Sega Genesis/Mega Drive 6 button pinout?

[Segasonicfan]

I've been reading a variety of boards and posts regarding the 315-5638, as well as controller convertors and cross-compatibility, and I have a few thoughts I felt were worth resurrecting this thread with.



1) Sharp X68000 in Reverse

If the x68k can recognize 3 and 6 button Sega pads by swapping a few output pins (the 2-button pad even uses the same off-the-shelf multiplexer IC that a Sega 3 button), wouldn't it make sense to reverse the scheme and see if the 6 button x68k controller works on a Genesis?

https://gamesx.com/wiki/doku.php?id=controls:x686button

2) Decap Attack

Preservationist, reverse engineers, and even Russian modders have been making their way though Sega's 315 series custom silicon via decapsulation. No one yet that I have found has gone after the 5638 (so far it has mainly been arcade board chips). I did some early reading on this today, and the cost seems to be between the cost of protective equipment + nitric acid for DIY and ~$150 plus postage for a professional service (minus the cost of donor 1653 controllers).
Would it be useful to spring for a decap?

3) CPS

While i could not find sufficient PCB images to know for sure, it stands to reason that the Mega Jet also uses this chip for its onboard controls, Ala the Nomad. All my research shows that every other 6 button compatible controls use different chips. Most are epoxy blob crap, some more advanced pads have bigger chips, but they incorporate turbo functions (including Sega's own offering.) One decent exception I'm investigating is the Capcom Pad Soldier GS/MD. I found images of the SNES pad, and that used an ASCII clone of the IC in a SNES pad in a nice normal sized package. Couldn't track down PCB photos of the Sega or 3DO versions, so I put in an offer on eBay for the Sega variant and i have it on the way.


It is 2025, and cannibalizing original controllers for these ICs is simply a poor option. Unfortunately the only confirmed alternative (Verilog implementation on CPLD) uses the EOL EPM7064 part, and the code hasn't been tested with the newer ATF1504 replacement. Additionally, the code would need to be run through a conversion tool called POF2JED which may not even succeed in translating the Verilog code correctly.


So, am I nuts, or do any of these thoughts make sense to the more technical-minded of you? I'm not a skilled engineer, nor a software programmer. I just push my soldering iron into things and sometimes they still work afterwards.

Thanks for the super detailed post!
Unless I am missing something though, I do not see what your goal is from this effort?  Is it simply to know more about the 315-5638?  IIRC its a fairly simple IC, and can actually be emulated via microcontrollers nowadays (although CPLD is more ideal for lag considerations).  I've thought about reverse engineering it completely with high-speed logic chips, but demand hasn't been high enough yet.

What is it that you want to use this for / accomplish?

[leftypem]

Ideally, I'd like to come up with an inexpensive way to implement the functionality without using original Sega chips. I'm designing some breakout PCBs to make joysticks (primarily for myself, but if they work I would put the designs out to the public). Current project is a series of half scale arcade cabinets, one of which is powered by a slightly modded and de-cased Genesis model 1, VA5. I'm creating a two-tier pcb stack to support 6-button, 3-button and SMS controller input with only 2 SDPT switches to swap between them. For my primary prototype I'm using the IC from an official 6-button pad that had a badly damaged cord, and a fairly nasty shell.

[yui]

Hello, as the schematics on the 1st page goes to 404 and i had downloaded it to reproduce in kicad i can share it again. i am willing to take it down if KillingBeans wishes as it his work.
If i complete the kicad version i will share it too, as it could be useful to someone. https://drive.google.com/file/d/1ySIl_keK4tUDEVnEDWFbkqGH-Xpcm6XR/view?usp=sharing

[dankcomputing]

It would be useful for fixing modern pads that don't implement the protocol properly, and for building dedicated fightsticks.