Digital bus architecture reduces audio system costs
Car manufacturers strive to make future generations of vehicles safer, smarter and more fuel efficient than their predecessors. To accomplish this, electronics systems continue to increase in number and complexity as more ECUs (Electronic Control Units) are deployed in the vehicle – enabling features such as active noise cancellation, internet connectivity and in-car communications. By Ken Waurin, Segment Marketing Manager, Analog Devices
As the number of ECU’s increases, so does the weight and cost of the required cabling harnesses that provide connectivity between the various ECUs. This added weight, in turn, decreases the fuel efficiency of the vehicle - much to the chagrin of car manufacturers.
Car manufacturers must balance their desire to deliver advanced, feature-rich infotainment systems, with the need to comply with government imposed fuel efficiency standards. Reducing the weight of existing cable harnesses can lead to significant improvements in fuel efficiency.
The current landscape
Traditionally, automotive audio ECU’s have been connected by either individual analogue cables or existing digital bus architectures – both of which present limitations, inefficiencies and unnecessary expenses. Car audio systems that use analogue wiring require dedicated expensive, shielded cables for each audio signal or channel. In today’s premium audio systems supporting multi-channel (5.1 or 7.1) Dolby or DTS decoding, the number of required cables rapidly increases. Additionally, the required ADCs and DACs increase total system cost while introducing potential areas of audio performance degradation.
Digital bus standards such as MOST or Ethernet AVB have been widely adopted in current generation infotainment systems because they greatly simplify the wiring complexities associated with analogue implementations. However, the added performance and flexibility of MOST and Ethernet AVB carry the added cost of expensive MCUs to manage their associated software protocol stacks. Additionally, these digital bus architectures are inherently non-deterministic – with variable system delays from node to node. This is not acceptable for latency-sensitive applications such as active noise cancellation.
Introducing the Automotive Audio Bus
The Automotive Audio Bus (or A2B) is an innovative and application-tuned technology from Analog Devices that has been proven to reduce the weight of cable harnesses by up to 75% while delivering high fidelity digital audio. The Automotive Audio Bus is optimised for audio applications delivering superior audio quality relative to analogue connectivity at significantly lower total system cost than existing digital bus standards. In its simplest form, A2B is a high bandwidth (50Mb/s) digital bus capable of transporting I2S audio and I2C control data together with clock and power using a single, two-wire, unshielded twisted pair cable over significant distances - up to 10m between nodes.
Automotive Audio Bus basics
From an implementation standpoint, A2B is a single-master, multiple-slave, line topology. The AD2410, the first member of a planned family of A2B-enabled devices, supports daisy chaining of up to eight slave nodes on a single UTP wire which eliminates the redundant wiring “home runs” required by some competing technologies. This daisy-chaining capability allows a maximum A2B bus distance of up to 40m with up to 10m supported between individual nodes. Using a line topology instead of ring, is an important element of the A2B technology related to overall system integrity and robustness. If one connection of the A2B daisy chain is compromised, the entire network does not collapse. Only those nodes downstream from the faulty connection are impacted by the failure. And the embedded diagnostics inherent to the A2B technology are able to isolate the source and cause of the failure.
The A2B master-slave-line topology is inherently efficient when compared to existing digital bus architectures. After a simple bus initialisation, zero additional processor intervention is required to manage normal bus operation. The A2B technology does however have an interrupt provision where audio nodes can signal when service and maintenance is required. As an added benefit of the A2B’s unique architecture, system latency is completely deterministic (two cycle delay) irrespective of the audio node’s position on the A2B bus. This feature is extremely important for emerging applications such as active noise cancellation where audio samples from multiple remote sensors must be processed in a time-aligned fashion.
The AD2410 A2B transceiver delivers audio, control, clock and power over a single, 2-wire, UTP cable. This reduces overall system cost for a variety of reasons:
- The number of physical wires is reduced relative to traditional implementations;
- The actual wires themselves can be lower cost, lower weight UTP as opposed to more expensive shielded cables; and
- Most importantly for particular use cases, the A2B technology offers a “phantom power” capability that delivers up to 300mA of current to audio nodes on the A2B daisy chain. This “phantom power” capability eliminates the need for local power supplies at the audio ECU – further reducing total system costs.
The total 50Mb/s bus bandwidth delivered by the A2B technology supports up to 32 upstream and downstream audio channels using standard audio sample rates (44.1, 48kHZ, etc.) and channel widths (12, 16 & 24-bit). This provides significant flexibility and connectivity to a wide range of audio I/O devices. Maintaining a completely digital audio signal chain between audio ECUs ensures the highest quality audio is preserved without introducing the potential for audio degradation via ADC/DAC conversion.
As described earlier, system level diagnostics are an essential component of the A2B technology. All A2B nodes have the capability to identify a variety of fault conditions including opens, wires shorted together, wires reversed, or wires shorted to power or ground. This capability is important from a system integrity standpoint because in the case of opens, wire shorts or reversed wire faults, A2B nodes are still fully functional upstream of the fault. The diagnostic capability also provides for the efficient isolation of system level failures which is critically important from the dealer/installer standpoint.
Designing an A2B-enabled system is greatly simplified using the SigmaStudio graphical design environment – the same development tool that supports the Analog Devices SigmaDSP and SHARC processor families. SigmaStudio provides the capability to initialise an A2B network and configure all registers via a user-friendly, industry-leading tool chain. An A2B bus bandwidth calculator as well as a BERT are also included within the SigmaStudio environment. A broad array of full-featured evaluation systems is available to quickly prototype an actual A2B network to facilitate early system proof of concept, test, verification and debug.
Target markets & applications
A wide variety of markets – both mature and emerging – will benefit from the efficiencies offered by the A2B technology. Target applications include:
- Audio ECU (Head unit, amplifier, smart antennas) connectivity;
- Microphone arrays for hands-free/speech recognition/in-car communications;
- Active noise cancellation; and
- Active speakers.
Wide Band ANC (WB ANC) is a derivative of the broader ANC application area and is now under investigation by many auto manufacturers. In WB ANC, tones harmonically related to a reference input as well as some non-harmonically related or random tones can be cancelled. These reference inputs are typically supplied via accelerometers physically distributed around the vehicle – most often in the four wheel wells. Additional inputs to the WB ANC system would come from error microphones, again physically distributed about the inside of the vehicle – with one required for each passenger “quiet” zone. A traditional implementation uses dedicated analogue connections from each accelerometer/microphone to the WB ANC processing unit. Clearly, this approach can be prohibitive in terms of both wiring cost/complexity as well as connector area on the processing unit. All of these issues are resolved using an A2B implementation.
In addition to ANC applications, efficient & cost effective microphone connectivity is quickly becoming a design priority. Bluetooth connectivity and hands-free / speech-recognition systems are becoming standard equipment and E-Call systems are expected to be mandatory is some regions. The trend from the OEM standpoint is to move toward multi-microphone systems configured individually or as modules. In either case, a system that uses A2B technology has a significantly lower overall system cost when compared to analogue connectivity, especially in multi-microphone arrays. Each AD2410 can support up to four PDM-input microphones, eliminating the need for three of the four microphone wires in a four microphone array.
Audio ECU connectivity is perhaps the most attractive, near term application area for using A2B technology. In a simple example where a head unit connects to a premium trunk amplifier, A2B removes the need for individual wires for multi-channel audio, plus NAV, plus Cellphone, plus Chime, plus… A2B replaces all of these wires with a low-cost, single two-wire UTP cable which has been tested and confirmed to pass the most stringent automotive EMC and EMI compatibility requirements.
Summary
A2B is a digital bus architecture that provides an optimised feature set, performance level and system cost for wiring-intensive audio applications:
- A2B delivers superior audio quality relative to analogue connections while providing the scalability of more costly digital bus architectures;
- A2B provides a low risk solution in many automotive applications. The AD2410 is fully released to production and qualified to the most stringent automotive standards;
- The first OEM series deployments start in 2016; and
- The AD2410 is the first in a planned family of A2B-enabled devices, with a roadmap to lower cost and higher integration.