第一个车载以太网应用-BMW 环视项目背景及历史 BMW Automotive Ethernet application background （1）

本文主要介绍，BMW当初在汽车上应用以太网技术的背景及历史；

此篇主要介绍BroadR-Reach/100BASE-T1技术的简要历史及背景；

The Breakthrough: UTSP Ethernet for Automotive

BMW decided to synchronize all hitherto knowledge with the future requirements on IPbased communication systems. A key learning was that a PHY usable with unshielded cabling was decisive for the future of Ethernet in automotive. Another was that the EMC properties, at least the immunity, had been surprisingly good the first time round. In consequence, BMW decided to look more closely at the possibilities to reduce the emissions of 100BASE-TX Ethernet when using unshielded cabling.

Together with Lear, who had supplied the central gateway, BMW performed measurements. Starting point was the existing gateway. The EMC performance of the ECU itself, when not doing Ethernet transmission, was very good. The hope was thus to be able to isolate the source(s) for the strong emissions. And indeed, the output driver stage was identified as the root cause. The gateway hardware allowed the output driver of the Ethernet PHY to be deactivated, while leaving the internal MII and all other interfaces live. In this case, the unit was well under the emission limit lines. Unfortunately, irrespective of the means taken – filters, ferrite beads, etc. – it was not possible to get below the emission limit lines when the output drive stage was switched on.

Thus, in summer 2007, BMW approached four well-known vendors of Ethernet PHYs and asked for their opinions and solutions. Colleagues at the automotive semiconductor supplier Freescale (now NXP) had suggested that this might be worthwhile. Of the companies addressed, only Broadcom responded positively, and in September 2007 the first meeting was
held in Munich. During this meeting the results of the gateway measurements were discussed and the automotive requirements were aligned with the performance value of a solution Broadcom had originally developed for Ethernet in the First Mile (EFM). In January 2008, the Broadcom technology called BroadR-Reach went into the EMC labs at BMW. below Figure shows the results of the first emission measurements performed.

Of the other three companies, one did not reply and the other two thought the BMW request impossible. Some years later, after BMW and Broadcom had proved that transmitting Ethernet packets at 100 Mbps over unshielded cabling was possible in the automotive environment and were promoting BroadR-Reach in order to attract other customers and suppliers, every one of the three other companies originally asked, developed other, incompatible solutions. One solution was even based on 100BASE-TX ，something BMW would have highly appreciated a few years earlier.

On one hand these solutions created confidence in the industry that transmitting 100Mbps Ethernet packet over unshielded cabling in the automotive environment is really feasible. One the other hand, for those not having yet decided on the use of BroadR-Reach, it caused additional validation and decision effort and some uncertainty for all. In a fragmented market, no one wants to have decided for the technology with the smaller and potentially decreasing market share.

In hindsight we know that BroadR-Reach/100BASE-T1 succeeded, but at the time the situation was not always that clear. All car manufacturers have a long lead time to introduce new in-vehicle networking technologies. The decision has to be taken at least three years ahead of SOP, meaning that another year before investigations on the technology have to have started. For BMW all other proposals were simply too late to consider. This meant that for BMW the other solutions were mainly a source of discomfort as they posed an economical risk.

With BroadR-Reach, the door opener to Automotive Ethernet was found. BroadR-Reach promised to transmit Ethernet packets at 100 Mbps at vehicle runtime over a single pair (100BASE-TX requires two) of unshielded cabling (Unshielded Twisted Single Pair, UTSP), i.e., the same cabling the industry used for CAN or FlexRay networks. This would be the most cost-efficient high-speed network in automotive, providing a higher data rate than MOST at a lower price level than MOST, any pixel link, or  consumer technology. Nevertheless, this was still only the beginning. In 2008 all that existed was a good technical prototype some engineers at BMW had had the chance to investigate. The technical and economic feasibility had yet to be proven over all levels of decision making within BMW. Also, the automotive industry as such had yet to be convinced that Automotive Ethernet was the right way forward. After all, the use of an in-vehicle networking technology has limited advantages to a car manufacturer when the car manufacturer is the only one using it.

 

本文节选自《Automotive Ethernet》

作者Thomas Konigseder 2017年从BMW离职，加入Technica Engineering担任CTO；