Andy Norrie
The crucial role of Automotive Testing
Quarch attended the Detroit Automotive Testing Expo under the banner of the Symbiosis Alliance. If you missed the show, then you can catch up with our experience and a summary of how we can help your testing.
On the Quarch side of the stand, we had:
-
-
-
Power analysis for EV charging with our 3-phase AC PAM
-
AC motor analysis is also possible with the 3-phase AC PAM
-
DC power analysis for anything that has a suitable shunt resistor (From small embedded shunts in R&D prototype files boards to high-power DC supplies) with our multi-channel PAMs
-
USB testing (hot-plug, physical layer interruption and switching) with our popular range of automated test devices
-
Automotive fault injection with our multi-protocol breaker, injecting physical layer faults into a wide range of communication interfaces from I2C, CAN and LIN, up to 1000Base-T1
-
-
On the Ellisys side we had:
-
-
-
Combined Bluetooth and WiFi protocol analysis with their industry leading tools
-
On the wired side, they had their Type-C Tracker product for USB protocol analysis
-
-
1 – Automotive testing of power
I got to see a wide range of power testing at the show. There were multiple companies showing their power analysis options, including Yokogawa, Keysight, N4L and more
-
-
-
AC motor analysis Recording a 3-phase trace to look at power consumption, phase, noise, and changes over time. This allows the tester to see how the motor is operating and if any of the phases are different from each other. The Quarch 3-phase AC PAM is very capable of this test. While sampling slower than some competing devices, it is cheaper and much easier to set up.
-
DC motor analysis This requires higher sampling (2 MHz or more), which is not something we can currently cover. This is used to analyse the power from the invertor going into the EV motor
-
Battery charging Analysis of the charge cycle of an EV can also be covered with the 3-phase AC PAM. DC charging is not something we have available to day though. There was little discussion of this in the show, probably due to the separate ‘battery’ specific show that took place a few weeks earlier
-
Battery to invertor. This should be a good application for the External Shunt PAM. The DC supply here does not require the high sample rate needed for the input to the DC motor, and there is less test equipment available to cover this point. Testing options here seem to be more limited, so this will be a point for us to fill
-
-
2 – Automotive testing of USB
The main use here is phone connection / media playback, although some vehicles have software updates and additional USB-connected peripherals.
Quarch has the USB switch. This allows 8 devices and one host to be connected. Each device can be connected in turn to the host. The switch can also separately control USB2, USB3, and power. This allows test sequences where multiple devices are individually presented to the car in sequence, saving a lot of manual testing.
Next, we have hot-plug. The switch can do basic hot-plug, but our ‘Breaker’ product can do a lot more. From a simple plug, we can go to connecting USB2 and USB3 at precise times, simulating pin-bounce during connection, and much more. This allows full corner-case testing of the plug-pull sequence and ensures the host and device negotiate correctly in every case. This hot-swap testing was originally developed for the data storage industry but is very applicable to automotive. Here’s a white paper that explains the problem around hot-plug and how to test for them.
3 – Physical Layer Fault Injection
A common form of automotive fault injection is to generate electromagnetic noise, which has the potential to flip bits in the data stream and add faults. This makes sense, as EMC noise immunity is important for vehicles. There is a second way, though: Quarch Breaker modules inject faults by briefly (down to 50 nS) disconnecting one or more lines in the bus.
This creates a very predictable length of disruption and on the exact line you choose (targeting a single data lane for example). It can also be used to create brief power disruptions or target specific sidebands, such as the USB Type-C CC pins.
In vehicles, a common risk is vibration. As a vehicle ages, loose wires and poor connectors can slowly degrade to the point that the connection is imperfect. Quarch breakers can precisely simulate the various states of failure on a communications bus. You can ensure that every possibility, from occasional errors to heavy link corruption and total failure, is fully tested.
Failure to test this on USB could lead to a poor user experience if a bad link causes the media system to crash. Similar errors on a critical bus related to ADAS can be a far bigger problem and could cause safety risks. Here is an example of fault injection on an ethernet link, showing a major drop in bandwidth without the link actually failing
Show summary
This was a great event for us to go to. We learned a lot about the requirements for automotive testing and met many other interesting vendors who also had great solutions.
The actual visitor count was rather disappointing; it would have been nice to see a larger number of automotive manufacturers present, though we did have good conversations with Ford and a few others.
Our power products were clearly interesting to the attendees, but few of them had much interest in the protocol side.
