CAN Newsletter September 2005

Focus on semiconductors

CAN transceivers in system basis chips

Over the last decade, the automotive industry trend has been to move to highly up-integrated, smart power integrated circuits (IC), also known as system basis chips. Systems previously containing multiple ICs, each providing a specific function, now contain one or two system-on-a-chip ICs containing a wide range of analog, power, and digital functions. Like many ICs today, they have a digital core comprised from a standard cell library similar to a digital application specific IC (ASIC). They also include analog building blocks such as operation amplifiers, comparators, data converters, and voltage/current references. Somewhat more unique is that power circuits are also incorporated into the IC to add the ability to control motors, switches, and solenoids, and to generate power supplies, both switching and linear, for internal and external circuitry. On a single piece of silicon, one can find low power circuitry biased only with few microamperes of current that may be next to a power device switching several amps of current. To build an up-integrated CAN transceiver on a highly integrated IC is a challenge as the semiconductor technology that is chosen is not based solely for the transceiver. Higher levels of integration force semiconductor process development engineers to choose fabrication steps and create components that offer higher levels of digital integration with high precision analog devices and high voltage power devices. The addition of power devices also benefits CAN transceiver design as higher voltage components allow the common-mode range specification to be met. However, the process must also be a low-cost solution. To achieve this goal, a junction isolated process is used instead of the more expensive dielectric isolated process. Junction isolation requires fewer process steps, and is hence less expensive to manufacture. However, dielectric isolation benefits from the lack of active parasitic components, which complicate circuit design and often cause unsuspected operation. These parasitic components often become active if the node is allowed to traverse above or below the supplies, thereby forward biasing PN junctions. Furthermore, it is important to realize that these PN junctions can be the base emitter junction of NPN or PNP transistors.

Regardless of these process technology challenges, the automotive environment requires a robust transceiver design. While needing to meet the ISO 11898 standard, additional requirements are needed in the areas of electrostatic discharge protection (ESD) and electromagnetic immunity (EMI). While the standard recommends protection to short circuit faults from battery and ground, as well as shorted load conditions, the use of chokes to improve radiated emissions can cause higher than expected transient voltages along the CAN bus which the transceiver must tolerate without failure.

Would you like to read more? Please subscribe to the CAN Newsletter. It's free!

To top

Speed up of CANopen using co-processor support

Usually CANopen is implemented in a standard micro-controller with a CAN hardware module. The CPU (central processing unit) of the micro-controller executes the application software as well as the CANopen driver. Depending on the traffic in the CAN network, the CPU is more or less busy with the communication task. For systems with heavy bus load it can happen that the CPU is not available for the application and cannot react to service requests properly. Also the time for the application is not deterministic because communication can happen at any time.

One solution to solve the dilemma is to use a high performance micro-controller (e. g. a 32-bit micro-controller). But if the application does not need 32-bit data width, a system design with 32-bit controller is usually more expensive and requires more development and circuit effort than 16-bit controllers. Also, increasing the clock frequency in order to enhance the performance can create additional issues: higher power consumption, EMC issues, dearer components (memory, etc.), and limitations in the speed of memory access. Freescale Semiconductor (www. freescale.com) has introduced a solution, which solves this dilemma in another way. The company developed a dual processor micro-controller, which is based on the HCS12 16-bit micro-controller.

Concept of co-processor design

The concept of adding a co-processor unit to an existing CPU is not new. Especially in computer designs and in signal processing systems, parallel processing is used very widely to increase the computational performance. But there are not only performance benefits. Usually the chip design is more complex and therefore also more expensive. A dual processor design makes the software development more complicated due to concurrent accesses of common resources and it increases development time and effort. This is not acceptable for low-cost designs. Therefore Freescale implemented a co-processor on the HCS-12X 16-bit microcontroller family, which reduces the load of the CPU but does not complicate the design or slow down the overall system performance. The first product of the family, the MC9S12XDP512 is available and in production. It is the first derivative from a whole low-end 16-bit microcontroller family.

Would you like to read more? Please subscribe to the CAN Newsletter. It's free!

To top

Surge suppression solutions for CAN

CAN system designers are being challenged to meet stringent surge suppression requirements while increasing reliability and reducing the size and cost of their products. Transient voltage suppression (TVS) devices can be used to increase the surge immunity without significantly adding to the cost and complexity of the transceiver circuit. External TVS devices have a higher surge rating than a transceiver’s internal IC protection circuit. In addition, TVS devices are tested with a direct coupled configuration that matches the severe surge environment of many applications. In contrast, the surge rating of a CAN transceiver can be misleading because their immunity level is determined with a capacitor that significantly reduces the energy of the surge pulse.

Internal transceiver protection circuits can be created using high voltage transistors, zener diodes, diode arrays, thyristors and over-voltage detection switches. The surge ability of a silicon device is directly related to its size; thus, it is not practical to incorporate large protection devices inside an IC. The relatively small die size of the IC’s internal TVS devices results in a protection circuit with a modest surge rating that is designed to handle only a few surge events. An over-voltage detection circuit provides a clever solution for relatively low frequency surges, but is not fast enough to protect against high frequency surges such as ESD.

Surge protection tests

Test setups: The test configurations used to determine the surge ratings for a system, transceiver and TVS device are very different. System level tests typically use a coupling clamp to induce a surge voltage on the data lines. The cable is placed between two parallel metal plates and the test voltage is applied to the plates. A capacitor of approximately 3 nF (pulse a & b) and 0.1 µF (pulse 1 & 2a) can be used as an alternative to the coupling clamp.

CAN transceivers are typically surge tested by using a 1-nF capacitor to couple the pulse to the IC pins. In contrast, TVS devices are connected directly to the signal generator. The transceiver and TVS device test setups remove the cable as a test variable; however, the test results must be analyzed with respect to the impedance of the coupling device. Also, system level and TVS devices are typically tested with 5000 surge pulses while the transceiver ratings are often determined using just three pulses.

Frequency dependence: The frequency content of the surge pulse and the frequency dependent impedance of the IC’s coupling capacitor are important parameters. The impedance of a 1-nF capacitor has a minimal effect on the high frequency pulse ‘a’ and ‘b’ tests; however, the impedance must be considered for the relatively low frequency pulse 1 and 2a tests. The capacitor impedance must be added to the voltage generator’s source impedance (Ri) to determine the maximum current of the surge.

Test results: Table 1 provides a summary of the surge ratings for an avalanche TVS diode and a typical CAN transceiver. The data shows that a 1 nF capacitor increases the apparent surge rating of the device under test. The energy transferred by the coupling capacitor is reduced because the capacitor lowers the current and shortens the duration of the pulse. A larger capacitance, such as the 0.1 µF value typically used for the ISO 7637-2 pulse 1 and 2a system tests, is required to reduce the coupling impedance. TVS diodes also provide the advantage of a high ESD rating. The transceiver’s ESD rating is adequate to avert damages that may occur during the board assembly procedure, but often is to low to prevent field failures.

Would you like to read more? Please subscribe to the CAN Newsletter. It's free!

To top

CAN and open protocols in commercial vehicles

In mid-June in Stuttgart, more than 100 users met for the second Vector symposium “The use of CAN and open protocols in commercial vehicles.” Among the numerous sensors and ECUs in a vehicle there is, according to the statement of a contributor, a large field in which networking with CAN, SAE J1939, CANopen, and LIN and FlexRay is the focal point.

Global unification

Wolfgang Appel, who is responsible worldwide as Lead engineer for E/E (electric/electronic) architectures for DaimlerChrysler trucks, demonstrated how important a uniform E/E architecture is for the company. For a broad spectrum of vehicles on all international markets with various requirements and existing standards, numerous technical, economic, and market-specific particularities must be overcome.
The main goal is, based on the new vehicle-spanning architecture, to use as many components as possible with the same interface, geometry, and mechanics. In particular, standardization thus extends from network topology to communication to standard software and network management.
In the future, the CAN bus will continue to serve as the basis for networking on-board electronics; for simple applications, it is complemented by LIN. On the protocol side, SAE J1939 is the standard. A gateway will handle the incorporation of proprietary messages and the mapping of the previous IES (a Daimler Chrysler proprietary higher layer protocol) from DaimlerChrysler to SAE J1939 requirements. With the increase in electronic functions in commercial vehicles, there is an urgent need to act to increase the reliable bit-rates for SAE J1939 to 500 Kbit/s or even better, to 1 Mbit/s.
How the design of a vehicle architecture actually looks is determined by numerous parameters, e.g. reliability, modularity, costs, compatibility, competitive advantages, safety, etc. An important criterion for the E/E architecture is how fast new, intelligent functions can be integrated into a vehicle.

Standardization and specific solutions

Joachim Lassmann highlighted the situation from the point of view of a large automotive supplier of motor and drive controllers, fuel-injection systems, combination devices, ready-to-install cockpits, etc. He heads the product management department for electronic network solutions for commercial vehicles at Siemens VDO. Networks reduce the cabling effort required in the vehicle and they offer greater reliability, safety, and diagnostic opportunities. Thus especially for the commercial vehicle sector, there are essential advantages since availability is the uppermost priority there.

Would you like to read more? Please subscribe to the CAN Newsletter. It's free!

To top