part of the mechanism, such that it would narrow the width of the exit slot and "push" the coins into the path of the switch actuator reliably. Once this was done, coin registrations were 100%.
Ball Guide Repair
I needed to repair one of the ball guides, too. The "spade bolt" at the front of the ball guide had snapped. I drilled out the rear side of the rivet until I could easily tap the rivet out with a punch. I had a spare spade bolt which I could install (RTBB, PSPA). Ideally you should rivet these back on, but I used a small, flat-headed screw and a nut to secure the post. It's important to check that the screw you use for this doesn't protrude into the path of the ball. Most screws won't work for this purpose as they stand too proud. The nut doesn't matter too much, as it behind the ball guide and not in the path of the ball. This is why I keep stock of a bunch of smaller and specialty screws that I can use for this type of application.
Hang-ups in Ball Trough
Occasionally, the ball would enter the outhole, the outhole coil would fire to send the ball into the trough, but it would get stuck somewhere. The left trough switch never got activated, and I confirmed the switch was working properly. So, where was the ball going? I removed the apron to get a look and I found the problem straight away. The ball trough runway was seriously bent, which was causing the ball to get stuck. Bending the metal runway back to its original shape got everything working properly again.
Reassembly One of the irritating things about reassembling early Data East pop bumpers is that each pop bumper lamp is soldered to one of the general illumination circuits. The wires need to be desoldered or cut in order to remove the socket from the pop bumper body. To make this process easier if it needs to be done again in future, I installed a screw terminal block which I soldered to the general illumination circuit. The pop bumper lamps are simply inserted into the other side of the terminal block, and can be unscrewed and removed as necessary. Regular connectors won't work here, as the wires penetrate the pop bumper mounting bracket through a very small hole, barely big enough for the wire.  Terminal strip connecting the pop bumper lamps to the GI circuit, to make disconnection easier. I also installed a playfield protector (Playfield Protectors) before reinstalling all of the playfield mechanisms and parts. The fitment of this protector needed some adjustment, which involved cutting some edges of the protector back so they wouldn't interfere with switches and posts. In particular, I noticed that the protector sits very close to the standup targets, and in some cases it abuts them. This is a problem because there is basically no clearance between the bottom edge of the standup targets and the playfield surface, so the protector can get jammed here and push the standup switches closed. I found the purple and red targets to be the most affected, so once I cut this protector edge there were no more phantom switch closures. A couple of small cuts had to be made around the pop bumper area as well, so that the protector fit nicely.
Half of the playfield uses metal posts to support plastics and other playfield mechanisms, while the other half uses fluorescent green star posts (PSPA). I think the star posts are way cooler then the metal posts, so I replaced all of the metal posts with green star posts. There are still a couple of places where you'll need to retain the original metal post so that other playfield parts can be installed or so that ball movement isn't negatively impacted. One of those locations is directly to the left of the captive ball. A plastic star post blocks the shot to the captive ball. Another spot where a metal post may be necessary is to the right of the double ramp entrance. I was able to squeeze a plastic star post into this spot but it was very tight. Your mileage may vary. The rest of the metal posts can be replaced with no ill effect. I think this makes the game look way brighter, and as this is the only game to use fluorescent green star posts, you may as well go nuts with them!
Conclusion Well, this is probably my most involved machine restoration to date. Lots of repairs and fixes, and lots of upgrades and fabrication of new parts. I've not had to reinstall prototype parts on a machine before, and it was an interesting learning experience to get everything working 100%. I love seeing a machine brought back to its former glory, but it's even nicer to see a machine be upgraded with prototype features that were removed prior to production. A lot of people think Teenage Mutant Ninja Turtles is an average game, and to be honest I do agree that most Data East games of this era are pretty simple. But the new ramp adds something new to the game that most people haven't seen, and the topper makes it unique. So, personally, I'll be keeping TMNT in my collection for some time. It's a simple game for kids and people new to pinball, so it's great to take to shows and other events where casual players may be present. I've since taken it to several shows this year, and it has always been a big hit with kids in particular. I would like to have a few Data East machines in my collection, and I think TMNT was a great way to start!
0 Comments
Leave a Reply. |
About
Here you will find logs of our pinball and arcade machine restorations, repairs, discussion about general pinball and arcade topics, as well as recounts of our random pinball adventures.
Check back regularly for updates! Blog updates
Archives
November 2024
Categories
All
Donate
Running this website is a hobby for me (just like pinball!). I like being able to show off my restoration work so everyone can learn from it and potentially fix their own machines. If you enjoy reading the site's content or it has been helpful to you, please consider donating to offset some of the website's operating costs. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
RSS Feed