Retrospective: PCs Powering Arcade Machines

April tends to be a very slow month for arcade news, since most of the info that pertains to the industry (particularly brand new games) takes place at Amusement Expo. So to help fill in the news gap, it’s time for another history article, although this one is all from myself and not the usual guest poster, Kevin Williams (although I do have to give him a H/T for some info & links on the Intel arcade connection). Our subject this time: PCs in arcades, particularly in regards to the early times and not their ubiquitous use in modern machines. How did it get to that point? That’s what I’m going to explore.

Arcades have the distinct honor of being the first place where the CPU would be used in a mass produced, commercial setting, but since coin-op machines were not “personal,” it would have been more appropriate to call them Public Computers.

No one really says that though, so when I say “PC,” it’s being used in the generally accepted way that everyone has used it since forever – the Personal Computer, or computer meant for home use. Generally speaking, such computers had a CPU, RAM, video & audio processors, ROMs for OS storage and methods for accessing software (cartridge ROMs, floppy discs, hard drives, etc.). It was a novel idea in the late 70s and early 80s, although usually not a cheap one – many on the market would cost over $1000 when they first went up for sale. This was mitigated to a strong degree by the likes of Commodore, Atari, Texas Instruments, Radio Shack, Apple and others who made personal computing more accessible to the masses thanks to low prices (especially the Commodore 64 and Atari 400). Of course, while those would be packaged in a much different way than an arcade machine, arcades hold the distinct honor of putting all of that tech into consumer hands first…but in such a focused way that no one would think of grabbing an arcade machine for their house and seeing if it could run spreadsheets or operate a printer.

A few notes: I am not going to be discussing consoles adapted to arcade hardware, as that would be a whole other post in itself. We’ll save it for a later time.

I am only looking at games which took a consumer product and adapted it into coin-op use(basically buying a PC from the store and throwing it into a cabinet with their own I/O board. I also don’t mind if they added more RAM to it but that’s as far as we’ll go).

I am not considering things like CPUs which were used in home computers as well as arcades, as then this post would have to cover 90%+ of the arcade games ever produced.

Also, there is a chance that I have missed something here, although I’ve made every effort to get my research right. I didn’t even known about the Intec Video System until this February, as I discovered it purely by accident, so it’s likely there’s some other obscure game release that also used a home PC to power everything.

If you’re interested in some other historical “retrospectives” on the site, check out The Tech That Nearly Killed The Arcade Industry Part 1 and Part 2.

In The Beginning…

PCs really started to reach the home market in the late 70s, although it would be a little while before one of the major manufacturers would decide to adapt such a thing over to coin-op use. Interestingly the first company to do so would be Bally, the company known best for pinball. How it worked out was a little bit backwards – in late 1977,  they created and launched the Bally Astrocade. This oddity wasn’t entirely a pure PC – it was part PC, part game console, something that many early PCs would do – but it did have a BASIC cartridge and programs could be saved to cassette tapes. As mentioned on old-computers.com, “For about 6 months, it was thus the world’s cheapest computer.”

The console/PC hybrid was a flop on the home market but with Bally still going strong in coin-op, they adapted the Astrocade to use in arcades, first doing so with Sea Wolf II that was released in 1978. By using a modified Astrocade computer, this also allowed the game to use extensive color – not a first, but certainly a rarity at the time where most games were still done in B&W.

Nine other games would launch using this hardware (not counting Demons & Dragons  since it was never released), with Wizard of Wor (1980) and Gorf (1981) being the most famous examples to be powered by the setup. That said, it didn’t appear that anyone at Bally really considered using the system to easily swap out software, as many other examples you’ll read about below focused on doing. That is likely due to the failure of the system on the consumer market – although in 1983 they did promise easy upgrades for Professor Pac-Man, only to not really pursue that idea when the game was a giant flop.

Into The 80s

By the time the market crashed in 1983, there were a plethora of PCs available for consumers to grab, few of which were compatible with the other. IBM-compatible PCs would soon dominate the scene, but aside from the Astrocade above, no one really thought about using one of the many PCs to operate an arcade machine until Exidy did so with their Max-A-Flex system in 1984.

’84 was a year where operators were demanding low cost kit solutions to refresh their older dedicated cabinets and Exidy thought they had a winner on their hands by taking an Atari 600XL (which was essentially an Atari 400 in a sleek new case) and adapting it to operate in a coin-operated environment. Given that the computer could easily switch out game ROMs, Exidy had hoped that they would be able to quickly and cheaply convert existing titles for the platform into coin-op games and make bank. This was ironic though, given that they used a major competitor’s platform instead of their own Exidy Sorcerer PC (which suffered a similar fate to the Astrocade).

As clever as they thought the idea was, it didn’t pan out, as they only managed to release 4 titles for the platform (Astro Chase, Boulder Dash, Bristles, and Flip & Flop) before abandoning it all together. Why is that, you might wonder? The games simply didn’t earn. This is likely due to a few issues, although I have to admit I have no hard data in this regard:

1) Players who might be familiar with the games might have had them at home and didn’t see the need to drop a coin into them

2) The games hadn’t been changed in any way aside from taking coins. Out of the four games for the Max-A-Flex, the one that really worked best in a fast-paced coin-op environment, and had some name recognition was Boulder Dash (thanks in good part to its Dig Dug-style gameplay). This title was good enough that a year later Data East released their own version of the game on their DECO Cassette platform then would later release a remake for arcades in 1990.

3) While cheap, PC hardware at the time couldn’t keep up with what arcade developers could do in creating their own systems. Often arcade makers would include more than one CPU or audio chip, could include custom chips and so on. The Atari 400 hardware as one example had been released in 1979. By 1984, far more powerful setups which were custom-made for video games could run circles around it, offering superior graphics and sound.

This failure of the Max-A-Flex would not dissuade others in the future from attempting the same model though, although it would be a few years before another attempt would be made.

Sweet 16 (Bits)

No, bits ≠ graphics like Atari, Nintendo & Sega’s marketing departments led us Gen Xers to believe, but the arrival of 16-bit computer to homes seemed like it, since the graphics and sound were superior to what had come before.  Funny enough, PC games on platforms like the Commodore Amiga and the Atari ST were looking more like arcade games and these platforms would get plenty of arcade ports. These improvements were down to several factors beyond bits (really bits just means that the more a processor can handle during a cycle, the more info it’s processing at a time; i.e., if you have an 8-bit and a 16-bit processor operating at the same speed, the 16-bit one can process double the data in the same timeframe. There’s more to it than that, but I’ll save the lecture for elsewhere) but it also meant that these popular and affordable new computers could be adapted for coin-op by companies who didn’t have the R&D budgets of Sega or Taito. To quote Atari’s PC ads of the day, you could get “power without the price.“

Interestingly enough, the infamous rivalry between the Amiga and the Atari ST would play out in the coin-op space, although it was a wholly one-sided battle. The Commodore Amiga 500 would be the most enticing platform on the coin-op scene, developers hoping that they could provide a cheap kit system with easy to swap games to the market. Bally Sente first toyed with the idea in 1986 with their game Moonquake, an unreleased game that got as far as having a flyer made for it that touted the equally never-released Sente Super System.

Even though Sente didn’t get the idea off of the ground, someone thought it still had merit, founding Arcadia Systems around 1987 and launching the Super Select System the following year. I can’t find much in the way of info about the people behind the company (aside from it being a division of PC-game developer Mastertronic that would be based out of California) but it wouldn’t be out of the question to assume that someone had been working for Bally Sente, liked the idea behind the Sente Super System, and decided to go get Mastertronic on board with it. The similarity in the names (The Sente Super System vs. The Super Select System) lend credence to this idea, although if I’m 100% wrong there, I’ll be happy to correct the record.

Either way, Arcadia’s Super Select System promised to do for coin-op what Nintendo’s PlayChoice-10 did for the market, but proffering much more powerful hardware that kept costs affordable. The Commodore Amiga 500-based platform could handle 10 games at once, all adapted for coin-op play from the Mastertronic catalog. They only managed to get 14 games out of the gate: Aarrgh!, Blastaball, Delta Command, Leaderboard Golf, Magic Johnson’s Fast Break, Ninja Mission, Pharaoh’s Match, Road Wars, Rockford, Sidewinder, Space Ranger, World Darts, World Trophy Soccer, and Xenon – but that was more than Exidy had done with the Max-A-Flex. That said, these didn’t review very well and quickly ended up finding their way to a dusty warehouse or a dumpster (or a dumpster at a dusty warehouse).

Arcadia’s Super Select System failed to materialize the same success that the PlayChoice-10 had found, but it did inspire someone in the UK to give the idea a shot with Commodore’s rival, the Atari ST. The Intec Video System had hoped to do the same thing as the SSS, but at a lower price point, a 21″ CRT and it would load games off of…floppy disks instead of ROMs.  I’m not sure why they thought that was a better idea than easy-to-swap ROM boards, but that was how it was touted, initially. It’s sort of amusing and sort of sad to see the contrast of the pre-test hype and the post-test crash on this one, though it was better that Mr. Nelson figured this out before he spent too much money on it.

Amiga fans would love this headline although given how the Amiga-powered Super Select also flopped, it’s a hollow victory.

Granted, that headline gets back to a point mentioned earlier and I think it still applies today – when a game hasn’t been designed with coin-op in mind, the chances of it doing well are very slim. It turned out that it would be easier to adapt console hardware and games to arcades, at least for a while.

It’s The 1990’s and It’s Time For Coin-op PC

Fast forward to 1997 and Intel was riding high off of the success of their Pentium CPUs, while Microsoft was taking a harder look at gaming in general. While interest in arcades was slumping from where it had been in the earlier part of the decade, both Intel and Microsoft saw an opportunity to bring development costs down for the industry via the Open Arcade Architecture Forum. The idea was simple – instead of arcade developers constantly creating unique hardware boards once every couple of years or just for a single game, Microsoft would create special arcade extensions in their DirectX APIs and pushing Windows as the OS of choice to make development more streamlined. This old news article is an interesting time capsule into the effort that would re-shape the arcade industry and set it up for how it is today, although this change isn’t something that a lot of arcade fans & collectors were too pleased with (I’ve always found a deep fascination with custom/specialized arcade hardware – PC’s always felt generic due to how common they are). That said, it is something that made both manufactures and operators happy, so that’s the route that things took.

Stepping back for a moment though we have to recognize 3DFX’s role in this too. For those who remember 3DFX, they likely recall that they were a creator of 3D graphics accelerator cards that could be installed into PCs. Few know however that 3DFX got their start in the arcade business, creating their cards for use on games like Atari’s San Francisco Rush and Wayne Gretzky Hockey. Arcades got to enjoy the Voodoo cards first but it wasn’t long before PC gamers could start enjoying arcade-like graphics either.

This kickstarted a short-lived trend at both Atari Games and Midway (which eventually became the same company) to create their own unique CPU & I/O computers, powered by 3DFX graphics accelerator cards. This started with the Flagstaff, and continued on with the likes of the Phoenix, Seattle, Vegas, Denver, Quicksilver II, and Graphite boards.

While it wouldn’t use a 3DFX card, Atari’s Media GX system would really be the first time (that I know of) where an IBM-compatible home PC would be adapted for use in an arcade setting. It used a Cyrix motherboard and x86 CPU, Cyrix being a minor competitor in the x86 war that was mainly fought between Intel and AMD (funny enough, I did own a Cyrix PC way back in the day, good times). The Media GX system is what powered Area 51: Site 4. It was the Midway Graphite however, which powered Arctic Thunder, where an arcade operator would open the back of their major-manufacturer made arcade cabinet to find a home PC case sitting there to run the game.

The Future Is Now

Before Arctic Thunder came along however, a couple of smaller developers jumped on-board Intel & Microsoft’s Open Arcade Architecture and themselves put those boxy PCs into arcade machines. One called Interactive Light you might not be familiar with (they created the unique dinosaur game Savage Quest as well as the popular Kick It! videmption game), but chances are that if you’ve visited an arcade in the past 20 years that you’ve seen a game from GlobalVR before. They started off with their Vortek VR machine, something that in light of today’s VR craze was way ahead of it’s time, and as they grew, they pushed the PC-for-arcades trend with them. It’s almost been 10 years since GlobalVR released a new game, although I understand that they are still selling parts and games to interested parties.

This video might seem random but it is relevant to the PCs-in-arcades discussion:

Sega and Namco would not immediately dump their own platforms for PC, but it didn’t take them too long to join with the fray. One could say that they both first made that jump with the Chihiro platform, which might have just been a modded Xbox, but that was also just a modded PC.  Sega slowly weaned themselves off of the NAOMI architecture, while supporting PCs via the Lindbergh system. Namco was content for a while to use their PS2 & PS3-powered boards until relenting with the N2 and ES1. Raw Thrills got their start in 2003 but like GlobalVR, they’ve never used anything other than PCs to run their games. Here’s one of their first ES1 games:

As a note on operating systems, developers will use either Windows Embedded or some flavor of Linux; It’s also become common to see games developed with mainstream engines like Unity and Unreal. The benefit to all of this is ease of development, which also means a lower cost – if you imagine that most games at the moment are $10k plus per cabinet, it would likely be much higher than that if the developer had to create their own hardware platform from scratch, while there being little benefit to them financially to do so. If anything, it would be a huge and stupid risk to potentially spend millions on developing unique hardware, when AMD/Intel/nVidia are already doing that for the wide market instead of a niche one.

It is a little sad that we likely will never see it go the way of custom hardware again, but it is understandable. Of course, there is always room for consoles to receive some upgrades and power some arcade machines, but that’s a subject I’ll save for another time.

While the hardware side of things is PC-dominated now, I think there are still lessons to be learned or re-learned when it comes to software. If a game starts on PC or consoles and is getting a port, the more it is customized for the arcade, the better its chances at becoming a success. Just putting a coin counter into a game and not changing anything else can be a recipe for an earnings disaster, something for any indie hopefuls out there to keep in mind.

Thanks for reading! If you know about some PC-powered arcade game out there that I missed (not console-powered), please let me know and I’ll have to add it to the history.