Creating an electric car: how to build an EV from R&D to production

Working for a company that designs and manufactures electric vehicles, their electronic components and software, I took a managing role in the Program delivery process. There, I encountered a problem that might be not so obvious at the first glance: being focused on functional design and user experience, engineers often completely lose sight of the project’s next stage: production, especially mass production.

Their main misconception is that a well-designed functionality predetermines the success of a device or a software system project. In real life, things are much more complicated. In this article, I will take a detailed look at the aspects that need to be considered throughout the entire life cycle of a project (from R&D to production), as well as ways to solve problems that arise in the process.

Car industry: the Gentle Giant

If the auto manufacturing industry were a country, it would be the sixth largest economy, the International Organization of Motor Vehicle Manufacturers (OICA) proclaimed in 2005. In that year, the industry produced over 66 million vehicles worldwide. Well, in 2022, it produced over 85 million vehicles. If Earth were flat, the car industry would have been one of the giant whales it rested upon.

Every $1 spent in vehicle manufacturing creates additional $3.45 in economic value. Automobile industry boasts #2 biggest job multiplier (i.e. how many jobs in other sectors are directly and indirectly supported by 100 jobs in a given sector) across the US economy: 464.08, surpassed only by utilities.  It accounts for almost 5% of GDP in the US and 7% of GDP in the EU.

How come the car industry has become one of the engines of the world economy? This is because it is so innovative. Of course, big tech like Google and Meta are now synonymous with innovation, but the automotive industry keeps up. In Europe, the automotive sector is the number one investor in R&D, responsible for 31% of total spending, or 59.1 billion euros in 2022, surpassing tech, aerospace and defence industries, according to the European Automobile Manufacturers’ Association (ACEA).

Worldwide, Volkswagen is #10 top R&D investor, right behind the big tech companies. However, unlike Amazon or Alphabet, car industry products are made of solid, palpable matter: metals, plastics and rubber. Thus the real-world transition from R&D to mass production is vital for the sector. But this process is complicated and very vulnerable…

A research by McKinsey shows that an average car is made up of around 20,000 parts (some sources bring this estimate to even 30,000 parts, 20 to 90 Electronic Control Units and millions of lines of code). As such, a global car manufacturer may have more than 18,000 suppliers involved in its production process, even more than an aeroplane manufacturer. Being very down-to-earth and material, car production is very vulnerable to real-life events and shocks.

Take semiconductors: a thrilling research made by Rabobank dives into the supply chain difficulties and dependencies. Different manufacturing stages take parts in different regions of the world: raw material extraction, purification, production of wafers, microchips and devices. In fact, semiconductors have to make a circumnavigation journey (or even more than one) on their way from raw materials to consumer devices. A strong gust of wind in the Suez channel can disrupt this entire chain and severely affect the auto industry.

The Achilles’s heels

This vulnerability is not fully understood and often underestimated by the staff responsible for the product design. For instance, my employer had a framework that engineers had to follow while building system designs and producing prototypes. Per this framework, management, and the engineers themselves, believed that the work they had done up to the end of design validation was 90% of the entire project. Of course, as the projects progressed to production, several crucial factors turned out affecting the entire process:

• Supplier choice. Most suppliers have already developed their technologies and toolings used for component production. This means that not every design can be adapted by every supplier. Often, a design has to be significantly altered in different ways to fit different manufacturers. For some designs, tooling can be prohibitively expensive due to the use of rare materials in the vehicle or the need to resort to emergency production safety measures. Sometimes the time to create tooling is so long that the production of a device becomes meaningless since the whole process does not fall into the timelines for entering the market. But even if your design matches the technological capabilities of a supplier, it may well be that the production plant can be loaded with work for several years in advance. This is especially true for those that make electronic components and particularly affecting electric vehicles. For instance, a typical combustion engine car has 2,000-3,000 multilayer ceramic capacitors (MLCC). But a single eclectic vehicle has as many as 22,000 MLCCs and the number is steadily growing. Given such an explosive growth in demand, it might be hard or even impossible to fit into a manufacturer’s schedule for many years.
• Lead time and availability of components and materials. For instance, using rare metals in the device can significantly increase the time required to start production due to lack of material. Similarly, production of devices can stumble upon shortage of components. A single fire at a Japanese chip manufacturer in March 2021 halted the production of nearly a third of the world’s microcontroller chips used in cars. Chip shortages related to COVID shutdowns in 2020 affected more than 169 industries, of which the automotive and consumer electronics industries took the largest hits. The war in Ukraine has made things even worse. In 2022, lead time for some raw materials reached a record high of 99 days. Alas, lead time increases as we go up the ‘food chain’ from raw stuff to components: last year lead time for analog chips, microcontrollers, FPGA, discretes and some passive components exceeded 40 weeks. For some semiconductor components it can even reach several years. Shortages are not limited to electronic components though. In the US, general repair shops' schedule backlog hit almost 5 weeks in 2022 instead of the usual under two weeks observed over the previous five years. Lead times have been excessive even for body parts, the simplest things a modern car features. This is an additional factor that breaks the balance between supply and demand. So now, more than ever, a supplier needs a very good reason to fit your order into the schedule. And even a better reason to readjust the manufacturing techniques or upgrade tooling.
• Cost and possibility of mass production. Some designs are not suitable for mass production or have a huge production cost: for example, due to the inability to provide the guaranteed iterability and quality of the device at the testing stage at the end of the line. If each device requires manual labour to validate, it is extremely expensive or simply impossible to produce hundreds of thousands of such devices.
• Defect rate, an indicator used by manufacturers to measure product quality. Producing some designs is possible only with a large number of defective items because it is impossible to establish an automated production with a guaranteed result for certain components. In this case, the cost of all defective devices must be added to the cost of the functional ones. In the 20th century, a defect rate of less than 1% was considered a good result, then the expectations increased to 0.1%. Twenty years ago, the average defect rate in the automotive industry was 0.0025%. Now, manufacturers are struggling to achieve a zero-defect strategy. You have to make sure your design can fit into these strict policies.
• Parts labelling. This aspect is often overlooked, but labelling is an integral part of the design. Lack of proper labelling to ensure the traceability of the device and all its internal components throughout the entire life cycle can lead to the complete impossibility of using the device in the automotive industry. “For automotive manufacturers, only a faulty part is worse than a labelling error”, says Marko Vrbnjak, product marketing manager at NiceLabel as quoted by the Labels & Labeling magazine. “A single digit misprinted is terrible and can have large financial implications for the manufacturer as well as the supplier”. In the development phase, this factor is often underestimated.
• Serviceability of devices. Similar to the previous point, the cost of repairing or replacing a device is often not taken into account. Ultimately, in this scenario, either over-complication (or overengineering) occurs (and the engineer tries to make the device last forever), or the cost of repair and replacement can be very high, which will defeat the commercial sense of the device.
• Design compliance. The car and its components should comply with cybersecurity, legal and safety requirements of the distribution region. But it is important to understand that the customers themselves are not always able to specify the safety requirements at the beginning of the vehicle design. The formation of such requirements and their implementation is often postponed to a later stage of the project. This means that a significant part of software has to be completely redeveloped to fit the project into the legal framework. In some cases, it also requires a revision of devices’ architecture or even the architecture of the entire vehicle. This is especially relevant for vehicles based on the 4th and 5th architecture generations. In addition, the need for maintenance and software updates for the already sold vehicles is not taken into account. These updates are critical not only in terms of user experience, but also in terms of security. Considering the need to update the software on hundreds of thousands of vehicles in operation, this aspect can be considered one of the most critical for the project success. However, it is also the most underestimated aspect at the start of development in terms of cost and importance to the success of the entire company.

Heal the pain, but do it fast

As a result of these misconceptions and omissions, time and costs required to create a device may double and sometimes even triple, leading to critical consequences for the project, the program and the company as a whole. In extreme cases, entire programs can get disrupted and entire companies can go bankrupt. How can managers avoid such flaws in the development and production process?

• Consider the hardware and software design process from the collection of needs to the end of a device or system's life cycle (in other words, to disposal or decommissioning). Skipping stages is possible, but must occur reasonably and consistently, with an assessment of the risks and consequences.
• Always look at the process of creating software and hardware not only from an engineering point of view, but also from a financial point of view. After all, the creation of systems and software is aimed, as a rule, at making a profit for a business.
• If someone in the team says something is impossible, always listen to that point of view and try to understand the logic behind it. If you disagree, then try to convince the employee of the opposite, but do not complain about the “lack of ambition”. Hearing a negative opinion at the right time can save you hundreds of millions of dollars, even if you don't like the opinion, or if it goes against the company's goals.
• Early involvement of Functional Security and Cybersecurity engineers along with after-sales service specialists in the creation process is crucial..
• Functional and Cyber Security must be considered in relation to the entire vehicle with on-board devices: considering the components separately leads either to overengineering and a significant increase in the cost of the device, or to the appearance of gaps that are not covered either at the vehicle system level or at the device level.
• Finally, it is vital to consider the design of the vehicle in relation to the design of devices — in other words, the formation of synergy between hardware and software.

And remember, we all live in the world of ever tightening deadlines. The automotive sector, as one of the innovation leaders, is no exception. In recent years, the average time of a product development (from concept design to start of series) was almost halved from an average of 48 months to about 25. It is high time to take all the mentioned issues very seriously, as our room for error is shrinking like La Peau de chagrin. Sometimes it may do us good to remember the words of Seth Mattison, co-founder of Luminate Labs and co-author of The War at Work bestseller (as quoted by The Edge Magazine): “Innovation doesn’t have to be launching a big new, expensive, or risky strategy, product, or service. It’s looking for new ways to improve small, internal processes. Where could we do this a little differently? Where could this be better?”