Oscilloscopes for debugging automotive Ethernet networks

29th March 2019
Posted By : Anna Flockett
Oscilloscopes for debugging automotive Ethernet networks

Developers of electronic control units (ECUs) with an automotive Ethernet interface must perform tests to ensure that their devices work properly. However, if problems occur during signal transmission, plain Ethernet protocol analysis is not adequate in most cases. The new trigger and decode bundle for oscilloscopes from Rohde & Schwarz provides a good workaround for this problem.

The bundle allows developers to trigger on the transmitted Ethernet protocol content, decode the content and correlate its timing to the electrical bus signals. During debugging work, this helps to significantly accelerate the problem analysis.

Automotive Ethernet is becoming increasingly popular as a fast bus system for in-vehicle applications such as driver assistance and infotainment systems. The automotive industry developed the 100BASE-T1 Ethernet interface for this purpose. It is based on BroadR-Reach technology and has been standardised by the IEEE 802.3bw working group. 100BASE-T1 uses full duplex Ethernet communications via an unshielded Ethernet twisted pair. The 100BASE-T1 signals have PAM-3 modulation and the levels of the differential signal are between –1 V and +1 V. At 100 Mbit/s, the data rate is significantly faster than traditional bus systems such as the CAN bus.

The transmitter modifies the frequency response of the 100BASE-T1 signals in order to ensure reliable transmission with minimal RF leakage from the unshielded cable. The 100BASE-T1 standard requires the presence of an equaliser in the transmitter. When establishing the connection, the 100BASE-T1 PHY chips measure the cable's frequency response. The equalisers predistort the signals for subsequent data transmission in the aim of ensuring reliable signal transmission while simultaneously minimising RF leakage from the cable. Compared to the Ethernet 100BASE-Tx standard which works without equalisers, signals in the 100BASE-T1 system are highly distorted due to the predistortion. As a result, developers can no longer judge the signal quality solely by analysing the levels of the electrical bus signals.

Testing automotive Ethernet interfaces
The IEEE has specified the characteristics of 100BASE-T1 interfaces. Using the standardised compliance test, developers can measure the electrical characteristics of the interface using an oscilloscope and network analyser in a laboratory. An Ethernet protocol analysis tool such as Vector CANoe or Wireshark is normally used to verify that the electronic control unit (ECU) correctly handles communications. Tools of this kind record all of the Ethernet data traffic and provide comprehensive analysis capabilities. However, transmission errors appear only in the form of telegram errors and in-depth analysis of the root cause is not possible. An oscilloscope with a suitable trigger and decode bundle is generally required for this purpose.

Using the new trigger and decode bundle for the 100BASE-T1 bus from Rohde & Schwarz, ECU developers can now for the first time directly correlate the electrical signals with the transmitted telegram content as part of their analysis work. For example, problems that occur on the bus in automotive Ethernet applications can now be debugged just as easily as conventional CAN buses (for which powerful trigger and decode bundle options also exist).

Special features of the automotive Ethernet trigger and decode bundle
In 100BASE-T1 communications, both data streams are transmitted simultaneously via a twisted pair. If the user records the bus level with an oscilloscope, then the superimposed data streams for both bus users are measured. Without separating these data streams, there is no way to perform the required analysis. The R&S RT-ZF5 Ethernet probing fixture from Rohde & Schwarz is equipped with suitable directional couplers for this purpose. After insertion into the Ethernet line section, it separates the data streams to allow nonintrusive recording of 100BASE-T1 communications with an oscilloscope.

However, the recorded signals have been highly distorted by the equaliser used in the 100BASE-T1 transmitter. Prior to further processing, the signals are first equalised using complex algorithms and then decoded. The oscilloscope unscrambles the telegrams in the decoding process and displays all of the transmitted data telegrams and idle frames. The decoded telegrams are shown as color-coded bus signals and in tabular format. This allows developers to correlate the live 100BASE-T1 signals with the transmitted protocol content in order to perform highly detailed analysis.

The extensive triggering capabilities also allow developers, for example, to display isolated telegrams with specific source or destination addresses.

Analysing telegram errors
The timing relationship between bus communications and other signals can be revealed based on 100BASE-T1 decoding. For example, users can determine the start time of an ECU for debugging purposes by triggering the oscilloscope on the 12 V supply voltage and measuring the time elapsed until the first valid telegram arrives. The stability of bus communications can also be verified just as easily: The developer configures the triggering specifically for short-term interruptions of the supply voltage and then analyses the resulting interruptions in the bus communications. If many interruptions occur, then the stability is significantly impaired.

Sporadic bus errors due to coupling from interference sources can be difficult to debug without additional analysis capabilities. By decoding the 100BASE-T1 communications, developers can analyse bus communications across all of the protocol layers with proper timing correlation in order to identify coupling from interference sources.

For example, in the measurement in the MAC frame and idle frames are correctly transmitted at the start of the recording. However, the data stream is abruptly interrupted in the middle of the recording. In the lower signal, the frequency spectrum of the interfering signal (area marked in gray) is plotted. A peak is clearly visible at 2 MHz. Obviously, this interfering signal caused the bus interruption. The combination of decoding capabilities with other analysis tools provided by the oscilloscope (e.g. frequency analysis) makes this type of debugging much easier. For example, the oscilloscope allows interference, that would be very difficult to isolate using other methods, to be detected at a glance. 

Summary
For developers of electronic control units (ECUs) with automotive Ethernet interfaces, Rohde & Schwarz now has a complete 100BASE-T1 trigger and decode bundle option including a probing fixture for nonintrusive signal access. During debugging work, developers are supported by the comprehensive triggering and display functions for transmitted telegrams. The displayed decoding information is correlated in time with the electrical signal. This allows users to analyse the protocol content during debugging and quickly identify the causes of any bus errors that occur.

Besides the 100BASE-T1 trigger and decode bundle option described here, Rohde & Schwarz has complete test solutions for 100BASE-T1 and 1000BASE-T1 automotive Ethernet compliance tests and link segment tests using an oscilloscope and network analyser.


You must be logged in to comment

Write a comment

No comments




Sign up to view our publications

Sign up

Sign up to view our downloads

Sign up

Agile for Automotive 2019
15th May 2019
United States of America Detroit, MI
ELIV 2019
16th October 2019
Germany Bonn World Conference Center