The Future of Onboard Computers

Teardown of Audi A8, Nissan Leaf, and Tesla Model 3 onboard computers



  This report presents an analysis of electronic components performed by Fomalhaut Techno Solutions (Fomalhaut) in partnership with Marklines. Starting with teardown, analysis, and cost calculations of mobile communications devices, Fomalhaut has expanded its business into the automotive field. Fomalhaut specializes in product overviews based on internal and external inspections of electronic devices, function estimates of circuit boards and mounted parts, and creation of block diagrams.

Tesla Model 3

  From the teardown analyses of onboard computers that Fomalhaut has so far conducted, this report outlines the features that can be seen from the computer circuit board configurations of the Audi A8, Nissan Leaf and Tesla Model 3 and investigates future issues of onboard computers.



Report created by: Fomalhaut Techno Solutions



Related reports:
Automotive Semiconductors: High-performance Systems for ADAS and Autonomous Driving(Mar. 2020)
Nissan and Honda Autonomous Driving: Hands-Off from Highway to Downtown(Feb. 2020)
CES 2020 Show Report: Renewed Expectations on ADAS vs. AV(Jan. 2020)
2019 Year in Review: Reality, Trends and Highlights of AV Industry(Dec. 2019)
Analysis of Tesla's Over-the-Air Software Updates(Aug. 2019)


Differences of onboard computers: Failure assumption or failure prediction

  Regardless of the price of the vehicle, one component which is always installed in any vehicle is the onboard computer. The computers go by various names such as “main computer”, “drive computer”, or “body control unit”. It is understood to be the component which has the role of controlling vehicle running, stopping and turning.

  Because a failure could lead to an accident, traditionally these components had the common trait of having a simple and robust structure. Once installed, they are required to operate without breaking for ten years or even longer. From the idea that fewer parts will reduce failures, the number of parts is kept to a minimum.


Audi A8 onboard computer

  The total number of parts on the onboard computer of Audi’s luxury car A8 (previous model) is 628. A feature of the computer is that advanced driver assistance system (ADAS) control mechanisms are incorporated. The ECU (engine control unit) of the current model is equipped with Nvidia Corporation’s (Nvidia) GPU (graphics processing unit) chip.

Audi A8の車載コンピューター
Audi A8 onboard computer


Nissan Leaf onboard computer

  The onboard computer of the Nissan Leaf features redundant circuits. The number of parts is 702. No connectors are installed on the circuit board and the individual parts are mounted directly on a single board.

Nissan Leaf onboard computer


  Both the Audi A8 and the Nissan Leaf onboard computers are encased in a strong, probably die cast, housing. Because processors are installed, heat is generated. Each has an air-cooled structure with no cooling fan. The size of each of these computers is about 20cm in length and width and 5cm height. Mass is less than 1kg.


Tesla Model 3 main computer

  In contrast, the main computer of U.S. company Tesla Inc.’s (Tesla) Model 3 has as many functions as possible packed into the computer, with a high density packaging never seen before in automobiles.

Tesla Model 3のメインコンピューター
Tesla Model 3 main computer

  There are two circuit boards about 30cm x 15cm plus one about half that size installed on the vehicle. With 6,905 mounted parts, liquid-cooling, and a mass including housing of under 3kg, its structure is in another dimension. If there is a defect in a part: out! A failure in the liquid-cooling system: out! The special characteristic of the onboard computer is that it daringly cuts down on redundancy.

  How can it be so different among cars? It is thought that it is related to the method of safety assurance followed by traditional automakers and emerging automakers such as Tesla. Traditional automakers have widely adopted a method which, by building in redundant circuits like the Nissan Leaf, maintains functionality even with the failure of one of the systems by using another. On the other hand, Tesla adopts a concept of “failure prediction”.

  AI is equipped in the two huge GPUs installed on the main computer, and it is thought that it has a mechanism that not only moves the vehicle but also simulates indicators of failure. The computer is connected to a data center by LTE communications and it is designed so that defects in other vehicles can be shared immediately. The onboard AI processes a great amount of data collected at the data center, it comprehends signs of failure, and if something is found, it directs to the nearest dealer; it is designed to bring the car to a stop safely before a fault appears.

  Whether “failure assumption” or “failure prediction”, the hardware which is required is dramatically different, and even the cost will change drastically. This difference is apparent between the onboard computers adopted by Audi for the A8 and Nissan for the Leaf on the one hand and by Tesla for the Model 3 on the other.

  It will be interesting to see which concept will become mainstream: “do not fail, do not cause failure, if failure occurs then how to proceed” or “failure prediction”. Which do you think?


Report created by: Fomalhaut Techno Solutions



Teardown, VW, Audi A8, Nissan, LEAF, Tesla, Model 3, ECU, ADAS, EV

<Automobile Industry Portal MarkLines>