Physical implementation of Conways computer

For general discussion about Conway's Game of Life.
Post Reply
waitmar
Posts: 3
Joined: April 19th, 2020, 11:58 am

Physical implementation of Conways computer

Post by waitmar » April 19th, 2020, 12:05 pm

Hey. I did not find the answer to my question, for this I write here. What are the simple physical or chemical methods for realizing a Conways game? Something related to semiconductors or chemical bonds? I dream of making a physical computer (hardware) on the principle of a Conways Game :)

Hunting
Posts: 4395
Joined: September 11th, 2017, 2:54 am

Re: Physical implementation of Conways computer

Post by Hunting » April 19th, 2020, 9:07 pm

Welcome to the forum!

Do you mean something like, this?

waitmar
Posts: 3
Joined: April 19th, 2020, 11:58 am

Re: Physical implementation of Conways computer

Post by waitmar » April 20th, 2020, 6:24 am

No, this is mechanical realization. I mean realization in microchip architecture, something like cellular transistors. So that it has real performance (at least like old computers of 2000)

Hunting
Posts: 4395
Joined: September 11th, 2017, 2:54 am

Re: Physical implementation of Conways computer

Post by Hunting » April 20th, 2020, 7:50 am

waitmar wrote:
April 20th, 2020, 6:24 am
No, this is mechanical realization. I mean realization in microchip architecture, something like cellular transistors. So that it has real performance (at least like old computers of 2000)
IMHO It's still trivial. It's just a combination of logic gates etc. Every cell is an JK trigger plus some additional circuitry, which is activated every time the clock circuitry sends out a signal etc.

You may want to do some 1D CA first.

GUYTU6J
Posts: 2200
Joined: August 5th, 2016, 10:27 am
Location: 拆哪!I repeat, CHINA! (a.k.a. 种花家)
Contact:

Re: Physical implementation of Conways computer

Post by GUYTU6J » April 20th, 2020, 12:08 pm

Just curious, do you have a basic knowledge about applications of semiconductors, chemical bonds or "cellular transistors"?
For the chemical part let's just start checking here and something related:https://en.wikipedia.org/wiki/Molecular_machine

waitmar
Posts: 3
Joined: April 19th, 2020, 11:58 am

Re: Physical implementation of Conways computer

Post by waitmar » April 21st, 2020, 8:47 am

Yes, I know what semiconductors are. By cellular transistor, I meant the simplest possible logical element from semiconductors, which will work according to the rules of the Conway cell.

gschadow
Posts: 4
Joined: August 17th, 2020, 6:27 pm

Re: Physical implementation of Conways computer

Post by gschadow » August 18th, 2020, 9:46 am

I just did one, I call The Life Machine™, a hardware implementation that produces generations in real-time at the speed of the raster image frame rate.

There are other ways to think about this though:
  • Independent from the actual image, this is what you would do if you want to have an infinite universe, where you would need to scroll and zoom the view-port to where you want to look; or certain curved surfaces, where you need to use some projection to see the action
  • Neighbor-aware cells, which I think is what you were talking about. Someone said a JK flip-flop, I say flip-flop or even an inverter loop as used in SRAM, with an 8:3 compressor + comparator, or saturated adder, arranged in a matrix. This method becomes costly though, because you need to put all these adder gates for every cell, proportional to the square of the side-length of the universe. If you can somehow squeeze the logic down to analog, maybe you could build a 16x16 prototype with discrete MOSFETs, which then you might be able to make a real chip from which scales. I would start from looking how RAM is built and see whether it is feasible to add the inter-cell connections to wrap the logic into the RAM.

hkoenig
Posts: 258
Joined: June 20th, 2009, 11:40 am

Re: Physical implementation of Conways computer

Post by hkoenig » August 18th, 2020, 9:56 am

Back in the late 1990s, I remember the Seattle airport had a Life display wall at one of the concourses as part of the various artworks. Display consisted of individual bulbs, maybe a 24x24 array. It would display patterns and run them for a while, about 1 gen per second. Not sure how they were generated, or how it all worked.

Was walking by and it took only a moment to recognize what the flashing lights were doing.

User avatar
pcallahan
Posts: 845
Joined: April 26th, 2013, 1:04 pm

Re: Physical implementation of Conways computer

Post by pcallahan » September 11th, 2020, 11:30 pm

hkoenig wrote:
August 18th, 2020, 9:56 am
Back in the late 1990s, I remember the Seattle airport had a Life display wall at one of the concourses as part of the various artworks. Display consisted of individual bulbs, maybe a 24x24 array. It would display patterns and run them for a while, about 1 gen per second. Not sure how they were generated, or how it all worked.

Was walking by and it took only a moment to recognize what the flashing lights were doing.
It would be interesting to have CA cells with LEDs that could snap together like legos and apply rules to neighbors. If you make the rule set programmable it would be a lot better, but i'm not sure what that would take. I can imagine either a rule propagation mode in which the units reset and accept new tables from adjacent cells before running the rules. Or if they had optical sensors, maybe you could even broadcast the rules at a really low transmission rate (a few hundred bits per second). How cheap could you make this kind of hardware I wonder. A 24x24 display is 576 units. At $0.10 per piece, you are under $60, which is pricy but not completely off the charts. It's also a pretty small grid.

Obviously this is not going to compete with the computation people can do on commonly available cell phones (which greatly exceed large computers of decades ago) but it works better as a craft project. I'm not sure about the optical sensor for programming. I feel that could actually be made cheap enough to replicate for each whereas something like wifi would be more expensive. Update: an LED can be used as a sensor so it should be pretty cheap. Because the rule set is minimal, you can get by with extremely low bandwidth. In fact, I wonder if you could come up with an inexpensive way to connect adjacent cells optically as well instead of with wires.

Or power the whole thing inductively and have no electrical contacts at all. Sorry just spitballing here. Some might break, but if they were all interchangeable, that would not be a major problem. (Wiring for power does seem a lot more realistic though.)

User avatar
Scorbie
Posts: 1692
Joined: December 7th, 2013, 1:05 am

Re: Physical implementation of Conways computer

Post by Scorbie » September 12th, 2020, 3:00 am

Previous FPGA Life Implementations, almost-off-topic but I think some visitors of this thread might find it useful.

viewtopic.php?f=7&t=4463&p=95520#p95520
viewtopic.php?f=2&t=1452&p=65126#p65124

Yoel
Posts: 384
Joined: July 2nd, 2020, 1:02 am
Location: Electronic jungle
Contact:

Re: Physical implementation of Conways computer

Post by Yoel » February 15th, 2021, 7:36 pm

Hunting wrote:
April 20th, 2020, 7:50 am

IMHO It's still trivial. It's just a combination of logic gates etc. Every cell is an JK trigger plus some additional circuitry, which is activated every time the clock circuitry sends out a signal etc.

You may want to do some 1D CA first.
It would be funny to build a large rectangle of microcontrollers with colorful LEDs that would play an arbitrary rule (Generations etc.) in a truly parallel fashion. This could be a school project or something. Making it larger than a few cells would be a ridiculous waste of time though (if soldered or even connected on breadboards by hand). More realistic, way faster and simple thing would be to program a GPU to do this (already done for CGOL).

Yoel
Posts: 384
Joined: July 2nd, 2020, 1:02 am
Location: Electronic jungle
Contact:

Re: Physical implementation of Conways computer

Post by Yoel » February 15th, 2021, 8:12 pm

GUYTU6J wrote:
April 20th, 2020, 12:08 pm
Just curious, do you have a basic knowledge about applications of semiconductors, chemical bonds or "cellular transistors"?
For the chemical part let's just start checking here and something related:https://en.wikipedia.org/wiki/Molecular_machine
Well, chemical or biological nano-computing is way beyond anyone's home or even average school project. This sort of equipment is very specific, very high-end and I doubt anyone who has access to it would be given an opportunity to implement CA on molecular level just for fun. Although, one may argue that chemistry and biology in general is about generalized (usually probabilistic) CA!

Hunting
Posts: 4395
Joined: September 11th, 2017, 2:54 am

Re: Physical implementation of Conways computer

Post by Hunting » February 16th, 2021, 2:29 am

Yoel wrote:
February 15th, 2021, 7:36 pm
Hunting wrote:
April 20th, 2020, 7:50 am

IMHO It's still trivial. It's just a combination of logic gates etc. Every cell is an JK trigger plus some additional circuitry, which is activated every time the clock circuitry sends out a signal etc.

You may want to do some 1D CA first.
It would be funny to build a large rectangle of microcontrollers with colorful LEDs that would play an arbitrary rule (Generations etc.) in a truly parallel fashion. This could be a school project or something. Making it larger than a few cells would be a ridiculous waste of time though (if soldered or even connected on breadboards by hand). More realistic, way faster and simple thing would be to program a GPU to do this (already done for CGOL).
I always dreamed of something like that. I will probably do it.

Perhaps even with transistors.

Yoel
Posts: 384
Joined: July 2nd, 2020, 1:02 am
Location: Electronic jungle
Contact:

Re: Physical implementation of Conways computer

Post by Yoel » February 16th, 2021, 3:55 am

Hunting wrote:
February 16th, 2021, 2:29 am
Perhaps even with transistors.
Using microcontrollers would be much simpler and cheaper than bare transistors. Plus resistors, diodes, capacitors, probably a few coils to regulate the timing, plus a lot of tenuous work for soldering or breadboarding all this together, testing... Unless you have unlimited access to electronics parts, even ready-made Arduino-like boards would probably cost cheaper.

Saying if after picking up an old TV from trash 2 days ago. For analog stuff though.

Hunting
Posts: 4395
Joined: September 11th, 2017, 2:54 am

Re: Physical implementation of Conways computer

Post by Hunting » February 16th, 2021, 6:25 am

Yoel wrote:
February 16th, 2021, 3:55 am
Hunting wrote:
February 16th, 2021, 2:29 am
Perhaps even with transistors.
Using microcontrollers would be much simpler and cheaper than bare transistors. Plus resistors, diodes, capacitors, probably a few coils to regulate the timing, plus a lot of tenuous work for soldering or breadboarding all this together, testing... Unless you have unlimited access to electronics parts, even ready-made Arduino-like boards would probably cost cheaper.

Saying if after picking up an old TV from trash 2 days ago. For analog stuff though.
But using bare transistors and resistors would look much cooler. It's probably a personal bias, but I feel that chips look less like electronics.

Yoel
Posts: 384
Joined: July 2nd, 2020, 1:02 am
Location: Electronic jungle
Contact:

Re: Physical implementation of Conways computer

Post by Yoel » February 16th, 2021, 8:46 am

Hunting wrote:
February 16th, 2021, 6:25 am
But using bare transistors and resistors would look much cooler. It's probably a personal bias, but I feel that chips look less like electronics.
I also prefer transistors and even tubes! I have a few good ones from the 1960s in my collection, as well as a Geiger counter, old-style neon lamps etc. However, I like bare transistors for things like Tesla coils, powerful transmitters and other high power analog things. For digital electronics we have nowadays cheap ready-to-use chips and boards like Arduino.

I do see how building such things from scratch is really cool. My dream is to make home-made FETs (difficult, but possible by some electrolytic oxidation reactions; crude diodes are trivial by whatever semiconductors like Al2O3 or PbS) or tubes (a vacuum pump and glass-working equipment is easy to get, in principle). I did make a few relays, fully home-made or based on coils from trashed TVs. Have you seen this (except that this guy is cheating by using a RAM chip)?

Yoel
Posts: 384
Joined: July 2nd, 2020, 1:02 am
Location: Electronic jungle
Contact:

Re: Physical implementation of Conways computer

Post by Yoel » February 19th, 2021, 4:06 am

Hunting wrote:
February 16th, 2021, 6:25 am
But using bare transistors and resistors would look much cooler. It's probably a personal bias, but I feel that chips look less like electronics.
I have 2 other ideas without logic gates at all:

1. Thyristors with neon bulbs as triggers switched by current comparators. The current pulse summed up from the 8 neighboring thyristors indicates how many neighbor cells are on, causing the comparators to switch (or not) the thyristor. It should be possible to power up such a thing simply but the mains electricity and to see it operating at 50 or 60 Hz. Since thyristors are high power devices, even things like small Testa coils could be attached instead of neon lamps!

2. Custom relays switched by pulses of particular current ranges. Some switching power could be dissipated by the relay itself to make it less sensitive for "off" signals (to implement B3/S23 instead of B23/S23 or B3/S3). The only semiconductor parts necessary for such a thing would be a few diodes.

Post Reply