# Tech Tuesday – Spooky Action Edition

Oscar Gordon

A Navy Turbine Tech who learned to spin wrenches on old cars, Oscar has since been trained as an Engineer & Software Developer & now writes tools for other engineers. When not in his shop or at work, he can be found spending time with his family, gardening, hiking, kayaking, gaming, or whatever strikes his fancy & fits in the budget.

### 51 Responses

1. For quaternions, multiplication isn’t commutative: i * j = k, but j * i = -k. For octonians, it isn’t even associative: (a * b) * c doesn’t in general equal a * (b * c).

(As the article says, I see. )Report

• Oscar Gordon says:

See, that right there would constantly trip me up if I tried to work with them. I am in awe of anyone who can do useful work with them (and maybe this approach will lead to useful work).

I feel like I should also add in a comment that the particle physics community is rather split as to whether or not such mathematical approaches to particle predictions are useful and a constructive use of time; as opposed to smashing things together at relativistic speeds.Report

• veronica d says:

If it help, the underlying mechanism for quaternions is literally the same as rotation matrices, just in slightly different form. (In fact, they have to be. They are describing the same process.) So if you can handle non-commutative matrix math, then quaternions are no harder.

For octonions, the analogy is a Lie algebra. Again, they aren’t actually harder. The complexity is essential to the natural phenomena they model.

Why use them? That is a good question. It’s usually nice to find some “algebra” for the objects you are studying. For example, Poison brackets are doing “algebra stuff” for rotational dynamics, just as the commutator does in quantum theory (in fact, they are the same thing really).

After you learn a few of these “algebras,” you start seeing the same patterns in many different places. It’s cool.Report

• Oscar Gordon says:

OK, I can handle rotation matrices, although they annoy me enough that once I figure them out, I code the little bastards into at least a spreadsheet.

Lie algebra is… I’d have to take a class, or read a Schaums guide, or something, but it sure looks similar to the stuff I saw underlying the Supersymmetry work.Report

• veronica d says:

I think one very basic “intuition pump” for this stuff is to note the following: the 3-space rotation matrices are non-commutative. This is obvious. If you rotate a book first on the X axis, then on the Z axis, you get a different result from if you do that in the opposite order. You can probably picture this.

Now, think of how the cross product works. Think of all the engineering equations that become way simpler if you use the cross product. In a sense, the cross product makes rotations “kinda algebraic.”

Now consider this: the finite rotations are non-commutative. However, the infinitesimal rotations are, to first order, commutative. You can see this by looking at them in matrix form.

That’s neat. But why?

I think you can use intuition to see why. After all, one finite rotation can follow another. But how does one infinitesimal follow another? Sure, if you’re doing a numerical simulation of a system of differential equations, then each time slice will follow another, but those are estimates. An infinitesimal move on X can neither follow nor precede an infinitesimal move on Z, unless there is some finite passage of time between them.

Neat. But there is more: that difference shows up in the algebras, yet the two algebras are related. But how?Report

• pillsy says:

They are similar, in that Supersymmetry involves specific kinds of Lie algebras.

Lie algebras (and the groups they generate) are often used to describe symmetries of physical systems, ones that can be continuously varied (like rotational symmetry)! Physicists find those particularly interesting because a system that has such a symmetry will have an associated conserved quantity.

In this particular case, rotational symmetry is associated with conservation of angular momentum. (If you had Lagrangian and/or Hamiltonian mechanics inflicted on you at some point this might sound familiar.)Report

• Oscar Gordon says:

Lagrangian, so yeah.

Like I said, quarternions I get, I just don’t use them (although I do use rotational matrices all the damn time).

Octonions are new to me. I get the general concept (if you can grok quarternions, it’s not a stretch to get to octonions), but figuring out to do math with them…

Yeah, I’m going to need to read a book, and figure out how to visualize them in space, because shit always confuses me until I can find a way to visualize it in a meaningful way.Report

• veronica d says:

Warning: you cannot really visualize them, since they are higher dimensional.

The first step would be to understand continuous symmetry groups (e.g. Lie group), and then Lie algebras. Next learn about quotient groups, and in turn the whole idea of “quotient structures” (such as ideals in a ring).

A good concrete example of this stuff is to then learn how Hamiltonian dynamics follows from Lagrangian. Then learn about how Poisson brackets work. (Note, they are an example of a Lie algebra.)

Then … well that’s where we hit the limit of my knowledge. But the point is, these physicists are trying to find a single algebraic structure that contains all of the symmetry groups from the standard model as either sub-structures or quotient structures. After doing that, they can look for other symmetries in the main algebraic structure and see what cool stuff that might imply about nature. Those other symmetries my suggest experiments.

Behind all of this is the intuition that nature should be singular, unified, and somehow mathematically inevitable.

But is that a reasonable thing to expect?Report

• Oscar Gordon says:

I don’t need to visualize them as they necessarily exist, but I do need to bring them down to 3, maybe 4 dimensions.

Kind of like how a tesseract doesn’t exist in 3 dimensions, but you can visualize it in 3D if you are willing to sacrifice accuracy a bit.

As for the rest, one of these days…Report

• Oscar Gordon says:

PS Thank you, Veronica, for at least pointing me in the right directions. I promise, someday I’ll actually have time to wonder down that path.Report

• veronica d says:

Same here, actually. I understand just enough about this stuff to sound smart on the internet, but I don’t really understand it the way I would like.

It seems the more I learn, the more I learn about how much I haven’t yet learned.Report

• Doctor Jay says:

When I read about Cohl Furey and this approach before, it seemed like it would be promising. I didn’t know the connection with Lie algebras, but I haven’t really studied them.

Also, “Cohl Furey” is the coolest name ever, and she has a striking look on top of that. I designed a superhero inspired by her. Codename “Cold Fury”.Report

• Oscar Gordon says:

Her superpower is to bury your mind in the cold logic of an 8 dimensional mathematical proof, thus rendering you inert while you tried to figure it out.Report

• For octonions, the analogy is a Lie algebra.

Yeah, that’s the ticket!Report

2. Michael Cain says:

TT13: Composite rebar dates back to at least the 1980s. The big delays in its adoption (aside from cost) were all the basic materials testing, then the development of the rules for how to use it, then the slow slog through final standards bodies approval, then getting it into municipal construction codes.Report

• Oscar Gordon says:

That explains why this is a story from Down under?Report

• Oscar Gordon says:

PS I used to work for a Civil Engineering department, the basic material testing doesn’t actually take that long, so I’d put it down to cost and regulatory hurdles.Report

• George Turner says:

I remember when civil engineers first stacked a rock on top of another rock. Decades later, another civil engineer came running into the cave and excitedly said “Og has stacked a third rock on top of the first two! It’s revolutionary!”

But looking at the pace of SLS development, aerospace engineers don’t have much room to poke at them. “We someday intend to stick an upper stage on top of the lower stage!”

Some of the ultra-high strength concretes dispense with rebar altogether, gaining extra tensile strength from finely shredded fibers.Report

3. Michael Cain says:

TT14: They had to solve a bunch of problems to make this work: optimization of fiber orientation, true 3D printing paths, and a big enough working volume. The print head is mounted on a six-axis robotic arm. The company is currently shipping carbon fiber bicycle frames that they claim are cost-competitive with hand-constructed frames from China.Report

4. Oscar Gordon says:

Well damn, I just looked at the calendar and realized today is 50th anniversary of the moon launch.Report

• You know, I didn’t remember that. The landing on the 20th, of course.Report

5. George Turner says:

TT09’s beyond the black body device:

The technical paper says:

We report a maximum NFRHT enhancement of approximately 28.5 over the blackbody limit with devices made of millimetre-sized doped Si surfaces separated by vacuum gap spacings down to approximately 110 nm.

28.5 is huge compared to 20, but it’s obviously less than 30. Unfortunately, the part of the article that might say whether it’s 28.5 times more, 28.5% more, 28.5 Megajoules more, or 28.5 electron volts more is behind a paywall. Anyway, it’s bigger than 20.Report

• Michael Cain says:

Here’s a link to a full copy of the paper.Report

• George Turner says:

Thanks!

I’m hoping the more efficient heat to electricity tech will enable a future of plutonium powered personal devices.Report

6. George Turner says:

Another tech item that took up most of my time yesterday was SpaceX’s announcement of what blew up their Dragon 2 on the test stand.

It seems a slug of liquid N2O4 formed in a helium line due to a slow leak in a valve. When they were 0.1 seconds from ignition, that slug was blown down the helium line like a bullet, destroying a titanium check valve. Titanium, interestingly enough, is also a shock-sensitive explosive if soaked in N2O4. NASA has a 1961 technical paper on it, and SpaceX, NASA, the FAA, the NTSB, and the Air Force now know a whole lot more about how to McGyver a titanium check valve into a detonator, and SpaceX is going to switch to burst disks for the abort system.Report

• DensityDuck says:

“SpaceX is going to switch to burst disks for the abort system”

which everyone else already used for pressurized helium systems but, y’know, SpaceX

(oh, and that’s three vehicles now SpaceX has blown up because of helium system foofery…)Report

• DensityDuck says:

” Titanium, interestingly enough, is also a shock-sensitive explosive if soaked in N2O4. ”

I think it’s not so much “explosive” as it is “it develops cracks that make it more likely to fracture or shatter”. So the frozen slug of NTO hit the check value and caused it to fail, which allowed uncontrolled mixing of fuel and oxidizer, which is what caused the explosion.

SpaceX statements are suggesting that the titanium itself is what caused the explosion, which is…concerning, because that’s not a thing that happens. Like, NTO tanks on spacecraft are typically made of titanium, if they were going to explode we’d know by now…

And again, that’s why everyone else makes check valves and oxidizer tubing out of steel…Report

• George Turner says:

The stress cracking of titanium in N2O4 is a different problem, and probably unrelated to what occurred with the check valve.

The valve got smacked by a projectile and the impact broke it. If you hit N2O4-soaked titanium with a hammer, imparting sufficient impact energy, the surface layer chemically explodes, even if the titanium doesn’t fracture. If the titanium does fracture (which is what eventually happened during the impact), the freshly exposed layer of titanium can also explode. However the explosion is not self-sustaining, unlike an alkali metal in water. Only the directly impacted surface or the fracture’s face blow up. Aluminum doesn’t do that, in part because aluminum has much greater heat conductivity so the impacted or fracture surfaces don’t get as hot.

In at least the initial explosion on the Dragon, the oxidizer was N2O4 and the fuel was titanium. No hydrazine seems to have been involved.

Interestingly, hot titanium and water also violently explode, which is something that has killed a lot of people. Industry set up safety committees and established standards to reduce the death toll in titanium processing, since some of the explosions were equivalent to 500 pound bombs. It took a while for us to realize that molten titanium presented an explosion hazard quite unlike molten iron and molten aluminum. Hot titanium strips oxygen out of water molecules to form hot titanium oxide (in an exothermic reaction) and free hydrogen.

There was a recent fire in a titanium scrap facility in Detroit that created at least two large titanium explosions. One of the explosions injured three firefighters and rained molten metal down on the rest, who were hundreds of feet away.

So, it’s a bit of rather obscure knowledge, but one that bit them.Report

• DensityDuck says:

I think I’m gonna need to see a cite for this one because I can’t find anything about titanium “chemically exploding” when exposed to NTO. AFRPL-TR-76-76 talks about stress corrosion cracking (several other documents concur, pointing mostly to the formation of nitric acid when the oxidizer reacts with water) but nothing implying there’s an explosive reaction hazard. Anything I can find that talks about titanium “burning” involved either heating past 600 degrees (such as in a jet engine) or titanium powder (explosions are a significant hazard for 3D metal printers, it turns out).

And, as I said earlier, nothing in my experience suggests that this is an issue. Titanium has been the standard material for oxidizer tanks for twenty years so, again, if titanium exploded when it contacted NTO we’d have known about it by now. (And if you’re asking “well why don’t those tanks have corrosion cracking problems” that’s a good question, and they answer it by taking a great deal of care to prevent water contamination of the tank before putting NTO into it; they use water instead of alcohol to clean it, and keep a dry nitrogen purge environment inside it until they actually need to put the oxidizer in there.)

It was corrosion cracking in the valve, and it allowed uncontrolled mixing of the fuel and oxidizer; technically a deflagration rather than an explosion. It’s what happened on Flights 19 and 29 when the helium tank failed and the expanding gas ruptured the two upper-stage tanks. There’s no need to invent exotic failure modes of materials that nobody ever saw before to explain what happened, it’s actually a well-understood issue with the material, and it’s rather surprising to me that this is having to be re-learned.

Like, to the point where if I were going to fly astronauts on the thing, I’d want to sit down with every drawing and procedure side-by-side with a flight-proven equivalent, and anywhere they aren’t the same ask for a complete explanation as to why SpaceX is different and how they’re sure it’ll work.Report

• DensityDuck says:

aw dang, the edit window expired as I was typing. Second paragraph parenthetical aside should say “they use alcohol instead of water to clean it”, I swapped water and alcohol while drafting.Report

• George Turner says:

I cited this paper elsewhere:

1961 NASA Technical Paper

Impact Sensitivity of Metals (Titanium) Exposed to Liquid Nitrogen Tetroxide.

Abstract:
The purpose of this investigation was to determine the impact sensitivity of commercially pure Ti, Ti alloy 6Al-4V, and precipitation hardened 15-7 Mo stainless steel when exposed to liquid N2O4. An explanation is given of the probable mechanism of the limited ignition resulting from impact.

Doug Jones, cofounder of XCOR aerospace, chimed in, saying:

Titanium can also do lower valences, but at Ti+4 has potential of -1.899, more negative than Al3+ at -1.662 and making it an enthusiastic fuel. But critically, Titanium is a lousy thermal conductor at around 20 W/m.K vs Aluminum alloys at around 160 to 180. Al is harder to ignite because it can carry the heat away, not so titanium. Breaking or grinding Ti can produce an impressive shower of sparks, and in high pressure liquid N2O4, it’s all over but the crying.

In a system that would only be used once during abort, burst discs make more sense than check valves- guaranteed leakproof, simple to install, lighter weight, no soft goods to degrade over time…

And another rocket guy said

When you crack the oxide coating of a piece of titanium submerged in liquid oxygen it will ignite, so it’s not just NTO. I’d like to know the manufacture date of the valve, and also how cracked it got (did the metal crack, or just the oxydation layer?). It takes a few years for the coating on titanium to reach the maximum of 25nm, but it starts out at around one nm.

Somewhere among my artifacts I have a discarded gas turbine vane made of titanium. It was binned because one of the corners somehow got bent under. I have no idea how it happened.

Like aluminum and many other metals, titanium forms a protective oxide coating. However, it’s hardness and strength, as used in aerospace, means that if it does break, there’s a whole lot of stress and energy involved, and it’s poor thermal conductivity means that heat stays concentrated at the impact or fracture site, so there’s extremely hot, energetic, freshly exposed pure titanium in direct contact with an oxidizer.

The implication I got from the SpaceX report, though it’s not directly stated, is that all this bad stuff happened inside a full, high-pressure NTO tank, and however much extra pressure the little detonator added, it was just too much, too fast.Report

• DensityDuck says:

Thank you for providing the paper. Unfortunately I don’t think it actually helps the assertion that titanium becomes “chemically explosive” when exposed to NTO.

The paper suggests that:

A) the ignition caused by the impact is limited to minor melting in the area of the impactor
B) the ignition doesn’t propagate through the oxidizer or across the metal surface, even when the impact region is soaked in oxidizer
C) alloys are less likely to ignite than pure metal (and aerospace components are almost never made from pure metal, certainly not ones that are part of a pressurized fuel system)

Sorry, dude, your idea of a titanium valve being hit by an ice cube and exploding like a bomb due to some hitherto unknown interaction of physics just isn’t there. They say in the report that at one point they heated titanium rods glowing-hot and shoved them into buckets of NTO and couldn’t make it burn. Their conclusion is that the “ignition” that results from impact is most likely from NTO caught between the impactor and the titanium, and doesn’t have anything to do with the metal at all.

“The implication I got from the SpaceX report, though it’s not directly stated, is that all this bad stuff happened inside a full, high-pressure NTO tank…”

If these check valves were inside the tank and they’re going to replace them with burst discs, that makes even *less* sense.Report

• DensityDuck says:

And my point in all this is not that SpaceX did the right thing, my point is that they’re kinda selling this as “weird crazy stuff that nobody could have forseen, we just fix this one issue and EVERYTHING IS FINE” which isn’t really the case. If you’re using a blow-down system to pressurize your tanks then you need to take into account what happens if stuff other than gas gets into the pipe!Report

• George Turner says:

Well, that’s what SpaceX, the NTSB, the FAA, NASA, and the Air Force concluded about the accident: A titanium check valve explosion in an NTO environment. And they said that their months of experiments blowing up titanium valves were quite illuminating.

Are you suggesting it was something else?

There are also many ways to make a failure that massively increases the surface damage compared to a simple percussive impact striking normal to a surface. For example, if you have a slightly over-sized central body, such as a plug, and you violently force down the center of an undersized pipe, and the subsequent explosive action in the fracture zone acts to increase the force driving the plug like a reamer, even to the extent that it blows the valve in half, things could get really exciting.

And on top of that, the baseline helium pressure coming down the pipe is probably up around 4,000 psi, well over 250 atm, so the reactions aren’t occurring at anything remotely close to atmospheric pressure, which is what the 1961 NASA study was examining.Report

• DensityDuck says:

“, that’s what SpaceX, the NTSB, the FAA, NASA, and the Air Force concluded about the accident:”

Cites please, old guy, these things need fucking links or you’re making shit up. As far as I can see there have been no statements concluding anything from “the NTSB, the FAA, NASA, and the Air Force”. No, “observers were present for the investigation” does not count as a statement by the organization.

“A titanium check valve explosion in an NTO environment.”

I agree that a check valve failure occurred, but I don’t agree that some kind of “chemical explosion” was what happened. I think that nitric acid was formed in the NTO when the tank was contaminated with water and not completely dried before being filled–a known issue. That nitric acid attacked the titanium material and caused it to become brittle–also a known issue. And NTO leaked into the helium system and froze, and when the helium system was activated (reducing the temperature and allowing boil-off to occur) this frozen chunk was shot into the fractured check valve at high speed and caused it to fail, which permitted helium to enter the NTO tank at higher rate than planned. This overpressurized the NTO tank and caused it to burst, which resulted in unconstrained mixing of oxidiser and fuel, with the result being a fire and explosion.

NTO isn’t stored at high pressure or cryogenic temperature — that’s really the whole point of the stuff, that it’s stable at most operational temperatures, it doesn’t even boil until 70 F — so you wouldn’t need a lot of helium going in at 4000 psi to pop that thing like a balloon. That’s what did for SpaceX’s Falcons, actually.
And NTO actually isn’t all that explosive just by itself. It just kinda sits there unless you mix it with something, and that will explode like all hell.

There was no mysterious interaction of physics here, it was a failure that resulted–yet again!–from SpaceX not paying attention to industry practices and experiencing the same problems that the industry had already learned to solve. Which is what I would expect from a company that’s being stood-up from zero, but it doesn’t inspire confidence in the ability to safely and reliably fly humans.

“Well if you agree that it was the check valve breaking then why do you care so much” I care because if you’re going to wave your hands and say “whurh, nobody could’ve forseen this” then it absolves you of the responsibility to do a good job. It’s the investment banker saying “well sure I put all my money into one energy company, but nobody could have forseen that Enron would crash” as though the problem were volatility in the energy market and not failure to diversify.

“the baseline helium pressure coming down the pipe is probably up around 4,000 psi, well over 250 atm, so the reactions aren’t occurring at anything remotely close to atmospheric pressure, which is what the 1961 NASA study was examining.”

That was your cite, so, if you want to say it’s not applicable go right ahead.Report

• George Turner says:

I assume you read their update on the investigation, which they released some time ago. They also had a press conference with Hans Koenigsmann from SpaceX and Kathy Lueders, who runs NASA’s commercial crew program. The NASA folks who are going to be in the Dragon 2 really want to know what happened, which is why NASA was part of the investigation team.

They had great high-speed footage of the explosion, and had no trouble at finding the valve. The key thing about the valve was that it had burned, and that was the obvious clue- not stress cracking, burning.

Nitric acid is such a non-problem that on the Shuttle they added about 2% nitric acid to the NTO to reduce corrosion rates, and titanium is commonly used to handle boiling nitric acid where stainless steels would have problems. However they don’t use it for red-fuming nitric acid. The corrosion rates of titanium in these nasty environments is usually under a mm per year, and the Dragon 2 probably only been fueled up that morning. Nobody from NASA or SpaceX even mentioned seeing any corrosion, anywhere.

this frozen chunk was shot into the fractured check valve at high speed and caused it to fail, which permitted helium to enter the NTO tank at higher rate than planned. This overpressurized the NTO tank and caused it to burst, which resulted in unconstrained mixing of oxidiser and fuel, with the result being a fire and explosion

Why do you say the check valve was fractured prior to the incident? What would fracture it, the weak little spring that maintains the seal? For stress cracking, you need stress. There isn’t any to speak of in a check valve until it’s having to resist a high reverse pressure.

Second, it wouldn’t matter if the valve was fractured because it’s a check valve. It’s job is to make sure that high pressure NTO doesn’t flow into a depressurized helium line. Once the helium line is at a higher pressure than the tank, the check valve opens, and from an engineering standpoint, might as well not even exist aside from probably insignificant the flow coefficient, which of course would have been selected to allow the engines to maintain full thrust all the way to fuel depletion.

The check valve does not regulate the pressure in the NTO tank, and has really nothing to do with that pressure. That’s why they can replace it with a burst disk. A failed check valve wouldn’t over-pressurize the tank.

Pressure increases the burning rate and decreases the ignition temperature of titanium, and in high pressure oxidizing environments hot titanium combustion becomes self-sustaining at much lower temperatures than are required at lower pressures. So the pressure adds to titanium’s burning rate, and the effects of impact ignition at high pressure will be much greater than the effects of impact ignition at atmospheric pressure.

And NTO is commonly stored for years under pressure, on spacecraft. To run the engines in a simple blow-down system, the NTO and hydrazine tanks are pressurized. After the burn is done, the tanks remain pressurized because they have no way of pumping the pressurant back into some other tank, and they’re not going to dump it overboard because that would cost mass, requiring them to carry enough pressurant to completely repressurize the tanks after every little course correction and attitude adjustment. Once the tanks are pressurized, they stay that way, sometimes all the way to Mars or beyond.

So this gets back to what they found, which was a check valve that had ignited. That big giant red-cloud of NTO that appeared over Florida? That was NTO, not the exhaust products of NTO and hydrazine combustion, which is colorless nitrogen, oxygen, and water.

As Kathy Lueders said, they were blessed to have this failure happen during a test. I said the same thing the day of the explosion. The failure mode could have easily gone undetected until it occurred during a flight.Report

7. DensityDuck says:

[TT10] the idea seems interesting (and I’ve seen it before, in other concepts) but my concern is whether they can meet crashworthiness standards with a passenger compartment that’s separable from the chassis. I’m assuming that the “swap” takes place at the end-user facility rather than in a factory, which means that whatever process they have for locking the chassis and body together needs to be able to be done by Just Some Dude but still keep the car from disintegrating in a wreck.Report

• Oscar Gordon says:

Or the shell itself needs to be a crash pod.Report

• DensityDuck says:

The problem is keeping the crash pod from detaching from the chassis and flying down the freeway at 75 mph. It’s possible to design a vehicle that keeps the occupants alive after that, but we’re talking full-body encapsulation and multi-point harnesses…Report

• Oscar Gordon says:

If my seat in a 737 is rated to keep me attached to the floor under 15Gs of acceleration with little more than some bolt plates, it’s not a question of can we do it. As you said, it’s a question of how do we make sure Cletus doesn’t screw up the connections when he’s doing a swap while suffering a hangover?

But I’ve seen designs for some pretty simple, but solid connectors that could do it (probably need 6 to 8 of them). You just need to make sure each connector has a sensor that can tell when it’s engaged properly, so you can’t forget one and still send the car out.Report

• DensityDuck says:

see my comment re: “end-user facility” and “Just Some Dude”. The seat that stays attached to the floor is put on by a trained technician using calibrated and tracked equipment, with a quality inspector looking over their shoulder the whole time, and it’s just put on one time.

“You just need to make sure each connector has a sensor that can tell when it’s engaged properly, so you can’t forget one and still send the car out.”

Which can certainly work, but it makes the whole thing a lot more expensive, and I thought that reducing overall expense was one of the ideas behind this thing…Report

• Oscar Gordon says:

Depends, is it cheaper to have sensors, or perhaps actuators (let’s take the hungover human out of the equation) that lock the body to the chassis; or is it cheaper to have a class of vehicles that sit generally unused, just in case a customer needs one.

If the desire is to replace most vehicles on the road with autonomous dispatched vehicles, you gotta handle those instances when someone just needs a damn pickup truck, or a passenger van, so one way or another you have to spend for that option.

I’m betting that a relatively inert shell with a handful of hard points that the skateboard can grab and lock onto is going to be significantly cheaper than a complete vehicle, not only to purchase, but to maintain as well.Report

• DensityDuck says:

“I’m betting that a relatively inert shell with a handful of hard points that the skateboard can grab and lock onto is going to be significantly cheaper than a complete vehicle…”

Sure it is! Then you have these sensors and actuators and whatever other complicated mechanisms lock the body to the chassis, which need to be tested for about a million cycles (both as independent components and again as a functional system) to be certified, and if you change the design it’s back to zero for your test program. It really does get expensive. There really is a reason that people haven’t already done things this way (I mean, we could already have a cab-only truck body with either a cargo bed or a “passenger pod” that converted it to a bus or taxi).

I’ve worked on this stuff, this is honestly not something you can just thought-experiment or handwave away.

“If the desire is to replace most vehicles on the road with autonomous dispatched vehicles, you gotta handle those instances when someone just needs a damn pickup truck, or a passenger van, so one way or another you have to spend for that option.”

You handle pickup trucks by having the Lowe’s offer them for rental, which is what they do now. Passenger vans are covered by your entire fleet being vans (after all, if you aren’t the owner of the car then why do you care how it looks, and if you want a fancy car to be driven around in then you’re probably rich enough to hire a driver who has their own fancy car).

And the thing is, manufacturers already MAKE standardized chassis with different bodies. The Toyota MC platform shows up as the Corolla and the Camry, the Prius, the RAV 4, the Celica sports car, several different vans, and I’m pretty sure if Toyota still made small pickups they’d have put one on this chassis as well.

I mean, there’s some nice CAD models and it’s amusing to imagine just swapping out a “pod” to get a different car, but I really think this is one of those things where there’s a lot more cost in doing what they plan to do than they imagine, and they aren’t really saving a lot by keeping the parts that they’re keeping.Report

• Oscar Gordon says:

I think a lot will depend on how the business/dispatch model winds up working. If the big box stores maintain small fleets of trucks, or if UHaul does (I’d bet more on UHaul doing something like that, then the big box), then sure, your commuter dispatch center won’t bother with anything but vans.

Then the swapping body pods become the domain of high end customers who want to be able to ride around in style, or something.

So yes, I see your point.

As a manufacturer, I can see the skateboard approach as still attractive, even if the pod linkages are depot level attachments (not something you want a chucklehead messing around with), provided, as Michael notes downthread, that the suspension kit is flexible enough to handle a variety of loadings.Report

• veronica d says:

How much actual variance is there in the day to day loadings that “typical people” put on their vehicles? Is there research on this?

(The rental companies and UHaul, and similar, probably have this data.)

I used to own a 1/2 ton pickup, and few times I actually loaded it down pretty well. However, those were all for work. I was hauling big spools of cable — the kind that took a forklift to get into my truck. (Getting the spools off the truck was much easier and kind of fun.) But the point is, those times were pretty rare. Most of the time the bed was either empty, or I was hauling the kinds of supplies that we could shove on by hand back at the shop. It wasn’t heavy by truck standards.

But really, a lot people will need a van/truck to bring home a new couch or big TV or whatever. That’s bulk, not load. Other folks will be hauling a bunch of people for a church picnic or wedding or whatever. Other folks may be taking someone on a date.

To a certain degree, those all seem like similar loads — at least all within the “normal passenger vehicle” range.

Hauling cable? Towing a big ass engine lathe that fits on a trailer — those sorts of things will probably need a different power train and or different suspension.Report

• DensityDuck says:

I think I might have some comments that got into a moderation queue, because I included links in them; could someone please check?Report

• DensityDuck says:

Ah, they are present now. Thanks to whoever did that, even if it’s just time and caching 🙂Report

• Oscar Gordon says:

No problem, happy to get them out of Limbo.

Busy day today, I will reply to everyone tomorrow or Friday.

Or at least sometime before the heat death of the Universe…Report

• JS says:

I’d imagine it has less to do with “keeping your seat in place” then with ensuring the bolt-on chassis is capable of handling crashes from arbitrary angles in a way that sheds kinetic energy, rather than letting debris (from the other vehicle or the chassis itself deforming, shattering, or breaking) into the human occupied space.Report

8. Michael Cain says:

If the desire is to replace most vehicles on the road with autonomous dispatched vehicles, you gotta handle those instances when someone just needs a damn pickup truck, or a passenger van, so one way or another you have to spend for that option.

I would have thought rent complete units, since the skateboard suitable for a compact car is not going to have suitable suspension for a van seating eight or a ton-and-a-half pickup.

Passenger protection will be… different. I’ve seen some system analyses for vehicles with a similar distribution of drive train and power components. The best approach is quite different than in a current design. The shell needs to be really rigid, without crumple zones, to protect the interior space. Remove the dash and steering wheel, protect the hell out of that repurposed space, and use it as a “ride down” zone for the passengers.Report

9. George Turner says:

Interesting Starhopper fire

Elon says there was no damage because Starhopper is stainless steel.Report