Tesla really deserves some kudos for pushing the 48V architecture and the Ethernet-based communication. It drastically simplifies the wiring mess that is in a typical car.
And it's not even close. Just watch the teardown of Cybertruck and compare its wiring to something like F150.
This has been a real head-scratcher for me. BMW and other automakers could have vastly improved their cars by switching to 48V a decade ago, but they still keep just plodding along with 12V.
It's a classic chicken and egg economic problem. BMW doesn't make the chips/electronics that support the 48V architecture - Bosch & Continental (with NXP/TI/Infineon/Renesas as their silicon suppliers) do and they're not going to support 48V unless ALL (or a significant majority) of the automakers will. So it's a game of chicken.
I designed some stuff along these lines 15 years ago. At that time, 12 volt stuff was not just available, it was available with great economies of scale and a huge range of options, off the shelf. You need an automotive-qualified relay? A light? A solenoid? A DC-DC converter module? A fan? You'd have 100 choices at 12v, 30 choices at 24v and 3 choices at 48v.
Are BMW cars less reliable, expensive, not ready for zonal assembly...?
Major car components like doors or front axles are assembled in parallel to miscellaneous parts on the main body, and all .join() at the final assembly. This had been the case for past 30-50 years, possibly more, in case this needs to be said.
They can be even more reliable. And they definitely _are_ expensive.
> Major car components like doors or front axles are assembled in parallel
And doors (and tailgates) are the biggest body component that is _sometimes_ assembled independently. Then workers manually route cables through the body.
Pre-routing cables inside panels that can then just be welded together can save a lot of labor.
> And doors (and tailgates) are the biggest body component that is _sometimes_ assembled independently.
Sometimes? What and when on Earth is this about? Pre-WWII?
They wash and paint and dry the whole body at once _for paint consistency_, then take off doors and trunk lids and bumpers and send them into separate assembly lines. Those major parts flow parallel "threads" in sync and converge near the end, where connectors are plugged in and those major parts are bolted back in and plastic trims are pushed in to tuck everything under. Cars were basically always done that way for a long time everywhere. I think even lots of hand made supercars are like that, only except tact times are magnitudes longer.
> Then workers manually route cables through the body.
> Pre-routing cables inside panels that can then just be welded together can save a lot of labor.
What do these even mean? Are you hallucinating workers crimping cables in-situ? They just clip on harnesses and plug in couplers in "the line". Never seen under a door trim?
It sounds like you're either extremely ill-informed, or worse yet, potentially, intentionally misinformed about car manufacturing that what you see is advanced manufacturing. I think you should... look more closely into what "legacy auto" have been doing forever.
> What do these even mean? Are you hallucinating workers crimping cables in-situ? They just clip on harnesses and plug in couplers in "the line". Never seen under a door trim?
Workers still need to pull the wiring bundles through the car body and clip them, after the body is welded together. The connectors are impractically bulky to put several of them along the cable routes.
Pre-assembled panels can have cable runs attached to them during the individual panel assembly.
They don't need to retool all the factories at once. They could have gradually introduced 48V systems in parallel with 12V, slowly phasing in new components as they replaced the old 12V.
Converting between voltages is not a free action, and running two systems is more complicated than one...
You really need some special component that is much better at 48 for it to be worth it, otherwise a delayed platform switch is better; one some competitors have moved and the suppliers exist.
On top of all that, almost all the wiring in the car can be made thinner, because of the greatly reduced losses. This saves a bit of weight, but also a lot of cost because copper is expensive.
There is no free lunch. You need to go to a finer wire strand, better insulation, better loom, better support for the harness etc, if you want that super fine wire to last. Some of those have pretty direct labor cost impacts too. That's gonna kill a lot of your cost savings, especially at lower production volumes where the design cost is harder to amortize. There's no free lunch.
It's very, very hard to get insulation that's not good for at least 100V and I suspect that just about any generic wire is good for more like 300V.
The only exception that comes to mind is wire that's specifically for "household low voltage" like 24V AC for thermostat, doorbell, sprinklers, landscape lighting. Also normal ethernet. But these are almost all what you'd call signalling wiring rather than power wiring.
Your average hook-up wire that you could buy at the auto parts store to make some repairs is almost certainly rated for 300V already. Mostly because of chafe resistance. Wikipedia says that the dielectric breakdown strength of PVC is 40 millions volts per meter https://en.wikipedia.org/wiki/Polyvinyl_chloride.
Divide both sides by 1 million and you get 40 volts per micron. OK so you need 1/3 of a micron to insulate enough for 12V and you need 1.25 microns for 48V. Now let's have a reasonable safety factor of say 10 or so and we're looking at 3 microns vs 12.5 microns. The only wire I can think of that might have insulation that thin is enamel coated magnet wire for the inside of motor windings. But even that is probably thicker.
Any kind of plastic insulation is going to be significantly thicker than this just to be able to be coated onto the bare copper wire and stick.
You're not wrong that the insulation needs to be thicker as the voltage goes higher. But you're unaware of just how ridiculously over-insulated everything already is due to other constraints of manufacture.
I'm sorry but the other comment is more correct. 48V standard was originally created for mild hybrid systems for ICEs during mid-2000s as a stopgap solution to full hybrid transition. Looks like the earliest mass-production 48V-class system was a 2001 Toyota that ran at 36V.
The integrated starter generator(ISG) is usually a pancake shaped motor that replaces clutch/torque converter in ICE car, nothing like the regular starter motor.
MHV was not even real hybrid, and is no longer relevant, so was 48V, at least for a while.
The special component was supposed to be the starter. With start stop systems essentially mandatory, the starter runs much much more often and therefore wiring savings on the starter are pretty useful…
How about comparing it to a teardown of a Rivian? Rivian uses 2-wire ethernet but 12V, so it seems like a more interesting comparison.
That said, it all seems like inside baseball to me. The BMW 850i pioneered the CAN bus, but that car was forgettable and although CAN bus took over the car industry that did not seem to create any durable advantages for BMW.
Ethernet seems like the inevitable replacement for CAN, in light of VW's investment in Rivian, and 48V vs. 12V for the low-voltage systems seems like a wash.
CAN is slow. At best it's around 1Mbit, but you get into electrical limitations. So you have to run multiple CAN buses in parallel and carefully manage bandwidth limitations.
My Chevy Volt had 4 different CAN buses and one additional LIN bus.
This can all be replaced with just two Ethernet buses: for safety-critical and non-critical uses. And the gigabit speed provides plenty of bandwidth for any reasonable sensor traffic, even including camera feeds.
The current architecture was justified in 90-s when LIN PHYs were an order of magnitude cheaper than even CAN PHYs. Now Gigabit Ethernet PHYs cost less than a dollar.
It's unlikely that the multiple CAN buses are being used to increase speed by, say multiplexing them. In general, vehicles use multiple CAN buses for enhanced security. For example: things like diagnostic ports are often on their own CAN buses so data can't be directly injected into onboard systems.
All but one CAN bus in my Volt were connected to the OBD port. The unconnected bus controlled the high-voltage battery contactors and some other critical stuff.
The "main" bus was saturated with data, more than 80% of bandwidth utilization at 512kbs. And it kinda had a mix of everything, from street names to be displayed on the dashboard to ECU messages. The other two buses had some random messages, with no rhyme or reason for the split ( https://vehicle-reverse-engineering.fandom.com/wiki/GM_Volt ).
FSD will benefit from high-resolution (4K or above) camera feeds (for things like reading signs and detecting small obstacles). You can do this in a 10Gbps network and have tons of headroom for every other function the car will perform.
Why would you not? Tesla is sending even the infortainment data stream through that bus. It's incredibly helpful having all data travel on a singular wire because you can tap in at one point and read it all out. Makes the entire system significantly easier to debug, understand and develop against.
It's a good thing we invented video compression and hardware codecs/encoders a long time ago.
What you'll actually be sending is a high bitrate mpeg stream, probably 54Mbps or thereabouts, you could probably fit 50x camera streams on a shared 10Gbps bus.
Ethernet can handle real time now even in bus configurations!
10BASE-T1S is a new standard geared for automotive. It uses physical layer collision avoidance instead of classic Ethernet exponential backoff. This provides deterministic maximum latency.
Though you can get max latency guarantees with switched Ethernet and the appropriate switch QoS and hardware.
This sounds like the real answer. Replacing an automotive standard with Ethernet is going to reduce friction onboarding junior webdevs with MacBooks, and enable a more stable higher turnover labor intensive organization.
You can already do this trivially with Linux vcan[1] so I don't buy this argument.
I think the bigger factor is that innovation in the CAN ecosystem has been lagging behind Ethernet for decades now. Only reason it's had such staying power is industry inertia.
The relative cost is probably a factor (which overlaps with inertia of course, but if the thing you already have implemented is also cheaper, you aren't going to hurry up and change).
CAN has desirable electrical properties (e.g. hardware-level prioritization) if you have life-critical devices and non-life-critical devices on the same network. But it's painful to deal with from a software point of view, compared to IP-based protocols, for anything that doesn't require the properties of CAN.
Not that I know of --- it's a figure of speech to compare it to an standardized method of electrical wiring in buildings that eventually got replaced with better standards despite similar concerns at the time.
And it's not even close. Just watch the teardown of Cybertruck and compare its wiring to something like F150.