Low-power radios enable medical wireless sensors

Low-power radios enable medical wireless sensors

WPANs and WBANs can support continuous data streaming with low-power usage

BY REGHU RAJAN
Technical Marketing
Microsemi
www.microsemi.com

Today’s sensing and monitoring solutions for wireless personal area networks (WPANs) and wireless body area networks (WBANs) can support continuous data streaming with extremely low power consumption. This is critical for wearable medical systems where frequent battery replacement would be difficult and expensive. While these systems previously required AA or AAA batteries, they can now run on smaller batteries. Making this possible are short-range radio transceivers optimized for power efficiency across a number of key parameters.

WPANs occupy a network space around an individual covering the living or working space nearby (typically 10 m) while WBANs occupy a smaller wireless space of about one meter around a person handle human body sensor communication. Both usually use Bluetooth or ZigBee protocols. Applications have expanded from low duty-cycle spot measurement to more data intense semicontinuous links. There are a variety of uses for this technology in hospital and clinical facilities, clinical home monitoring and ambulatory applications, and consumer health and fitness (see Fig. 1 ).

Fig. 1: External sensing use cases and technology requirements.

Radio requirements for WPANs/WBANs

Many issues must be considered when selecting a short-range radio transceiver capable of the power efficiency needed in WPANs and WBANs. Among these, power supply voltage is particularly important. Most sensors run on a single-cell battery, depending on chemistry, so sub-2-V supply voltages are preferable. Transceivers designed to operate down to 1.1 V are preferable to optimize design flexibility and reduce power management constraints. Radios that operate at 2.5 V will consume twice as much power as those operating at 1.25 V. Higher operating voltage is only required when output power in excess of 5 dBm is needed. In short-range applications, output power rarely exceeds 0 dBm.

Systems with a current profile without excessive peaks aid noise reduction and fit a higher supply impedance. Wireless-based sensor networks always rely on some level of duty-cycling to save power and restrict the usage of radio space, which can generate peaks in the current consumption profile of the sensor. Low peak current consumption in the transceiver reduces constraints on the wireless sensor’s power supply.

Transceiver output impedance is also important, as it has a major effect on power amplifier power consumption. Most radios have output impedance below 100 Ω. Low impedance is only required for high-output-power, long-range applications, and results in up to five times higher current consumption than higher-output impedance options that are more suited for short-reach wireless interconnect applications. Overall, assuming a similar receiver sensitivity and PA efficiency, a high impedance 900 MHz radio would use only 1 mW in its PA to achieve the same range as a 50-Ω 2.4-GHz radio that might use 25 to 40 mW.

The choice of carrier frequency also influences power consumption. The two available options within the medical (ISM) radio band are 2.4 GHz or sub-gigahertz frequencies. Protocols such as Wi-Fi, Bluetooth, and ZigBee often use 2.4 GHz. In low-power and lower-data-rate wireless medical monitoring applications, however, sub-GHz wireless systems offer reduced power consumption and longer range for a given power.

The Friis Equation quantifies the superior propagation characteristics of a sub-giahertz radio, showing that path loss at 2.4 GHz, for a given setup, is 8.5 dB higher than at 900 MHz. This translates into 2.67 times longer range for a 900-MHz radio since range approximately doubles with every 6 dB increase in power. Sub-gigahertz carrier frequencies also reduce the risk of interference from airways that are crowded with colliding 2.4-GHz Wi-Fi, Bluetooth and ZigBee signals used in everything from wireless hubs and computers to cellphones and microwave ovens. Sub-gigahertz ISM bands are mostly used for proprietary low-duty-cycle links and are not as likely to interfere with each other. The quieter spectrum means easier transmissions and fewer retries, which is more efficient and saves battery power.

Furthermore, the narrower sub-GHz bandwidth allows efficient operation at lower transmission rates. For example, at 300 MHz, if the transmitter and receiver crystal inaccuracies are both 10 ppm, the error is 3 kHz for each. For the application to efficiently transmit and receive, the minimum channel bandwidth is two times the error rate, or 6 kHz, which is ideal for narrowband applications. The same scenario at 2.4 GHz requires a minimum channel bandwidth of 48 kHz, which wastes bandwidth for narrowband applications and requires substantially more operating power.

Carrier frequency also has a major impact on the average power budget at the network level. ZigBee and Bluetooth offer highly sophisticated link and network layers, but these stacks account for up to 50% to 75% of the radio power consumption, with larger overheads. For ultra-low-power systems, the “one size fits all” standardized option is rarely the optimum solution. Instead, designers developing solutions for these applications should consider using the protocol best suited for their need.

Finally, link data rate is one of the most important factors influencing power consumption in duty-cycled wireless links. The average power is almost inversely proportional to the link data rate; so a 100-kbit/s radio will consume almost half the power of a 50 kbit/s radio for the same payload. When comparing RF transceivers, “energy per bit” is a better indicator of power efficiency than current consumption. But high-data-rate radios are often those with the higher peak currents, and these are highly undesirable for most small batteries as they need large, leaky, storage capacitors.

The aforementioned factors are critical for applications where power is at a premium and payload is greater than 10 bits/s. Whereas previous body-worn wireless sensors could only be used for slowly varying parameters, new RF technologies can be used to help observe more rapidly changing physiological parameters, such as heart and brain electrical activity or blood oxygenation, that require data rates on the order of 0.5 to 5 kbits/s to extract meaningful waveforms.

Balancing transceiver power and performance

It is very challenging to achieve radio performance capabilities while meeting requirements for extremely low power and a careful choice of radio architecture and building blocks are critical. Today’s ultra-low-power RF transceivers achieve high performance from low current by incorporating circuit development techniques such as weak-inversion CMOS-based design.

When operating in a weak inversion region, there are numerous trade-offs to be made related to parasitic capacitance, noise, and transistor size. Smaller device sizes reduce capacitance but increase noise and circuit mismatches.

The latter is particularly troublesome for ultra-low-power transceivers that process signals at low drain voltage and, hence, low signal amplitudes, and must keep mismatches to a minimum. Matching is costly, and consumes silicon real estate and/or voltage head-room to be effective.

Other trade-offs must be made related to bandwidth, which may also be a key parameter in certain radio blocks, and potentially leads to compromises elsewhere in the circuit. In the typical transceiver, dozens of these trade-off decisions may need to be made across as many as 100 functional blocks, or more.

Fig. 2: Block diagram of a typical wireless sensor based on the ZL70250.

One example of a solution derived from a careful balance of these trade-offs is the ZL70250 transceiver from Microsemi (see Fig. 2 ). Housed in a 2 x 3-mm chipscale package, it has standard two-wire and SPI interfaces for control and data transfer using any standard microcontroller. The microcontroller’s A/D converter connects to the ultra-low-power analog front-end device. The resulting solution can be used to develop a wireless ECG solution that can run continuously from a CR series coin cell for up to a week. Similar power efficiency can be achieved with such devices as a three-axis accelerometer or pulse-oximeter for patient respiration measurement, and a variety of other wearable health-monitoring platforms. ■