By Chris Anderson, CTO, Taoglas. Article featured in Wireless, Design & Development Magazine, Vol. 25, No. 5
The connected car of the future will have more options than ever, expanding beyond basic infotainment and navigation systems to offer the latest in safety features, and (in the-nottoo-distant future) autonomous driving. The requirement for more sensors and antennas to deliver high-bandwidth, low-latency connectivity is seemingly at odds with another requirement of auto manufacturers—fewer cables and connectors that cause noise and vibrations, while being complicated and expensive to install. In today’s connected cars, antennas and electronics are increasingly being forced in closer proximity. What does that mean for design engineers?
For a modern car, electronics represent a little more than 30 percent of the vehicle’s total costs, and that percentage is expected to rise. The more electronics a vehicle deploys, the greater need for electrical power, system components, and need to interconnect those system components. Electronic connections (be it the power supply or control wiring harness), digital communications connections, or radio frequency (RF) interfaces add complexity, cost, and weight.
Automotive designers spend considerable effort to minimize these factors. Consumer expectations are such that even base model vehicles are now expected to have features like smartphone connectivity via Bluetooth, entertainment, and safety systems like the Event Data Recorder or European eCall. Hence, significant design efforts are required in this area every time for all models. Old cars dealt with the power and control needs of facets like power windows and seats, using very complicated discrete wiring systems. Their weight and cost kept these features in premium vehicles.
Bosch created the Controller Area Network (CAN) bus in the mid- 1980s to address these issues. Today, CAN allows for power and communications using simpler and cheaper connectors and wiring. Some new car features such as the entertainment center, cameras, and cellular/Internet connectivity, use variations of USB and Ethernet. However, none of these address the need to run radio signals around the car. The GNSS receiver, eCall cellular radio, SiriusXM receiver, new DSRC radio (for talking to other cars and roadside infrastructure), maintenance cellular radio, remote keyless entry radio, and tire pressure monitoring system receiver all need RF cabling, connectors, and antennas.
Strategies For Elegance
The first step in simplifying vehicle interconnects was the use of a multi-drop digital data bus instead of discrete wires. The next major step has been to consolidate functions into clusters of electronics. This often includes radio communications devices like GPS or cellular radio systems. One trade-off is how antennas for the radios in question often need to be remotely mounted elsewhere in the car for an appropriate radiation pattern to communicate with the radio link’s other end. GNSS or SiriusXM obviously need a clear view of the sky to see satellites, while cellular antennas need a clear view of the horizon.
This is where physics starts to complicate matters. Running radio frequency signals around a car from a radio and antenna is normally accomplished with coaxial transmission line cables. The thinner and lighter those cables are, the cheaper and more flexible they become. Unfortunately, this also means sacrificing performance and introducing more signal loss, which is proportional to both frequency and length of the cable.
In the case of GNSS, this situation can be easily addressed using an active antenna. The GNSS antenna contains a receive filter and Low Noise Amplifier (LNA), which is powered by a DC voltage over the coax cable. This removes the coaxial cable losses and helps retain the best possible performance. While this is normal practice for GNSS and SiriusXM receive-only systems, the process is increasingly difficult to do with radio systems that are bi-directional and also transmit, such as cellular, WiFi, and Bluetooth.
The filter and LNA added to a receive-only radio are duplications of parts already in the receiver.
In that context, they’re an added expense over what would be needed if the coax losses were low enough. For a radio that transmits, however, this sort of solution would require transmit and receive filters, a transmit power amplifier big enough for the signals in question, the receive LNA, a pair of RF switches, a power supply system, and dedicated transmit/receive control signal from the remote radio.
This, in turn, would require a coax cable between the radio, its active antenna, and also a control signal cable. Just as with GNSS, all this is already built into the cellular radio, so it’s a lot of extra expense and complication to mitigate coax losses. Therefore, it’s very rare to see an active antenna for a cellular system because of the added cost. The brute-force way of dealing with coax losses in cellular or other transmitting radios is often to use lower loss, higher-quality coax as the transmission line. There are trade-offs, however, in that the higher-performance coax is thicker, heavier, stiffer, and expensive.
The trend of co-locating the radios and their antennas will only continue, as the cost, weight, complexity, and RF performance benefits far outweigh the added design and development complexity. In some cases, however, as additional antennas are added, a certain minimum separation between antennas is required. This is likely to push most antennas into a single area of the car. It will also require that an antenna deployment area will take up more space. An example here would be the need to have four cellular antennas, GNSS, DSRC, SiriusXM, and two WiFi antennas all located in a single enclosure on the roof of a car. Car designers are not likely to accept the 400-mm diameter dome that would be the optimal size solution for this, so a lot of effort will go into understanding the complex interaction of co-locating all these antennas and the radios that use them. Some antenna companies have already been doing this sort of product for other markets. While it would be nice to get all the car antennas into a single package, some smaller number of antennas will still need to be located away from the main antenna cluster with coax; for example, AM/FM antennas, because of their size and the common need to have two of them for receive diversity. Another example would be cellular antennas that need to be physically separated to ensure maximum MIMO throughput performance. Even in these cases, it’s most likely that the radios in question would still get integrated into a TCU-type solution and the location of the antennas constrained to somewhere that keeps the coax runs short. Longer term, it would make sense to create a digital interface standard for the FM broadcast radio, GNSS, and other radios such that the radios could be integrated directly with their antenna. This would allow those radio and antenna units to be distributed around the car wherever the vehicle designer has a place to put them that lines up with the radio’s performance need while also minimizing the use of coax cable and RF interconnects. The connections would be limited to high-speed digital data and power, and the lower-cost wiring and connectors that involves. One trend that’s already in discussion is running power and communication over the same wires. Historically the power system in a vehicle was viewed as being so noisy that to make any attempt at communication over the same wires would be highly unreliable. Using new digital spread spectrum communication techniques, viable solutions have already been created that could one day allow the interconnection of vehicle systems with only two or even one wire(s).
Go Where the Antennas Are
As the number of radio systems in cars grow, a better overall solution to the coax issues is to simply locate the radios very close to their antennas. This has resulted in the radios moving into a combined electronics package. An example of one is called a Telematic Control Unit (TCU). The TCU is then physically located near the antennas, wherever they’re placed on the vehicle.
This has a number of effects:
- The longer runs of coax cables that used to go from the radios to the antennas are now effectively replaced with cheap digital communication wiring between the TCU and rest of the car systems.
- Active antennas are no longer needed because the transmission lines between radio and antenna are so short, their losses are negligible.
- The antennas all need to be in roughly the same area of the car, or the above benefits are lost.
- Co-locating radio antennas creates a greater possibility for interference between the radio systems that must be carefully designed around. When the antennas were farther apart, this issue could often be ignored.
- Co-locating the radio electronics near their antennas creates more opportunity for RF emissions from those electronics to interfere with radio reception performance. This requires additional design effort and testing.
This co-location of TCU electronics and antennas also creates a need for new RF interconnects. While using a small, short coax cable to connect the TCU radios and antennas is an obvious solution, all those connectors are potential long-term failure points. When there can be up to eight or 10 coax connections between a TCU and its antenna system, there is also potential for assembly mistakes and a non-trivial amount of assembly labor costs.
Most RF connector systems haven’t changed for decades. It’s uncommon having to connect an array of eight to 10 RF feeds between two electronic boards in a small physical space. As such, there haven’t been a lot of products to meet this need, especially in a low-cost, high-density, highvibration environment. Most of these interfaces are still proprietary custom solutions not commonly available off the shelf. Since both sides of such an RF connector interface are specific to the product to which they’re being deployed, there’s no driving need for an interconnecting standard. A generic solution for this application is a current point of research for connector and antenna companies so that future products can simply reuse a known good solution.
One of the most interesting areas of investigation is the use of a selective axis of conduction elastomer materials to create PCB board-to-board interconnects with what looks like a simple layer of rubber. There are still significant issues to be solved, such as insertion losses, isolation, and cross-talk, but the concept looks promising.
The state-of-the-art in cabling for automotive electronics focuses on minimizing the number and length of all cables in the vehicle. This has led to consolidation of the electronics into packages with similar physical needs in the car. The future of cabling in cars will continue to push towards higher-speed data buses and minimizing interconnects other than power and data. This all conspires to further complicate the design and test of the vehicle electronic systems. That additional complication highlights the need for experienced expert partners when it comes to specialties like radio electronics and antennas.