FPGA developer Jotego asks players where his arcade cores should go next

Arcade FPGA developer Jose Tejada, known across the retro hardware community as Jotego, is asking players and preservationists to help choose the next direction for his open-source project. Instead of dropping a poll with specific arcade boards to vote on, Tejada laid out three broader technical paths and opened the floor for feedback.

Since 2019, the Valencia-based engineer has built out the jtcores framework in Verilog, recreating classic arcade hardware at the circuit level for the MiSTer FPGA and Analogue Pocket. His work already covers more than 1,250 classic coin-op games, spanning Capcom's CPS-1, CPS-2, and CPS-3 architectures (including heavyweights like Street Fighter III: 3rd Strike), Sega's System 16 and System 18 boards, and custom FM sound chip implementations like the Yamaha YM2151 and YM2612.

Cabinets full of arcade PCBs and cheaper automation

The first two options on the table both come down to falling hardware costs for reverse engineering. Tejada notes that PCB validation has become much more affordable lately, and he is sitting on cabinets full of physical arcade boards that could be used to double-check older assumptions.

In the early days of the project, his team focused mainly on converting physical boards into KiCAD schematics. Now they can probe deeper into the silicon to improve cycle accuracy, and Tejada is even floating the idea of releasing internal tools so the community can speed up its own PCB-to-schematic work.

Cheaper automation also changes the math on what systems he can reasonably tackle. Historically, Jotego approached complex hardware in small increments so he could maintain a steady drip of weekly game releases for his Patreon backers (who get early beta access via the jtbin repository before cores go free to the public).

"Many measurement, verification, and testing tasks can now be automated at a fraction of their previous cost. This opens up opportunities to take on more ambitious systems that would have been difficult to approach before."

Adopting the orphaned cores on update_all

The third pitch is less about uncracked silicon and more about housekeeping. If you use the community "update_all" script on a MiSTer, you end up with dozens of standalone arcade cores from various contributors, and the user experience across them is wildly inconsistent.

Tejada pointed out that several cores sitting in MiSTer's Arcade folder don't work out of the box on his own CRT television setup. He is considering folding some of those abandoned or unmaintained cores into the JTCORES framework so they share the same standardized menus, video options, and display handling as his existing releases. He also expressed interest in building clean, well-documented, open-source replacements for cores that are currently closed-source or AI-generated.

Save states for those of us who need them

Tejada deliberately left specific game titles out of his post so he wouldn't steer the conversation, and the comment section quickly filled up with over a hundred replies. Most of the community responses are pushing for him to double down on hardware accuracy and deeper PCB validation.

I asked for more save states across the library instead. What can I say? I'm not a good gamer! Save states really help me.