Reproducing the Konami 005273 Resistor Network

Motivation

I came across a Teenage Mutant Ninja Turtles (TMNT) board recently. The TNMT games on Konami boards are very iconic both in their 4 player and 2 player forms. 

According to the listing the board was "Tested, working" and the seller even did a retest prior to sending which reported on "Retested, working". So all great... until it arrived. 

Upon testing, I realized the inputs were not working properly. Buttons were mostly okay, but only the left direction was working, and the same afflicted the input 2. I thought that it could be some sort of non-standard JAMMA input mapping, and even replaced the EPROMs with the 2 player version of the game, but problem persisted. 

Understanding the problem

Finally I decided to take a look at the schematics. On page 17 of a PDF manual I downloaded, I found the input section, which looked like this:

Okay, so it looks like all inputs go through a component labeled 005273, which are labeled RN1 to RN10. And guess what? it seems to be grouping together similar inputs for player 1, 2, 3 on each component, for example, RN3 has UP inputs for player 1, 2, 3 going through. Then, I inspected the board and...

"Well, there's your problem"

So RN1 to RN4 are missing. That's exactly where 1P to 3P Up, Down, Right are mapped to. RN1 and RN6 are Player 4 exclusive, so that means I could move RN6 to an empty slot and then I would need 2 more empty positions filled so I could play the 2 player game without issues.

Coming up with a solution

So I only need 2 but here lies the problem: 005273 is a Konami custom chip, and obviously cannot be bought easily. The solution is to rebuild it then, and for that, I searched the Internet for schematics but none were there to be found... I could find a few pictures of reproductions that already existed, but none had component values, etc. So I would have to do this on my own. 

In comes your run-of-the-mill component tester that you can get anywhere from ebay to Amazon. These devices while not the most accurate, military grade, certified calibrated thing around, are very useful and versatile. So off I go to start testing.

From the pictures I found on the internet I could see that there were capacitors involved, so I decided to test every leg and see what it would produce. One thing I did first was to test each pin on an RN slot on the PCB to see if they would "beep" to ground or power. That's pin 9 (Vcc) and pin 10 (GND). 

Now, looking at the schematic I could see that there are 4 inputs and 4 outputs, so there has to be some connection between pin pairs 1-2, 3-4, 5-6, and 7-8. Looks like the odd pins are outputs, and even pins are inputs. 

Measurements

So here was my testing method: Test every leg against Vcc, every leg against GND, and then each leg pair. Since JAMMA inputs are with negative logic, that is, 0v = on, 5v = off, there has to be pull-ups for each input, so most likely Vcc connections will be resistors, and the only reason to have GND on the component would be to use filter capacitors, so most likely these are the components connecting between the legs and ground. 

Okay so top row gave me 350 nF between ground and Vcc, so that's probably the power filtering capacitor. the other leg-to-legs all returned something around 220Ω. So likely a resistor between those. 

On row two, each leg on even pins gave me also around 350 nF indicating a series capacitor with those pins, and similarly, 3.3KΩ on the odd pins indicating series resistor for them. 

On row three (excuse my getting a bit lazy on the annotations) the even pins to Vcc gave 3.52KΩ, indidcating the path with the 220Ω and then 3.3KΩ resitors, similarly tests in between legs of the same type game me doubled values, which again indicates two components of the same type on the path between them. I did some other exotic measurements to confirm the results I was coming up with.

With all of that information, this is the schematic I came up with:

Which sort of makes sense. One pull-up resistor on each output after a RC filter on each input. 

Prototyping the solution

To test, I figured I could cut some corners. For instance, all capacitors would go, and so the 220Ω resistor between input and output, which would basically just end up in a jumper for each pair with a pull-up resitor. I used 4.7KΩ because it was what I had in hand at the moment. 

It's not elegant but it proved to work just fine

Doing it right

Well I obviously don't see that as a final solution, and at this point, I figured I was way deeply invested into this to just buy a replacement online. Time to design my own PCB. 

There isn't much to it, really just the same connections as the schematic. I managed to get all components on the same side, so the other side just holds Konami 005273. 

With a black solder mask it fits right in.

Tiido suggested me to design with the little teeth, so they could be soldered directly to the vias, but they can also have a header soldered to them for a firmer bond. As a note, I later reversed the design so that the components are all on the bottom side and the logo would be facing the same direction as the original components. 

I haven't manufactured these yet and I don't plan on selling these if I do. If and when I make them, I will post some pictures of the end result if there are some interest. I've also considered getting some dipping epoxy and coating it so that it looks even more genuine. 

But why is this component like this?

Why the complexity and not simply a resistor array for the pull-ups? Well this is made as such to address noises that the input harnesses can pick up, especially when the wires are long. So each input goes through a RC filter and there's an additional power filter for the pull-ups. I obviously didn't know much about arcades so this all a journey of discovery for me. Glad you're coming along. 

See you on the next post. 😁👍