Hacker News Reader

> 1. Extremely high duty cycle 100 gigawatt laser. (edit: now apparently it's an "array" because someone with a brain mentioned that a 100 gigawatt laser is ludicrous) That's right 7 peak summer output Grand Coulee Dams and 12 Palo Verde nuclear power plants with all reactors operating simultaneously COMBINED would be needed to power this laser.

That's not how this stuff works. You would use some sort of large capacitor/battery bank which is charged by a much smaller power draw over a long time, and then dump the energy over a short time.

The laser array currently operating at the National Ignition Facility has a peak output of 500 TW, more than three orders of magnitude higher than the number you called ludicrous. Obviously Starshot would be applying that power for a much longer time interval, and indeed would be a global undertaking dwarfing NIF. But it doesn't need 100GW of dedicated power plants.

NIF's laser works for only 1microsecond, took over a decade to build and the price was over 3.5 bln. "Peak current for the system exceeds 100 million amperes and the peak power exceeds 1 terawatt (1 trillion watts)" https://lasers.llnl.gov/about/how-nif-works/power-conditioni...

Peak NIF power is actually only ~a nanosecond. And I acknowledged that Starshot would be a global undertaking that dwarfs NIF. I'm just saying that it's wrong to say this is going to require 100 GW of industrial power plants.

Do they not need to run it/them continuously for years for a solar sail approach to make sense? NIF is not a good comparison for exactly that reason.

No, the acceleration phase is only for a few minutes. (I agree with you that if it requires 100 GW of power for years, his criticism would be accurate.) After a few minutes, the probe is already so far away from Earth that you can't focus the lasers effectively on it.

From wikipedia:

> ground-based lasers would then focus a light beam on the crafts' sails to accelerate them one by one to the target speed within 10 minutes, with an average acceleration on the order of 100 km/s2 (10,000 ɡ)

Well damn, then. I missed that when I was reading about it before. https://en.wikipedia.org/wiki/Breakthrough_Starshot

> > high duty cycle

This means that the laser is on most of the time.

If the laser has 70% efficiency, and is on 35% of the time, you need 100 / .7 * .35 = 50GW average (and 100 / .7 = 142GW when lasing). Any storage you have would have to be long term, because you need to lase when the spacecraft is overhead and not at night.

He's using it to mean "high duty cycle over a time-period of minutes". That would be relevant if his criticism was "we don't know how to build lasers that use that much power for that much time" (which is true, currently), but in fact his criticism was that it would require industrial power plants capable of outputting 100 GW (which is wrong). The big power draw of 100 GW over ~minute timescales is for the purposes of launching one probe. After that your power draw stops and you can recharge for as long as you want before the next probe launch. So it's not high duty cycle over timescales longer than hours or days or years.

The thing is, it'd fundamentally different from lasers that have timescale of microseconds or less. He mentioned "capacitor banks."

Launching 20 per day over a 10 minute span each is 200 minutes; you still need 12GW average over the day.

And you need an insane amount of storage-- if 12% of the energy comes directly from the power plant, you need to store 53 terajoules, which is about 60,000 tonnes of lithium batteries per daily shot. (It's more than that, because you're power/discharge rate limited, and it's going to wear out that minimum amount of batteries quickly).

> He mentioned "capacitor banks."

No he didn't. I did.

> Launching 20 per day over a 10 minute span each is 200 minutes; you still need 12GW average over the day.

Where are you getting 20/day from? I didn't see it on this OP link or the Wikipedia page. I'd have guessed more like once a week, which is 100 MW. That's a reasonable size for a gas turbine power plant, which seems like an appropriate expenditure for a project like this.

As a point of reference, the entire globe launches something to space every couple days on average.

> Breakthrough Starshot is a scam and I can't believe that NASA is falling for it.

I know some of the folks involved and they are working on 50 years to launch. It's a long term (multiprong) research project.

I wouldn't call it a scam, I would call it "blue sky" (if it weren't a space project!)

> I wouldn't call it a scam, I would call it "blue sky" (if it weren't a space project!)

Uh, don't we already have "moonshot" for this kind of endeavor, which the name "starshot" already plays on / speaks to?

Seems likely some of the tiny probes will end up in the star so I suppose the name would be appropriate :-)

50 years to launch is a scam. I literally calculated rough trajectories using basic software (NASA EMTG) on how to get a probe out in the rough direction of proxima centuri now. A 50 year timeline on relatively simple things like this should be laughed out of the room. Even though Elon Musk is morally reprehensible now, his “unrealistic” schedules of 2 years may go over for moonshots but even twice the amount of time means the project is done in 4-5 years. Not some ridiculous 2 generation long schedule.

If it takes 2-4% of the world's current electricity production to propel the things alone, we're maybe not ready to do it quite yet.

But we're ready to start thinking about it. Just like we dreamt about rockets going to the moon decades before we attempted going to the moon.

let me understand: you played with a software vaguely related to the topic for 5 mn and it was enough to decide that 50 years to make happen a mission that was never done before is a scam ?

>I know some of the folks involved and they are working on 50 years to launch.

I.e. far enough in the future to safely assume all current employees will happily retire. Isn't that brilliant?

Is it any different from building a cathedral?

It's one of many competing research projects, the sum total of which add up to less than a single shuttle launch. I'm content with a dead end or two.

After writing my comment I looked it up and it’s the financial equivalent of a couple of post-docs for a year or two so I’m a lot less negative.

Post-docs need to write papers. Let them write some about an impossible thought experiment.

If they go past that though…

You announce this kind of project so Congress Critters can get all up in arms about wasteful spending and is something that can then later be sacrificed to make them feel like they've done their job. Meanwhile...you have other projects that have been protected from the axe because the sacrifices of this farcical project.

So you're saying it's like the Queen's duck? https://bwiggs.com/notebook/queens-duck/

This is a tactic as old as the review/inspection process itself. The reviewer always needs to feel like they've done their part, and no product can be perfect on the first pass, right? So put the obvious thing there for them to find, or they will find the thing you are trying to squeak through the process.

I had one building inspector tell me that when he found the obvious thing that was obviously obvious, he would then look for the thing trying to be hidden.

Given 1.6 sq meter 500W solar panels, that's an area of 320 square kilometers. That can be constructed at $1/w at $100 billion.

Also it's not necessary (or technically, it's actually impossible) to have 100% duly cycle laser for propulsion.

It's definitely a megaproject, but it's not a "scam"

For comparison 320km2 is about 2.5 San Franciscos

I asked chatgpt to make a quick demo to demonstrate the size of the 18x18 km2 square:

https://lxe.github.io/map/

Neat. How did you ask it to make that? Do you mind sharing the transcript?

Here you go! https://chat.openai.com/share/024a7117-f3d0-4a23-8186-86eeea...

yeah, when you look at that over the default location over London, it looks big. Moving it closer to the Western US, it looks tiny

Size of a Texas ranch. Call it “reverse solar” and you might have a winner.

Lasers are running at 30-40% efficiency so for a 100GW laser you need 400GW power plant. But the power plant and transmission has some losses so you probably would need 500GW

> That can be constructed at $1/w at $100 billion.

Is this in space or on the ground?

I think for the foreseeable future we'll have better uses for the electricity than powering a laser to accelerate some probes.

I'm curious about whether it's even feasible based on dispersion issues. My understanding is that it's hard to beat dispersion over long distances, regardless of your optics on the laser (and we're talking long distances here...)

Any optics experts able to chime in?

Not an optics expert but I consult for a company that makes lasers. Afaik not currently possible. The beam divergence is only one of many limiting factors

I am not a physicist. Does beam divergence decrease in a vacuum? Could we put the lasers in space at L5 to power the swarm? Or L5 for Venus for the increased solar flux?

Beam divergence, while happening at greater rate not in a vacuum, is a property of how photons propagate, so it's not something that can be eliminated

It's almost entirely a function of wavelength and aperture diameter. It's basic wave mechanics, and is almost impossible to improve upon.

I just recently watched a great video explaining how lithography machines push the limits of what's possible: https://www.youtube.com/watch?v=rdlZ8KYVtPU&t=1151s

Note that the best ASML machines can eke out an improvement on the order of a factor of two at best, not two thousand, which is what would be likely needed to power space probes at light-year distances.

most laser systems get close to diffraction-limited divergence, which is governed by the size of your optics. vacuum vs. air affects it, but the effect is so small that you need sensitive instruments to detect the effect

https://en.wikipedia.org/wiki/Gaussian_beam#Beam_waist

the divergence angle is inversely proportional to your aperture diameter

https://en.wikipedia.org/wiki/Airy_disk

in theory by putting a bunch of laser emitters in solar orbit and operating them as a phased array you can make an aperture the size of the solar system. 1.22λ/d at, say, 400nm and 2 astronomical units of diameter is about 1.6 attoradians of divergence; at a radius of 4.3 light years (the distance to proxima centauri) this works out to an airy disk focal point of 66 microns. optical phased array lasers have been demonstrated https://en.wikipedia.org/wiki/Phased-array_optics but nobody knows how to make one out of satellites yet

i've only glanced at the proposal https://arxiv.org/pdf/2309.07061.pdf and they're proposing 4.1-meter-diameter 3.6-gram probes, which would allow you to use a phased array about a million times smaller, about 5 million km, 13 times the distance to the moon. i don't see where they're proposing how they propose to beam 100 gigawatts onto the probes so i don't know if this is their proposal

dispersion, by contrast, which means wavelength-dependent indices of refraction, doesn't affect lasers and doesn't happen in a vacuum; probably the person who mentioned it was using the wrong term by accident and meant divergence

Less than 100 microns! You can lithograph on Proxima Centauri with that setup. Maybe PKD…

One simple trick von Neumann probes love? Just build more probes remotely using a fricking laser?

> and they're proposing 4.1-meter-diameter 3.6-gram probes,

If you put 100 gigawatts on a 4.1 meter diameter very thin reflector, you're going to melt it. That's 10 gigawatts/square meter. Most of the light misses the targets.

> i don't see where they're proposing how they propose to beam 100 gigawatts onto the probes so i don't know if this is their proposal

A phased array on Earth, several kilometers in scale.

> but nobody knows how to make one out of satellites yet

Probably no one ever will make a huge (solar system scale) coherent power transmission system out of satellites, because of the thinned array curse-- https://en.wikipedia.org/wiki/Thinned-array_curse

It would be cool if we made a very large scale telescope, where a sparse array is less of a problem.

thank you, i didn't know about the thinned-array curse, and now that i understand it i agree that it would not be useful for this

It would definitely depend on the area of the reflective surface. If it is few hundreds of km^2, which is definitely possible as we don't need any special mirror, it would be fine I guess. Even shining beam on patched up aluminium foil would work fine for sailing it.

Although we would need heavier support structure then. And the risk of micrometeoroids collision will be higher.

Beam divergence has to do with your output aperture.

The smaller the hole the laser light comes out of, the more it spreads.

This is a big, several kilometer phased array, which has an effective hole size of a few kilometers. As a result, divergence is less.

100 GW isn't absurd, compared to global human energy consumption of around 20000 GW.

The problem is, until we figure out how to stop burning fossil fuels and fix climate change, the amount of spare power we have is actually a large negative number.

The power is so high so that the probes can be accelerated up to 20% of the speed of light while still in the solar system. The array does not run for long periods of time.

Your comment is a scam, and I despise shills.