Optical data transmission: What’s already quite common in data centers will revolutionize data transmission in the automotive sector
For years, Ethernet technology has been increasingly making its way into the automobile world, a phenomenon that is causing great changes in on-board network architecture. The reasons for this shift are the constantly expanding data quantities due to ADAS, automated and autonomous driving, and the increasing number of sensors, cameras, and displays in vehicles. The transformation in the way vehicles are used in the course of autonomous driving also requires large quantities of data. It may be possible to use the vehicle of the future as an office or living room, for example. Another reason for using Ethernet technology is to reduce the complexity of cable harnesses and make it possible to manufacture these automatically.
There has long been a trend toward replacing classic Ethernet systems with optical data transmission in the world’s large data centers. This trend is already quite advanced in the telecommunications sector, which leads us to the questions we want to answer in this blog post: Why are classic Ethernet systems being replaced by optical data transmission, what are the differences between these systems, and what are their advantages and disadvantages, not to mention the limits of both. Here’s what we found out.
Optic or copper – transmission technologies and their physical properties
If you examine the structure of the transmission channel, there are essential differences to how each transmission technology works. While with electrical data transmission the electrons handle the transmission of the various voltage values, with optical data transmission the photons do this, sending a kind of “Morse code” where the light source switches on and off at very high frequency.
If you consider transmission speeds, at 230,000 km/s, electromagnetic data transmission can transmit data faster than optical transmission, since the light speed in fiber optic cables is inferior, at just 200,000 km/s. But this isn’t the whole picture, since with electromagnetic data transmission, physical effects play a large role.
A significant factor is the damping, which increases disproportionately to transmission frequency during electrical data transmission. An additional factor is the length of the transmission link, which also amplifies this effect. Therefore, it is necessary to restrict the line length for data rates beyond 10 Gbit/s for electrical transmission (e.g., 10 Gbit/s to 15m or 25 Gbit/s to 11m). The compression possibilities of the so-called pulse amplitude modulation (PAM 4 – PAM16) are used for transmission of higher data rates in order to send the necessary data rates across the connection lines.
Furthermore, it is necessary to restrict the number of so-called inline connections. Currently, the chip industry is trying to find solutions to improve signal integrity to prevent multi-laning (several parallel lines) at data rates > 25 Gbit/s. Here, approaches from digital signal processing (DSP) such as equalizers, common mode choke, DC block, EMI filters, and ESD-proof lanes are used to enable the signal integrity via measures on the control unit. With higher data rates, these measures also affect the required board space, EMC susceptibility, and thus the costs.
If you examine optical data transmission, it quickly becomes clear that its physical principles open up a broad field that is already being used in data centers. Thanks to the low damping of fiber optic lines, for example, transmission of 25 Gbit/s across at least 40m with NRZ (non-Return to Zero) is possible. This means that transmission can occur without compression, which reduces the complexity of the transmission on the control unit, for DSP and PAM modulation are not required at this data rate. Another advantage is that less space is required on the control unit since such measures can be omitted. Furthermore, optical data transmission is ahead when it comes to electromagnetic compatibility (EMC) since neither plastic nor glass are subject to the effects of electrical and magnetic fields. Therefore, this technology is very suitable for use in sensitive areas in the vehicle such as battery management systems, for galvanic separation, or HF antennas.
Optical data transmission as a technology of the future in automobiles?
With all the pros and cons, it’s necessary to ask when it will make sense to consider a paradigm shift in automotive data transmission.
According to an MD internal survey of experts, opinions about this vary considerably since only a few companies on the market are focusing on this technology. In the survey, the technical effort required, costs, and customer acceptance were compared with the levels of data rates from the IEEE 10 Gbit/s, 25 Gbit/s and 50 Gbit/s.
The results with regard to technical effort required were determined using the necessary measures for EMC, limit values (e.g., insertion loss, crossview), chip complexity, the number of inline plug connectors, design effort for minimizing reflections, and the achievement of the required transmission frequency. The achievable link length was also incorporated into the assessment.
To evaluate the costs, an estimation of the costs of a transmission link were examined. This includes the chip-to-chip connection and all required measures on the board that enable secure data transmission per data rate under the required conditions.
Another important point of the evaluation was customer acceptance, which was quantified with the acceptance of the transmission technology and with the applicability at OEMs and Tier1s. Here, the estimation was of the effort required for approval inspections, approval costs, and effort required for installation in vehicles.
From these questions, average values were determined and used for the individual charts. There is a relatively precise statement about the point at which the effort required for electrical data transmission exceeds that for optical data transmission – at the latest when that is the case, the automobile industry will start moving in the direction of optical data transmission.
Technical effort required to achieve the data rates
The reason for this is that in the range between 10 and 25 Gbit/s, the costs for achieving signal quality increase exponentially. They might even exceed the costs of implementing a new technology in the medium term. The breaking point will be beyond 25 Gbit/s, since starting here, 2 cables (e.g., H-MTD cables) in parallel are required for a functioning copper link.
Costs of the data rates
An important issue for the introduction of optical data transmission is customer acceptance. If you examine the survey (see Fig. 3), it is clear that optical technology requires more promotion in order to prepare the technology for serial use in the automotive sector. Assuming implementation at 25 Gbit/s, time is growing short for creating a controlled serial start since the approval of a new technology takes somewhat more time than the approval of known technologies.
Acceptance of the technologies for the respective data rates
Early acceptance can only be achieved if the many advantages, such as mechanical robustness, EMC independence, simple galvanic separation, installation space savings and weight are emphasized.
Copper – a resource in short supply
Another significant issue for the paradigm shift is the availability of copper, a resource that is in shorter supply now more than ever due to increasing demand. Demand for copper is increasing in part due to the greater requirements of on-board networks, which may mean that it is time to consider alternative solutions. Here, optical data transmission would be the optimal solution since there are no copper components and this topic would then no longer affect on-board networks.
When will optical data transmission make a breakthrough in automobiles?
The final question is when optical data transmission might be deployed in automobiles. This topic was also covered in the survey and the result was somewhat surprising since opinions here varied considerably. From the analysis of the earliest deployment points, it is possible to see in Fig. 4 that the higher the data rate, the more the people surveyed relied on the optical solution. It can be assumed that the deployment of optical multi-Gigabit transmission will become part of serial applications in 2026 and this technology will then conquer the market step by step.
When will optical transmission achieve a breakthrough in on-board automobile networks?
Taking all of these points into account, the conclusion is that it won’t be long until optical transmission is deployed as a new technology since the need to solve all these problems is increasingly urgent. The standardization of IEEE802.3cz will also be completed at the beginning of 2023. At this point the physical layer of the Open Alliance will start, so that presumably in 2024, the standardization of the optical multi-Gigabit transmission will be fixed.
This is why MD ELEKTRONIK is already focusing intensively on optical data transmission in vehicles and working on solutions for tomorrow’s future technology.