SuperChis Prime: How I Fixed the SuperCard SD's Flaws on a Budget
Hello players, I'm one of the designers behind the SuperChis project. I’d like to share, from an engineer’s perspective, the thinking and execution behind our new GBA flash cartridge, the SuperChis Prime.
Our goal was clear: to create a genuinely "good" budget cartridge within a $15~$25(depends on tax and shipping) price range by solving the core issues that plague the classic SuperCard SD.
This wasn't just about patching an old design; It's a reconfiguration for a better future.
Our goal was clear: to create a genuinely "good" budget cartridge within a $15~$25(depends on tax and shipping) price range by solving the core issues that plague the classic SuperCard SD.
This wasn't just about patching an old design; It's a reconfiguration for a better future.
1. Hardware: Targeted Solutions
The root causes of the classic SuperCard's problems—"slowdown" and "white screens"—lie in its SDRAM timing compatibility and the need for "slowdown patches" for every game. Our approach was twofold:- Using Large NOR Flash to Bypass SDRAM Issues: I replaced the NOR Flash with a larger-capacity chip. When a game is written to the Flash, the GBA CPU reads directly from it, completely bypassing the problematic SDRAM access timings. This eliminates "slowdown" and the "white screen/freeze" caused by failed patches at a fundamental level. We also deeply optimized the write speed. Burning a 32MB game takes about 105 seconds on the standard version, and only about 85 seconds on the "Micron" version with higher-performance Flash.
- Premium PCB Process: To ensure long-term reliability, especially for the contact fingers, we use a 2U Immersion Gold (ENIG) process for our PCB manufacturing. It's an invisible upgrade, but we believe it's essential.
- Standardized and elegant PCB layout: It can improve product stability and avoid some anomalies that are not caused by component quality. Thanks to Mori, for his perfectionism.
2. CPLD Firemware: A Clean-Slate Approach for a Better Future
The original SuperCard's CPLD firmware was "black box". So, we made a decision: we started from scratch with a completely new, forward-designed firmware. Thanks to davidgf for the many suggestions, which have benefited me a lot.- More Developer-Friendly: The new firmware has clear logic, avoiding potential register conflicts and other hidden issues. This provides a much better foundation for long-term support and feature development for excellent third-party kernels like SuperFW.
- Multiple compatibility interfaces: We are compatible with SuperCard SD's SD Write/Read, and also support SuperCard Lite's SD Read. This helps to quickly migrate existing excellent software.
3. "Hybrid Architecture" and Fragmented Flash Management
Making a simple NOR Flash cart isn't hard. The challenge was balancing flexibility and features at a low cost.- Fragmented Flash Management: Users don't need to worry about a game's physical location on the Flash. You can add or delete games as easily as managing files on your phone, and the system handles the space allocation automatically.
- Retaining SDRAM for Unique Value: We didn't discard the SDRAM just because we added Flash. We repurposed the SDRAM as a "preview/launch area" and, crucially, as the memory for some feature. This "Flash + SDRAM" hybrid architecture allows us, at a low cost, to offer both "instant game launch with no slowdown" (from Flash) and "advanced features like Instant Saves" (from SDRAM).
4. GBA Firmware
At present, SuperFW has fully supported SuperChis and may be released in next version. Because SuperChis are forward compatible, there will be more custom firmware support in the future. It's worth noting that the time to enter the SuperFW was reduced from 6 seconds to 1 second because we had a larger Flash and avoided asset decompression. davidgf's rotating meta data design largely avoids uneven flash writes.5. The Necessary Trade-Offs
Within our strict cost targets, we had to make some compromises:- No RTC/Rumble/... Support: This was to control core chip complexity and cost. The software clock in the SuperFW kernel keeps time while in-game.
- Standard Battery-Backed Save: The standard version uses a CR1616 battery with low-power SRAM, offering a theoretical save retention of up to 20+ years (conservatively estimated at 10). We solved the problem of SuperCard batteries dying prematurely by using low-power SRAM and designing strictly according to the manual.
- Compatibility issues: We have found that the some GBAs on the motherboards labeled 02 have power supply issues, and cannot use the common NOR Flash, and need to use the more expensive low-power, high-performance NOR Flash, for which I have already funded research, and hope that the 02 motherboards can be resolved in the future with a simple mod. Or you can replace it yourself with a more expensive flash like the MT28.
- Save Game Handling: Support for games with EEPROM/Flash saves relies on the SuperFW database and patching engine. While this covers virtually all commercial games, some homebrew titles (e.g., certain Pokémon ROM hacks) may require manual intervention.
- Use recycled ICs: Whether it is mass production or DIY, for GBA game scenarios I recommend using tested recycled ICs. Especially for those NOR Flash that still have more than half of their life left. This is more economical and environmentally friendly. Of course, the test process includes full erasure, writing, stress testing, etc. Of course, you can also replace it with a new NOR Flash by yourself, which will cost $20~$30. As long as the speed is faster than S29GL01GS12, it will be a good Flash choice.
I'm just curious
