As the automotive industry accelerates its transition to automotive Ethernet technology, comprehensive design validation is critical to ensure interoperability and reliable operation among multiple ECUs. The concept of automotive Ethernet was proposed by the OPEN Alliance SIG, also known as IEEE 802.3bw (formerly BroadR-Reach), an Ethernet physical layer standard designed for automotive networking applications, such as advanced safety features, comfort and infotainment Function. With Automotive Ethernet, multiple in-vehicle systems can access information simultaneously over a single unshielded single-pair cable. For automakers, the technology reduces networking costs and cable weight while increasing signal bandwidth.

To achieve higher signal bandwidth, automotive Ethernet uses a full-duplex communication link over twisted-pair cables, supporting simultaneous transceiver functions and PAM3 signaling. Using PAM3 for full-duplex communication can complicate viewing automotive Ethernet traffic and signal integrity testing.

The OPEN Alliance has developed automotive Ethernet test specifications for components, channels and interoperability. The test system incorporates an electronic control unit (ECU), connectors and untwisted pair cables. The test requires the system to work under the harsh environmental and noise conditions inside the vehicle. To do this, users must be able to characterize and view signal integrity and traffic at the system level in order to perform reliability testing. Examples of applications where customers require signal integrity testing at the system level include:

• TC8 Signal Quality Testing • ECU Component Characterization and Testing • Automotive Ethernet Cable, Connector, Cable Length and Routing Characterization and Testing • Electromagnetic Noise or Gaussian Noise Testing • High Current Injection Testing • Production Unit Testing • Automotive System Versus Automotive Ethernet impact on network performance

DC motor on/off engine on/off Automotive Ethernet system debugging

Tektronix recommends performing signal integrity testing during the design phase to identify potential problems before system integration.

Full-duplex communication and testing challenges

Full-duplex communication and PAM3 signaling add complexity to verifying ECUs under real-world conditions. Most serial standards work in simplex mode, where only one device communicates at a time, some communication standards use a separate link for transmit and receive, while in automotive Ethernet, master and slave devices can communicate over the same The links communicate simultaneously. (See Figure 1)

Therefore, the signal from the master and the signal from the slave are superimposed on each other. The master device knows what data it is sending, it can determine the slave device's signal from the superimposed signal and vice versa. Although transceivers are designed to handle this situation, isolating the signal on an oscilloscope for signal integrity testing or protocol decoding is nearly impossible.

In order to perform signal integrity analysis on a link, using an oscilloscope for protocol decoding in a real system environment, the automotive designer must look at each link separately, and the user must separate the signal before analyzing it.

It should be noted that it is best to perform signal integrity testing during the automotive integration phase, selecting cables, checking ECU performance under electromagnetic noise conditions, determining optimal cable lengths and routing, etc. For this type of analysis, the eye diagram test can be a very important tool to see the health of the system, which we will discuss later.

Separating Automotive Ethernet PAM3 Signals

Currently, there are two ways to separate the master signal from the slave signal. The first is the traditional method, which requires the user to disconnect or cut the automotive Ethernet cable and insert a directional coupler to separate and test the signal. This approach is inherently flawed in achieving accurate testing with minimal disturbance. The second method, the Tektronix Signal Separation Method, is a new method that uses advanced software and probes to separate signals non-intrusively, allowing the user to see the true signal more clearly. This approach overcomes the shortcomings of the traditional directional coupler approach. Below we will discuss and compare the two methods.
Directional Coupling Method As mentioned earlier, the directional coupler method requires disconnecting the automotive Ethernet cable and entering the directional coupler to separate the signals. Cutting cables at the system level is not an easy task, so this method is not suitable for system level testing.
With this method, the user can view the master and slave signals, but it introduces insertion and return loss, making it difficult to determine whether the error is caused by the system or by adding hardware. Also, while we may be able to remove the effects of directional couplers, de-embedding may amplify noise in the system, affecting measurement and characterization accuracy.

The setup we used included a clamp to convert automotive ethernet to SMA connectors, a directional coupler, and a clamp to convert SMA to automotive ethernet cable.
An eye diagram showing the effect of insertion and return loss on an automotive Ethernet signal with a directional coupler installed. The maximum amplitude is 100 mVpp because the directional coupler operates on the directional principle and the insertion and return loss results in closing the eye. Until recently, the directional coupler method was the default test method for automotive Ethernet because Tektronix' software-based signal separation test method was not available.
Tektronix Signal Separation Method Tektronix Signal Separation Method, introduced in July 2019 , separates full-duplex signals by viewing voltage and current waveforms from both the master and slave test points, and uses a proprietary software algorithm to provide the separated signal . The Tektronix Signal Separation Method is a software-based solution that allows the user to see the real signal without cutting the automotive Ethernet cable. One of the advantages of this method is that it can display the master and slave signals without the added insertion and return loss and de-embedding effects of the directional coupler method.

The eye diagram below uses Tektronix Signal Separation software. Compared to a directional coupler eye diagram, the signal quality is higher and the eye diagram is "clearer". Users can accurately represent automotive Ethernet signals, enable signal quality measurements, and be able to identify potential performance issues faster.

Comparison of Signal Separation Method and Directional Coupling Method

We use the two test methods mentioned above to conduct measurement tests and compare the test results.
For the test, we first set up and ran the test using Tektronix Signal Separation Technology, a current probe and a voltage probe. For the directional coupler method, we cut the automotive ethernet cable and plugged in a directional coupler with an SMA connector. Then we run the test under the same test conditions as the directional coupler method, and then invoke the signal separation method waveform to compare the two test methods.

The comparison shows that there is a clear difference in the magnitudes of the two methods, illustrating the effect of the directional coupler. With the directional coupler method, the amplitude of the master signal is about 90 mVpp (peak-to-peak voltage) and the amplitude of the slave signal is about 85 mVpp. In contrast, the signal separation method has an amplitude of about 1.5 Vpp for the master signal and about 1.45 Vpp for the slave signal. In this example, the directional coupler adds 20 dB of loss.

To eliminate the breakpoints introduced by directional couplers, de-embedding is necessary to compensate for insertion and return losses. As mentioned earlier, while it is possible to eliminate the effects of directional couplers, de-embedding can amplify noise in the system, affecting measurement and characterization accuracy. It should also be noted that de-embedding can be time-consuming and extremely challenging. Additionally, cutting cables and installing directional couplers can be extremely challenging for system-level testing and maintenance of an automobile.

In contrast, the signal separation method shows the real signal without disturbing the system. With this new approach to automotive Ethernet testing, users can characterize signals with greater accuracy and in less time without the added expense and measurement challenges. Users can use this method to perform signal integrity testing at the system level, performing all the tests available in the application environment.