FPGA in 30-40 Years

[TimePiloteer]

I sure love what FPGA technology has done for arcade technology, but I can't help but wonder what some of these FPGA-based replacement boards will be like in 30-40 years. Will these current-gen FPGA chips fail and need to be replaced? Will these current-gen chips even be manufactured anymore or available in 30 years or beyond (tech seems to evolve ever more rapidly these days)?

In 30 years it'll be even harder to keep the original arcade PCBs going, as various TTLs, CPUs, RAM, EPROM and other chips will have been out of production for many years and obsolete or incredibly hard to find, while more original parts will be stretched even further beyond their life expectancies and fail more. Will our JROK or BitKit boards be in the same situation 30 years from now as the original arcade hardware?

With many of these FPGA systems on the market having the hardware language implementation closed-source, also creates a situation where it could be difficult to port these to whatever newer platforms come along between now and then. Who will understand old Verilog/VHDL and have an old Windows 10 machine to compile it on in 30 years? What do you all think? Is this a bit of kicking the can down the road?

Hardware emulation is more accurate now than software emulation, but comes with more of a dependency on - hardware - which is what we are struggling with on original boards. Would a bigger investment in making software emulation more accurate be a better or more portable strategy for preservation? Curious what all the smart folks here might think about this topic.

 

[Yoeddy]

Most of us will be pushing up daisies by then.

Jason

 

[Arcadenut]

In 30 years it'll be even harder to keep the original arcade PCBs going, as various TTLs, CPUs, RAM, EPROM and other chips will have been out of production for many years and obsolete or incredibly hard to find, while more original parts will be stretched even further beyond their life expectancies and fail more.

You don't have to wait 30 years, we are here already. I suspect we'll be able to keep these running for another 10 - 20 years for most games with what's currently available. Some parts are reproduced (Custom Chips usually) and others can be reproduced if it's cost effective (i.e. the SC-01) Some reproductions will be perfect, others not so much (see SC-01 again). You can also steal most TTL parts off of other "parts boards". At some point most games will probably be available in FPGA form as it becomes more challenging to repair existing games.

 

[Tornadoboy]

It may seem impractical or even ridiculous now, but I'm hoping manufacturing technology is eventually going to make reproduction on demand of pretty much any component easy, sort of like what 3D printing is the beginning of for a lot of things. That's why it's REALLY important to document everything possible right now while we can: Schematics, data sheets, ROMs of course, reverse engineering custom chips, scans of artwork, exact dimensions and blueprints of cabs, control panels and their parts, etc, hopefully ALL of that we be remade later which is the way most games are only going to survive in, their ability to be reproduced.

 

[bman7418]

I think 3D printing will get to the point that you can reproduce chips in demand. I bet they will have 3d X-ray scanning and be able to reproduce the chip. we are almost there with this technology anyway.

 

[awbacon]

in 30 years ill be 66 and in 40 ill be 76. I hope I am still playing games by then

 

[alfonzotan]

Nobody knew what an FPGA was in 1981. Nobody now knows what will be everyday technology that we haven't even thought of yet in 2061.

 

[KaBoom1701]

We will all be in the matrix and everything will be VR.....

 

[VectorCollector]

It all started here, the first demonstrated version, ->

The transistor was successfully demonstrated on December 23, 1947
at Bell Laboratories in Murray Hill, New Jersey. Bell Labs is the research arm of American Telephone and Telegraph (AT&T). The three individuals credited with the invention of the transistor were William Shockley, John Bardeen and Walter Brattain.

 

[CraftyMech]

As long as the code is preserved, FPGA implementations can always be ported to successive new generations of FPGA or equivalent technology. When the BitKit reaches the end of its lifecycle, I do fully intend to release all the source code.

Regarding the commercial lifespan of FPGA products, Xilinx Spartan3 series debuted in 2005, and the chips are still on the market. The BitKit is built around the Xilinx Spartan6, which was first released in 2009. The Spartan7 series was launched in 2017, and would be a fairly easy port for the BitKit platform.

The MiSTer project is more vulnerable right now, as it depends on a powerful FPGA platform (DE-10 Nano) that is heavily subsidized by the chipmaker. Should that relationship end, MiSTer would need to be ported to a different platform and that could dramatically drive up the cost. Most of MiSTer's cores could be ported to a less capable FPGA board, but late-gen systems like Neo Geo & CPS2 seem to be pushing the DE-10 capabilities already.

 

[ida34]

I hate to say it but in 30-40 years most of us that were around for the golden years in the 80's will be gone. I see this in my other hobby, military vehicle collecting. Thirty years ago the Vietnam era jeeps were cheap and easy to find. Then the Vietnam vets were getting older and many had more money to buy back their youth. Now those jeeps (M151s) are up there with WWII jeeps in value. I was born in 1970 so I was a teenager when most of the popular games came around in the 80's. Most of us from that era now have money to buy back our youth. I think you give it 20 years and our games will be going down in value and the number of collectors of the 80's era games will be way down. As a kid of the 80's I really like those games and the 90's and later games do not really appeal to me with some exceptions. I think we will see a boom in the 90's and up games in the near future as the kids that were teenagers in the 90's start buying back their youth.

 

[parism]

It may seem impractical or even ridiculous now, but I'm hoping manufacturing technology is eventually going to make reproduction on demand of pretty much any component easy, sort of like what 3D printing is the beginning of for a lot of things. That's why it's REALLY important to document everything possible right now while we can: Schematics, data sheets, ROMs of course, reverse engineering custom chips, scans of artwork, exact dimensions and blueprints of cabs, control panels and their parts, etc, hopefully ALL of that we be remade later which is the way most games are only going to survive in, their ability to be reproduced.

^ This

Electronics can be substituted/emulated with newer components but cabinet hardware and art will be hard to reproduce without good documented examples. Let alone finding anyone that cares enough to do so in 40 years time, and does it well. That is a problem today too!

This is the only reason I hoard parts for my games.

p

 

[bitburn]

In 30-40 years there could be an EMP or solar flare event that frys all circuitry causing massive fires. Need to chisel the ROM code onto granite slabs for future generations a 1000 years from now. Most ancient writings only exist on stone.

150 Years Ago: The Worst Solar Storm Ever

On Sept. 2, 1859, an incredible storm of charged particles sent by the sun slammed into Earth's atmosphere, overpowered it, and caused havoc on the ground.

www.space.com

Get a chiseling...

 

[Onyx Jaguar]

Yeah, without documentation all could be lost. Which is what grinds my gears about modern computer and car companies locking everything behind a "service". In some cases these companies would even seek to erase history to protect whatever "rights" they have locked away in their vaults. All this could be lost without proper documentation. It's happened before, and I'm not saying with certainty that it would happen again, being sure that we are safe from tech degredation, natural disasters, is foolish

 

[Bishop]

Collecting and value ebbs and flows in relation to the collecting audience.
When all the folks that were in arcades in the 80's-90's die off the value will free fall.
And all this stuff will be forgotten by future generations. Enjoy it now, don't worry about the future.

 

[XXI21]

in 30 years ill be 66 and in 40 ill be 76.

 

[toledoflyer]

Collecting and value ebbs and flows in relation to the collecting audience.
When all the folks that were in arcades in the 80's-90's die off the value will free fall.
And all this stuff will be forgotten by future generations. Enjoy it now, don't worry about the future.

Unfortunately, I think this is reality. As much as we all hope it’s not true.

 

[DZA]

As long as the code is preserved, FPGA implementations can always be ported to successive new generations of FPGA or equivalent technology. When the BitKit reaches the end of its lifecycle, I do fully intend to release all the source code.

This. Great to hear that the code will be released.

Even if toolchains are obsolete, deprecated, etc. people will still be able to read the code and refactor it as needed to a new platform.

CRTs are the major thing I think will be an issue in my lifetime.

 

[MikeHartman]

CRTs are the major thing I think will be an issue in my lifetime.

It would be awesome if screen technology progresses to the point where we don’t even miss CRTs. That will take a while though - we’re not even close yet.

 

[DZA]

It would be awesome if screen technology progresses to the point where we don’t even miss CRTs. That will take a while though - we’re not even close yet.

We'll see. I think it is less the actual "result" as much as it is the actual technology for me. It sucks that tubes are basically unable to be fixed once they are done for. Plus they are fragile so we lose some every year to negligence.

My understanding is that CRT production is unlikely to ever return for various economic and environmental reasons. Whereas some older technology is experiencing a resurgence, this won't happen for tubes.

Anyway, there are threads on threads about CRTs.

As for FPGAs, I think they are a great technology and wholly support arcade emulation on them. The toolchains are horizontally integrated enough that I suspect they will be supported for some time to come. The HDLs and other specification languages are likely to be supported on future platforms regardless of the underlying semiconductor tech.