EV virtue signallers learn a rather harsh lesson

https://youtu.be/ZxjoHUcZdwk

tomo,
Hadn't thought of the Hyperloop tie-in. It was close to truth in naming though: pure hype of a loopy idea.

As to needing fibre-optic networking in a car, I don't see a need for its throughput (in a sane world), though its resistance to noise would be helpful.

CAN bus was invented in the mid-80s. That was when ethernet was getting going and fibre-optics were also early in a long take-off run. In context, a hard disc was considered to be performing pretty well if it could put through 1 megabyte per second and ethernet managed something like 0.3 megabytes per second. Even at that time, fibre was good for hundreds of megabytes per second, and that over huge distances. Surely it was what we'd all want.

Cut to today: the solid state NVME drive in my cheap desktop puts through nearly 800 megabytes per second, my gigabit ethernet pushes a bit over 100 megabytes per second, and still nothing I do hits fibre until it reaches my ISP.

CAN has seen pretty minimal progress in comparison. Give it the same sort of improvement as hard discs or ethernet and it should be enough. That seems to be what Tesla has done (underneath the marketing facade).

It seems that there are a number of forces involved; it's not a relentless drive for more and faster. LIN bus was invented as a simpler, lower performance and cheaper car networking approach. Odds are there's not one best answer.

In other car news, Extreme-E gives way to Extreme-H. Not having to lug such a heavy battery around could be a win, somewhat offset by having to lug a heavy hydrogen tank around. At least refuelling should be quicker.

And on the US Supreme Court ruling on presidential immunity, this partisan lawyer's opinion gives a bit more depth than the news headlines. Some interesting points, particularly that it has identified areas where a president *is* answerable.

tomo,
Etherloop strikes me as merely a marketing term. The Wikipedia article on Etherloop is pretty vague, but shows that the name is from the '90s. There is a small amount about Tesla in there, but little to indicate anything more than a shared name.

The small amount of technical stuff I've turned up doesn't indicate any physical loop structure; looks like a two-wire terminated bus to me, just like CAN. Not that that's a drawback. A loop structure is nice for flexible configuration — each node only has to worry about its two neighbours — but there's a price in latency and fragility.

I think you put your finger on the real advance: Tesla's willingness to build their own controllers. If you can make a single controller that combines the functions of half a dozen off-the-shelf ones, network traffic is going to go down.

You also put your finger on the danger: the software is necessarily more complicated than in the previous modular approach. That doesn't worry me much for a combined wiper/washer/headlight/blinker controller (or whatever), but let's hope they don't glom in brakes and steering as well.

And for something completely different, I thought this video on speaking English with German grammar was good fun.

tomo,

Thanks for the Etherloop pointer. Interesting, but it doesn't seem all that much of a revolution. It has the feel of a recurring swing in computing between centralised and distributed processing. It is a good point about Tesla building their own controllers, thus being able to combine tasks that are done separately in other vehicles.

The 48V part is probably the bigger win, but also a mixed blessing. It's great for high power items: wiper motor, demister, headlights. But you'll need a 12V setup for off the shelf items like radio, and the digital electronic bits probably run at 3.3V, so it's trading copper conductor against power supply switching. 48V might also challenge insulation breakdown in wiring looms.

BH stylesheet seems to have gone haywire (at least in Firefox). Looking in the HTML inspector says that "common.css" isn't being loaded because it's tagged HTML rather than CSS. Strange.

tomo,
If he looks in each year (or month) when the Squarespace bill comes in, it's hard to imagine he can see much value in it. Can hope he got something out of looking in during COVID. Maybe he's expecting more interesting times ahead. I have occasionally thought of hitting the tip jar here, but I don't like Paypal (or the currency exchange charges), so soothe my conscience by contributing more to Jo Nova.

It's marvellous the breadth of expertise that climate scientists have. I listened (very briefly) to Joelle Gergis on ABC radio, telling us what a costly mistake it would be to pursue nuclear power instead of wind and solar. Of course economics and engineering are as nothing to one who commands the deep knowledge of climate.