The secret upgrade: How expansion chips in video game cartridges pushed performance

To push past the inherent limitations of early game consoles, developers quietly built chips into the game cartridges themselves.

cartridge boosters 01
Credit: Benj Edwards
A secret weapon

During the 1980s and early 1990s, when video game titles shipped primarily as cartridges, software developers eked out extra capabilities in their games by including special expansion chips inside the game cartridges themselves. The most simple of these enabled bank switching (a technique that allowed a CPU to access more RAM or ROM than usual), but the chips went on to grow dramatically in scope: adding RAM, extra sound-synthesis capabilities, graphical tweaks, and eventually 3D polygonal graphics using DSP co-processors. In the slides ahead, we’ll take a look at a handful of classic secret hardware expansions that allowed game developers to push game consoles past their inherent design limitations.

cartridge boosters 02
Credit: CBS Software

The Atari 2600 was a stunningly primitive machine by today’s standards, shipping with a mere 128 bytes of RAM and no frame buffer. Those limitations actually increased its lifespan in the long run because they opened the door for new programming techniques and in-cartridge hardware expansions to keep the system relevant. The CBS RAM PLUS chip was one such hardware expansion chip; it added 256 bytes of available RAM, which allowed for slightly more sophisticated game design—when tucked into copies of the pseudo-3D Tunnel Runner, seen here.

cartridge boosters 03
Credit: CyberRoach
Videopac C7010 Chess Module (1983)

What we know as the Magnavox Odyssey 2 (1979) in the U.S. first launched as the Philips Videopac G7000 in Europe. It included a fairly wimpy Intel 8048 CPU and 64 bytes of RAM (even less than the Atari 2600). When it came to playing a computationally intensive game of computer Chess, the Videopac couldn’t cut it, so developers devised an add-on computer module permanently wired to the game cartridge. The module included an 8-bit, 4.43MHz NSC800 CPU and 2KB of RAM. It was an expensive solution, but at least it made playing chess possible.

Unlike every other example in our survey, this co-processor wasn’t small enough to hide inside a game cartridge itself, but it does provide the most extreme example of single-use game expansion hardware I’ve ever seen.

cartridge boosters 04
Credit: Tomark Edzwyn
Activision DPC (1984)

When it came time to create a sequel to his Atari 2600 platforming-classic Pitfall!, David Crane craved a dramatic musical soundtrack to accompany the game. Unfortunately, the 2600 hardware wasn’t powerful enough to play the song and handle the graphics at the same time, so Crane devised a chip called the “Display Processor Chip” (or DPC for short—which, not coincidentally, are David Crane’s initials). The DPC acted as a sort of CPU assistant, fetching data quickly and putting it where it needed to go so more things could happen at once—such as improved graphics and sound. Crane intended to use the chip in other 2600 games, but the 1983-84 video game crash put those plans firmly to rest.

cartridge boosters 05
Atari Embedded POKEY (1987)

When General Computer Corporation created a follow-up for the Atari 2600 in the early 1980s, they stuck with its predecessor’s sound processor (TIA) for backwards compatibility. By the time the console was actually released in 1986, the 2600 sound chip had aged poorly. GCC had a backup plan, which included embedding an Atari POKEY sound chip into every game cartridge for enhanced sound. Atari engineers had originally designed the legendary POKEY sound chip in the 1970s for use with the firm’s original 8-bit computer line, the Atari 400 and 800, but it was still fairly competitive by 1986. Only two 7800 games shipped with a POKEY chip inside: Ballblazer and Commando.

cartridge boosters 06
Credit: Nolan Pierson
Nintendo MMC5 (1989)

If you count its Japanese origins, the Nintendo Entertainment System’s hardware design remained on the market from 1983 all the way to 1995. The secret to the NES’s amazing lifespan was its reliance on embedded cartridge chips called Memory Management Controllers (MMC), of which Nintendo released seven over the years.

The most advanced of these, MMC5, included an extra 1K of RAM and a large grab-bag of memory manipulation tricks to allow games with superior graphical richness. The U.S. release of Castlevania III notably included a MMC5 chip, which allowed detailed stained-glass backgrounds and smooth vertical scrolling, among other benefits.

cartridge boosters 07
Credit: CobraSA
Nintendo Super FX (1993)

Like its predecessor, the Super NES also relied heavily on cartridge enhancement chips to extend its lifespan by providing interesting new graphical and sound capabilities. The most famous of these is undoubtedly the openly hyped “Super FX” chip, which brought high-frame-rate polygonal 3D graphics to the 16-bit console for the first time through the game Star Fox. Revisions of the Super FX chip also allowed advanced 2D sprite-scaling effects like those seen in Super Mario World 2: Yoshi’s Island.

cartridge boosters 08
Credit: Martin Petersson
Sega SVP (1994)

Upon its release in the arcade in 1992, Virtua Racing became a landmark in 3D polygonal gaming with its smooth frame rate and detailed visuals for the time. When it came time to bring that magic home to its popular Genesis console, Sega turned to a beefy co-processor called the SVP—a DSP chip running at a rumored 23MHz that enabled enough on-the-fly calculations for the Genesis to render an enjoyable home version of the 3D racing game. The port had its drawbacks, though: Aside from not being as detailed as the arcade release, its embedded chip proved costly, pushing the MSRP of the game up to $100. Probably due to cost concerns (and the impending 32X enhancement), Sega never used the SVP chip in a Genesis game again.