Articles Architecture
The physics of the automobile
A hundred years ago the car was the cutting edge of applied physics. Its electrical architecture has not kept pace, and that is where the next step is.
First published 2019, revised 2026
Today the transport industry has a well-earned reputation for conservatism. Yet only a hundred years ago the car industry was on the cutting edge of technical progress: a melting pot in which thousands of inventions from the most varied fields were tried out, from new thread standards to scientific management. The people who set the principles of car-making understood that the industry had to keep innovating.
The car won its place on physical efficiency. In a railway carriage, up to 1.5 tonnes of carriage weight moves with each passenger; in a car that figure is 300 to 350 kg, roughly five times less. It is no accident that the car became the model of a compact, personal technology.
Two peaks, and a third
At the start of the twentieth century, new knowledge from physics and chemistry was turned into vehicle systems in record time. Looking at how transport innovation developed, two large peaks stand out. The first is tied to the rise of electrical engineering, from the late nineteenth century to the 1960s. The second began in the 1970s with microelectronics. A third is next in line, tied to digital technologies, smart onboard grids and new methods of energy management.
The self-regulating onboard network
As late as the 1950s and 1960s the onboard network served a minimal set of loads: lighting, wipers, the starter, the radio. Its architecture was a few power-distribution boxes holding switching relays and circuit fuses. Assemblies were connected on a functional basis: one power circuit feeds one assembly, or one functionally related system.
That philosophy is still with us. So is the practice of over-specifying every assembly: this is why 14-volt network parts are tested against voltage surges up to 150 volts and other anomalies of a self-regulating network. Despite strict standardisation, the range of adverse processes in the onboard network has only grown worse over time.
Electronics reached the car with less capability than it had
Electronic communication channels arrived in the car in the 1980s as descendants of industrial channels, and that lineage reduced their technical capability. To monitor a production line it is enough to watch an event: a part moving from zone A into zone B. A modern car needs thousands of parameters managed in real time, not the primitive watching of events.
Meanwhile the power system faces a move from 12 volts to 48 volts. The previous move, from 6 to 12 volts, took more than fifty years. Since the 1960s the number of electric drives in a car has grown by roughly a factor of twenty; a large luxury car can carry more than a hundred. Network self-regulation, box architecture and the endless growth of the wiring harness have reached the limit of what can be improved.
The smart onboard grid
The intelligent grid that will power hundreds of assemblies in a vehicle is unlikely to resemble today’s multi-kilometre harness. It has to handle energy flows optimally, on a power circuit that is simple in topology.
With this approach the total length and weight of the harness can be cut by 70 to 75 per cent against a conventional one, while the capability of the systems grows, thanks to real-time control, a small time metric for a full digital picture of the whole network, about one millisecond, and the availability of that picture at any point in the network.
| Conventional onboard network | Smart onboard grid | |
|---|---|---|
| Onboard network length | Up to 2.8 km of wire | Up to 500 m |
| Cables in the car | 1,000 to 2,000 | Two wires, or two buses |
| Contacts in the network | 1,800 to 2,500 | A fraction of that |
| Power distribution | Central boxes of relays and fuses | In the control block, next to the load |
| State of the network | Known only at points, at moments | A full picture every 1 ms, readable anywhere |
From the source essay, Table 2. Figures are vehicle-dependent.
Energy is only an intermediate state
Electrical energy is an intermediate state: the useful effect is mechanical torque on the shafts of the drives. So full control of the energy flows is, in effect, precise control of the entire mechanics of the car. A large amount of data about the state of the car’s assemblies can be read without sensors, from the state of the power network itself.
This technology was developed specifically for transport onboard networks, not adapted from another market. The same direction is visible in aviation, with the all-electric aircraft concept, and in marine and rail transport. The history of the car is full of dedicated assemblies invented for it; the smart onboard grid is the next one.
The architecture this points to is set out in the technology section, and a worked integration of it on a real vehicle is on the compact-car page.