Some basic laser logic?

I have other schematics. But won't bore you / clog this space. 5 more diagrams. G pass ( RB ref), B pass (GR ref)... ( like above but not)

RG pass (B ref), RB pass ( G ref), GB pass ( R ref). I will post one of these three and you can do the other two "in your head"

You get the idea.

Here is an interesting question.

For a given electrical input into a NMOS 6502 chip, how many transistors does that input pass through in the space of one clock cycle ( system cycle? And the "internal cycle?).

Similarly, if we assume my laser mirror-filters above are designed to be 90%+ or 95%+ or in some cases 99% efficient reflectors or pass-ers, we might ask "Is the system I envision fundamentally feasible. A sound idea? What wattage should the laser be? How miniscule an output signal is too miniscule?"

And what is the average and/or worst case scenario for fanout of an average electrical signal input into a 6502?

And could laser-mirrors meet this concern as well?

I'm not seeing any logic there. None of those diagrams require fancy mirrors: if they're looking down at the beams from above, all you need is to put red at one height, green at a different height, and blue at a third. Then placing regular mirrors at some of those heights will do the same thing. It also makes it obvious that the beams do not interact with each other in any way.

What you're missing is a way to control one beam with another. There are more exotic materials that will do this, but dichroic mirrors will not.

And how to "catch light in a bottle"?

For flip flops and latches?

And for synch-ing purposes?

The mirrors aren't that fancy. Thin layer films like in regular semiconductor manufacturing/engineering type stuff. Maybe use SiO2 and TiO ? Already a standard in optoelectronic chips?

I don't get your point? About top and bottom or offset beams? Mirror/filter must "discriminate", I.e. choose the path of the incoming beams? Right?

Here is a diagram I made six months ago.

Probably incomprehensible? But I remember proving it and setting it aside?

Functional completeness needs just one more logic block? Is it NAND and NOT? I forget?

That would be all combinational logic. But sequential logic is another story.

That should be self-explanatory?

0 is 3 separate beams. 1 is 3 combined beams.

1 in, 0 out.

0 in, 1 out.

Too many mirrors? Maybe? Somebody good with puzzles want to break it down simpler?

Could probably just use 2 beams. RG or GB or RB ... And appropriate mirror filters.

Spatial separation or combination ; path separation, is the 1 or 0.

For a two beam; say, just RG , from my diagram, you can get a 3 mirror inverter.

But I'm thinking I might try BG or BR.

My B laser ( purple violet? Listed 405 or 410 nm) seems to have significant portion in the UV? Or maybe my glowey-putty (the kids?LOVE it !) responds to blue/violet light too?

Oh? The package says, "...Expose to bright light ..." But they mean broad spectrum light? I think.

My red and green laser don't light it up at all?

So at the risk of clogging this, superfluous, I have, a few pictures of me "catching light in a bottle". The old delay lube technique. For 1.25$ at the ineptly named dollar store.

This is supposed to be a picture of 3 4-bit, nibble codes ; 1010 and 1110 and 1111.

But my hand is a bit shaky.

And a photo next to an old but trusty tech. For comparison .

P.s. sorry upside down photos. My phone is smarter than me?

1111 , 0111, and 0101 ... Top to bottom. Left to right.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

"In logic, a functionally complete set of logical connectives or Boolean operators is one which can be used to express all possible truth tables by combining members of the set into a Boolean expression.[1][2] A well-known complete set of connectives is { AND, NOT }, consisting of binary conjunction and negation. Each of the singleton sets { NAND } and { NOR } is functionally complete."~Wikipedia entry, Functional Completeness

Garth. Was just this moment thinking about this exact topic.

Half silvered or 1/4 or 3/4 mirrors can do a similar job.

Think of an AND with laser inputs. If both are logic high; RG on same wire, say. That's 2 unit power in on one wire, 2 unit power in on another wire, and assuming some interesting combination of 90-99% reflective / passing mirrors, the correct color combo cones out the other end.

But also the correct power level. It would probably not be useful to have 0.5R and 0.5G come out the other end. And 2R2G might be just as confusing.

Perhaps some systematic way of getting a 1.25R and a 1.25G signal out the other end, followed by a reduction filter mirror, might ensure that a 1G and 1R always comes out of our putative AND gate.

Really that's all my kooky idea needs. Is an AND gate. And then it's a functionally complete combinatorial logic system.

Then onto sequential logic.

Certainly "gating" or "switching" has been a problem for light technologies before.

In electricity an "active system" is one possessing transistors.

I am proposing a passive optical system .... Wherever possible.

I am playing with a DMD / DLP at the moment.

Must write a little raspberry pi code to get it running programs beyond the splash screen.

But I don't want to sample a light field ( dot matrix of laser points) with CCD and make electrical pulses to drive the DLP that will ACTIVELY direct the laser beams to the proper target.

That smacks of cheating to me.

I am looking into fused silica wafer waveguide manufacturing methods. Looking for low tech way to make them.

Also need a microlense array to couple beams into the fiber optic image conduit.

PM me if you have a cheap / free one. Otherwise I'm 4 or 5 months away from acquiring one 200 or 300 or 400 $ was the?cheapest I found?

It is possible that the system I describe, having no way to practically use on light pulse to turn on or off another light pulse, is a deal breaker. I am hoping it is not. And hoping I can imagine a way around it.

Time or Space or Heat, Garth?

Choose your poison?

"BTW, the smallest silicon geometries today have die feature sizes that are less than 2% of a wavelength of visible (blue/indigo) light, at about 6 nanometers, and they're working on 4nm.".

I am not sure I can get a 50-50 duty cycle laser. But I do know that 1 picosecond pulses are fairly common and fairly easy to attain and that there are tremendous volumes of post-2000s literature with "Femtosecond Lasers" in their book titles.

I have some TTL red lasers that will run at 15 kHz and some that should work at 50 kHz. Cheap from eBay. But some real equipment would be nice.

I am not super keen on my flip flop or latch solution so far. As "green putty" is not very scientific sounding.

Although the early astronomical CCD experimenters ... If I remember lore correctly and am not making up a new myth?... Quite literally drew on their CCDs with highlighter ( lumigen chemical) to make them UV sensitive. Effectively downsampling the frequencies of the impinging photons to the visible spectrum.

3D geometries might be easier with light, as it has a tendency to want to fly in all directions anyway.

I assume little problem with heat dissipation, though if I am clocking it with ultrashort light pulses at 500 GHz or 1 THzthis assumption might be flawed.

Glass is a good thermal reservoir? Fused silica has a low coeffcient of thermal expansion. Perhaps some manner of underlayer thermal heatsinking is in order?

But 3D NAND flash memory is doable. But 3D microprocessor architectures are less doable?

Flip chip RAM. To processors are used on raspberry pi.

Any too? I just think light might be better and faster and of COMPARABLE size.

FYI my FOICs have either 25 um or 24 um or 12 um cores.

One of the companies in CA that makes these, custom, puts in dark fibers inbetween to lessen light leakage / cross coupling.
EDA fibers? I forget what the acronym stands for?

But there is a company in Boston selling GPUs? And or logic accelerators that are using light chips.

Active light chips. In Silicon? I believe.

Light matters. Is the company / product.

It does certain math faster. But is not very good at general computing for the reasons we mentioned.

Can't get light to switch light ... Easily.

At a 1 picosecond pulse there should be about 500 cycles of EM radiation in the light pulse.

One of the books I was perusing said " 1 ps or lower, .... 0.5 ps theoretically" can be achieved with "mode locking" of lasers.

I gather it gets tougher after that.

I am no certain that it is capable of a 50-50 duty cycle. I was sort of just hoping.