Hacker News Reader

>But, contrary to popular belief investors aren’t total idiots who neglect basic questions

Investors, particularly tech investors often do neglect basic (tech DD) questions.

Sometimes it’s because they’re investing in the “team” and neglect DD. Sometimes it’s because the area is hot and they need to make a play... or just because they happen to know or like the people involved.

Whatever the reason, it’s dangerous to assume that the investors must have done their DD (see Theranos for example). Not saying that this is the case here, and often things work out anyway...

Theranos isn't the best example. DFJ invested in their seed, but the vast majority of their funds were less sophisticated non-tech investors. Using them as a poster child for Silicon Valley dysfunction is a bit off to me.

Your broader point is well taken though, there are certainly cases where investors completely skip tech DD. Eg, Juicero's absolutely comical "engineering" that killed them with obscenely high costs.

Theranos had a bunch of larger investors [1] many of them as sophisticated as any tech investor. I’m not saying SV investors are particularly dysfunctional, just that it’s pretty normal for tech DD to not get done, or get done badly. I use the Theranos example because it’s big and public (most disappear without any news).

[1] https://www.crunchbase.com/organization/theranos/investors/i...

While the high engineering costs are definitely comical, wasn't Juicero's failure because nobody wanted their revenue-driving product, the juice bags? Seems like a $600 device could still be profitable (and subsidized) if the core revenue driver was viable.

> We have loops much bigger than this with much more complex magnetics and vacuum components in our particle accelerators, so this wouldn’t be terrifyingly novel tech

Last cost estimate I saw for a reasonably-sized slingshot was $10 billion, which is in the neighbourhood of the LHC's €7.5 billion accelerator + accoutrements cost [1].

Some of the skepticism in this thread, particularly regarding the technology's core viability, is premature. But economic scepticism is warranted. I would be skeptical of a space elevator project without mass-manufactured carbon fibre; I am skeptical of a slingshot proposal without cheap superconductors.

[1] https://en.wikipedia.org/wiki/Large_Hadron_Collider#Cost

Keith Lofstrom wrote about the electronics costs associated with a mass driver back in the '90s [1]. It's dated so some of the base costs have gone down, but it's still a good primer.

[1] https://settlement.arc.nasa.gov/Nowicki/SPBI1SI.HTM

My first thought was that maybe this would lower the cost of particle accelerators like the LHC, due to more mass production of superconducting magnets.

The other consideration is that apparently we now have much more powerful, higher temperature superconductors, at least the talk about MIT's current fusion project says so.

https://www.youtube.com/watch?v=L0KuAx1COEk

Not sure if they are cheaper, but since there would be less bulk and no need for liquid helium, that would be my guess.

So instead of indulging in the “hurrr it’ll never work” superiority stimulus, I’d Like to point out some rays of hope

Thank you.

This doesn't seem impossible. There are some PoCs to be found in old technology.

The shock loads are surmountable. We made an atomic bomb that worked after enduring at least 10KG acceleration. [1]

The aero loads (mechanical and thermal) also seem feasible. We built guided interceptors that launched from sea level and accelerated at 100G to over 7,000 mi/hr. [2] These endured incandescent skin temps within a few seconds of flight.

Aero drag loads can be mitigated mechanically. The aerospike on the Trident missile shows one way of reducing drag by putting a small mechanical probe ahead of the vehicle. [3]

No idea about the economics, but I wish them every success.

[1] https://en.wikipedia.org/wiki/M65_atomic_cannon [2] https://en.wikipedia.org/wiki/Sprint_(missile) [3] https://en.wikipedia.org/wiki/Drag-reducing_aerospike

I am rather hoping the article misspoke and they're actually talking about a centrifugal tether or "skyhook".

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

They aren't. Here's another article.

https://techcrunch.com/2018/02/22/spinlaunch-2/

Indeed! Assuming that your centripital accel. numbers are right, I wonder why not even just use linear accel.

Assuming that you want to exit the tube/system at 5000mph, you could keep acceleration to 9g (tolerable for a human)with less than a 9 mile long acclerator and 18 seconds. If you were ok with 40g, you're down under 4miles and 6 seconds.

(I'm not estimating the deceleration Gs from suddenly hitting MaxQ on exiting into the lower atmosphere. So perhaps that shock puts us up into a high G situation anyway, so might as well go for the smaller real-estate circular acceleration option?)

edit: "tolerable for a human" might better read "survivable for trained humans"

Yeah I'm not up on my supersonic aero anywhere near enough to work out max-Q for their very pointy projectile, but I'll point out that 'suddenly hitting MaxQ' is a jerk problem, not an acceleration one. MaxQ is MaxQ, whether you've been cruising at that speed and air density for hours or just hit it out of a near-vacuum.

I think centripetal / launch acceleration dramatically outweighs air drag though, by analogy to the SR-71. The max thrust divided by its dry mass gives a maximum thrust/weight of about 1g, and that thing could cruise at Mach 3+ which is about halfway to Spinlaunch's 5000 mph.

Their little teardrop projectiles have got to be way lower drag than anything with wings and intakes, and it's probably much heavier for its size than the SR-71 as well, so I can't imagine MaxQ on these things gets past 10-15g.

I don't see the term "MaxQ" very applicable here. The idea of Max Q is that you start at low aerodynamic forces and then accelerate via rocket thrust, with dynamic pressure increasing until the atmosphere density causes it to fall.

For a projectile from this launcher, MaxQ is the second it opens the airlock of the vacuum chamber and releases it to atmosphere. Then it hits a wall of air, and dynamic pressure spikes to the maximum instantly. It is true that you get rid of the centripetal acceleration upon release, so to a degree you trade the centripetal acceleration for drag acceleration.

Converting everything to metric and working out the centripetal acceleration, a radius of 1 mile yields a centripetal acceleration of 232g, to an arbitrary 3 significant figures. A radius of 2 miles yields 116g. The acceleration would drop linearly along the length of the "arm" as you move toward the center, so you could calculate (with calculus, over which I can't be bothered at the moment) the optimal way to taper the arm from root to tip for a given end load.

Drag losses on the arm at a 1 or 2 mile radius would dominate, and ridiculously so. Frankly, roughly 100-230g is not severe at all for a properly engineered vehicle at the payload ratio you could expect to achieve if you get the first 2.2 km/sec "free", as you would in this scheme. That's the majority of a factor of exhaust velocity for a kerosene/lox engine, and all of a factor of the exhaust velocity for a solid rocket motor. Multiples of exhaust velocities are factors of e, so you're probably better than doubling the payload ratio. Perhaps counterintuitively, going from 3 percent mass fraction to 6 percent brings your vehicle size down by half. You go from ~33 times the payload to ~16 times the payload. And, while others are talking a good game about drag, this saves you loads in gravity losses, as the rocket is spending far less time fighting gravity with this initial boost. (1g for two minutes is 1.2 km/s, give or take centripetal effects from your curved launch trajectory. Not negligible, at all.) The other commenters seem also to be mostly neglecting that the motor/engine will very quickly reach near-vacuum conditions, not immodestly increasing its ISP.

Much better than a mile-long tether would be, in my not-so-humble initial opinion, to reduce the tether length by a factor of 10 or 100 from "a mile", to give a centripetal g load of 2k-20k g. Again, calculus tells you how to taper your (probably carbon fiber, possibly consumable, most definitely streamlined) attachment rod. And don't bother with the complexity of a vacuum. The Nike program launched vehicles at Mach 10 in the lower atmosphere. That's 3.4 km/s, 1.2 km/s faster than SpinLaunch's proposal. Given that heating goes something like the 4th power of velocity, Mach 6 or so is going to be much, much less severe than the Mach 10 environment encountered by Nike. Why, SpinLaunch vehicles won't even get to glow white hot!

As I said above, I highly doubt it’s a literal arm / sling being used. I think it’s a magnetically guided and accelerated loop, so the payload is the only part of the system moving at all. You’re right the drag/tension on an actual arm would be ludicrous, and would certainly force you to a smaller radius.

> at 1 mile you’re looking at ~300g acceleration, 150g at 2 mile.

Could you explain? Centripetal acceleration is v^2/r, so naively I'd think something moving at LEO orbital velocity in a radius of 1 mile requires ~4,000g because LEO has a radius ~4k miles under 1g.

Article says 5000mph, so the launch is well below orbital velocity.

That is only about 1/4 of the velocity needed for low earth orbit [1]. The goal here seems to be to replace the first stage of a rocket by flinging it above most of the atmosphere.

[1] https://en.wikipedia.org/wiki/Low_Earth_orbit#Orbital_charac...

Yes — rocket is still needed for circularizing burn (i.e, achieving orbital velocity).

(Which makes sense — if you managed to accelerate something to orbital velocity at sea level, it would (1) shed much of that speed before it actually reached a near-zero-atmosphere altitude and (2) burn up.)

25% the velocity but 1-.75^2 = 43.75% of the energy.

Further the rocket equation is a bitch so your saving even more fuel. The problem is you also end up with a lot of atmospheric drag and heat so the final savings are not as huge.

I don't think you did that right. The energy to get to (1/4)V is (V2)/16. The energy to get to V is (V2)/1. So your payload is only 1/16th of the way there or 6.25%.

Agreed about the rocket equation though - that first 6.25% of payload energy is much harder than the rest

If it makes you feel better picture the same energy accelerating the space shuttle it's solid rocket boosters and that huge fuel tank to some velocity vs just the shuttle to a much higher velocity. Or say to yourself it's all relative. Like tossing a ball between you and your friend on a train does not take more or less energy from you as the train is moving at higher or lower speeds, but it does take different amounts of energy from the train.

Rockets get energy from their fuel directly from burning it, but also from the kinetic energy of their fuel. So in space a rocket that can add 100km per hour of delta V from fuel before running out can do that at 0MPH , 1000 MPH, or -10,000MPH all the way up until relativity becomes an issue.

So, first find out how much speed/energy you need as a baseline it's velocity squared (100% velocity)^2 = (1v)^2 = 1e = 100% energy. Now instead of that we need to go from V1 to V2 you need (V2 - V1)^2 energy. That's (100%v - 25%v)^2 = (1v-0.25v)^2 = (0.75v)^2 = 0.5625e so we still need 56.25% as much energy, and we gained 100%e - 56.25%e = 43.75% energy.

PS: Unfortunately, these things don't start in space so we need to consider air resistance.

One question I've never been able to figure out with this concept is what happens when this fast moving launch vehicle hits the stationary air. Would it be like hitting a wall? I'm a bit physics ignorant.

You'd probably have a stationary launch ramp exit, built up the side of a mountain. If you can put the top of the tube at 10-15k ft that will reduce air drag _substantially_. Even a 5000ft exit would be in substantially thinner air.

I have a hard time believing that the reduction is anywhere near substantial enough. Friction heating was a major issue for high speed aircraft like the SR-71, and they "only" fly at ~3500 km/h at 20+ km altitudes.

The world record top speed at sea level is less than 1000mph. It's like flying in soup by comparison.

I assume you'd expose it to gradually increasing air pressure on launch somehow.

You're talking about minimizing jerk:

https://en.wikipedia.org/wiki/Jerk_(physics)

To my knowledge, there is no reason to minimize jerk for machinery. Only the maximum acceleration value matters.

>> To my knowledge, there is no reason to minimize jerk for machinery.

Maybe not for a blob of homogeneous metal parts, but with complex structures that jerk can multiply into bad things. As the nose hits air (the sudden high G-load) a shockwave moves down the vehicle at the speed of sound (sound through metal, not air). Shockwaves can do funny things in complex structures. Imagine that this thing had a solid rocket motor for circulization. As the shockwave moving down the metal walls at one speed, slower through the fuel, the fuel might crack in ways that are not-good for a rocket.

Fluid-filled things like pipes can also be subject to water hammer-like effects when g-loading changes abruptly as opposed to a gradual onset. The fluid gets moving quickly as the pipe stretches at one end, then must stop abruptly as the pipe hits its limit.

Is https://en.wikipedia.org/wiki/Max_Q relevant?

That's actually a very interesting question. Are there electronics or machinery affected by jerk? How would that work?

Another comment mentioned "hammer-like effects"... I can see it. Mentally, I picture a half-filed milk jug being pushed around the counter. Smooth acceleration changes cause the water level to safely assume the new steady-state, but high jerk values can cause the container to collide with the water hard. If you have hardened electronics, hopefully none of the components are loose.

It sounds a lot harder to make a rotary sling, the arm of which will be put under extreme loads, than it does to create a long linear motor.

> I worked at Hyperloop One for a time,

You just lost your credibility.

If you won't stop breaking the guidelines we'll ban the account.

https://news.ycombinator.com/newsguidelines.html

Oh fuck off already! This place is full of idiots, that I am not banned already is an insult to me.

Why?

Long ago I heard that the reason why the Jules Verne space cannon was infeasible was not so much the acceleration as the aerodynamics; 5000mph is "hypersonic". I'm not quite clear whether the spinlauncher's rocket is just for circularisation or whether it's needed to clear the atmosphere too?

Why are you skeptical? HN community always been this negative, or even more! Remember the initial Drew's DropBox idea that majority bashed him that its a stupid idea because... everyone would rather built its own software? Drew Stock in Drop is $1B worth right now.

If anything, I think its opposite - the more HNers are negative the more your idea has a strong standing!

Coincidentally, on the subject of catapults, just recently I was bashed by "always_good" that "I'm sure a bunch of children also wonder the same thing" [1]

[1] https://news.ycombinator.com/item?id=17174196

Edit: made it less personal :)