Author Archives: K J

Improving Context Awareness in Hearables

On-ear devices, also known as hearables, are one of the fastest-growing devices in the consumer electronics market today. Although at present their role only covers hearing aids and a tool for listening, the on-ear devices have progressed to being wireless. Now there is a brand-new way for engaging this technology to the wider world around us.

Qualcomm conducted a survey in 2019 and found that more than half the respondents were interested in hearables that are context-aware. One of the most useful capabilities the respondents were looking for in hearables was background noise reduction, and the other, dynamic volume adjustment.

The interest that users are showing for these features is for the next-generation of hearables. They are looking for a better, more intensive listening experience. With an increasing interest in hearable, users are now expecting these next-generation features that are currently not available.

For instance, traditional hearables may be wireless but controlled by the phone. While jogging or working out at the gym, it is inconvenient for the user to stare at their phones for adjusting the volume of their headphones. Even with buttons available on the headphones, they are likely to be tiny and not visible when the headphones are on the user’s ears. That makes it very difficult to locate and use the buttons.

One way of improving the user interface would be to add gesture control. Simple gestures and motion tracking can provide instructions for specific actions and controls. For instance, a simple tap on the earbud could mean an increase in volume. Tapping the entire headphone is much simpler than finding and pressing a specific button on it.

A gesture for detecting in-ear presence could automatically pause the audio as soon as the user removes the earbuds from their ears. The audio can resume the moment the user inserts the earbuds back in their ears.

As the range of movement of the human head and ears is relatively consistent compared to that of the pocket or the wrist, hearables can be ideal for tracking fitness. However, motion tracking needs to be precise to not generate false positives and negatives. Therefore, with proper fitness algorithms, it is possible for hearables to track whole body movement such as when running, biking, or standing in a queue. Accurate classification is necessary to convert step counts to calorie counts, providing a more complete picture of the user’s day.

Hearables with spatial audio can change the sound as the head turns, and linked with the accurate head tracking, can put the user right in the center of the orchestra, leading to a truly immersive listening experience. However, this also requires the latency to be low, so that the sound changes with the head movement without delay. This can help and elevate the user’s experience with XR or gaming as well.

Today’s headphones cannot provide the user with the above experience. Users may have to turn down the volume or remove at least one earbud when they want to listen to the external world. This is because the design of the hearables blocks most external sounds to enable the listener to focus on the audio.

Sensitive Magnetic-Field Sensor Has Low Noise

Although applications for magnetic sensors cover a vast field, ranging from the gigantic magnetic resonance imaging or MRI systems to sensing tiny gear-teeth, they are one of the most overlooked or misunderstood among the modern sensors in use. Researchers are constantly on the lookout for increasingly small but more sensitive magnetic-field sensors. However, sensitivity alone is not the only qualifying parameter for such sensors—low-level transducers require to be low-noise as well.

That is exactly what researchers at Brown University have developed. Their magnetic sensor is not only sensitive, it exhibits a very low noise level. With support from the National Science Foundation, the researchers have developed a device that, as a part of an arrangement of a magnetic immunoassay, looks for pathogens in fluid systems using magnetism. They claim that as the device is extremely small, millions of such sensors can fit on a single chip.

The basic principle behind the sensor is the Hall effect. In a Hall effect sensor, passing a direct current through it when the sensor is perpendicular to a magnetic field, causes the development of a voltage at right angles to the current path. The presence and magnitude of the magnetic field directly influence the presence and magnitude of the voltage.

The researchers at Brown University have developed a variation of the Hall effect sensor and have named it the Anomalous Hall Effect or AHE, and this occurs in ferromagnetic materials only. The difference between the two effects is that while the conventional Hall effect is the result of charge on electrons, the anomalous Hall effect is due to electron spin.

As electrons with various spins orient themselves in different directions, the AHE detects this with a small but definite voltage. Incidentally, magnetic fields cause many interesting phenomena on atomic particles. For instance, MRI systems capture signal source emissions related to the magnetic moment of the hydrogen nucleus.

The researchers fabricated the device as an ultra-thin film made of ferromagnetic materials like boron, iron, and cobalt, with electron spins arranged in in-plane anisotropy—meaning, the electron spins align themselves in the plane of the film. However, exposing the film to a high temperature and a strong magnetic field can change the spin of the electrons to perpendicular anisotropy, and their alignment turns perpendicular to the film.

Equalizing the two anisotropies results in a reorientation of the electron spins when the material encounters any external magnetic field, providing a reorientation voltage across the AHE. Compared to a conventional Hall-effect sensor, an AHE sensor is about 20X more sensitive.

The thickness of the AHE device offers a tradeoff in performance. A thick film requires a strong magnetic field to reorient the spins, resulting in a reduction in sensitivity. However, in a thin film, the electrons tend to reorient their spins by themselves, reducing the usefulness of the sensor. The researchers tried many thicknesses and found 0.9 nm thickness worked the best.

As magnetic anisotropy is highly dependent on temperature, researchers are using temperature to fine-tune a single magnetic AHE sensor, thereby achieving very low levels of intrinsic noise during its operation.

Using Integrated Power Switches

Power switches are most commonly in demand for their simplicity in turning on and off a voltage rail or for protecting a power path. Engineers find load switches easier to use compared to discrete power MOSFETs. For complete power protection of the system, eFuses offer an integrated approach. The combination of load switches and eFuses offers more than significant PCB space savings. Compared to discrete circuits, the combination of load switches and eFuses, also known as integrated switches, offers substantial improvements in performance, while resolving common power management challenges such as faster current limiting, detecting, and responding to mistakes in field wirings, and improving battery life and power density.

In fact, using the right integrated switch helps to reduce EMI and heat generation, while improving the power efficiencies to 90%. Bad power management leads to several side effects such as the generation of excess heat, electromagnetic interference, inaccurate voltage control, and these can lead not only to poor device performance but even to its outright failure. For the above reasons, designers are using integrated switches in electronic equipment such as desktop computers, LCD TVs, and plasma TVs.

Using integrated power switches offers several advantages over solutions of discrete controller and MOSFET. The loser component count leads to lower cost and higher reliability.

With electronic products shrinking in size, PCB space is almost always at a premium. The integrated power switch with its lower footprint has a better advantage over discrete components. Several manufacturers offer a variety of integrated power switches, and these include Fairchild Semiconductors, Power Integrations, ON Semiconductors, and ST.

Fairchild Semiconductor offers their new Green FPS e-Series of integrated switches as a replacement for conventional, flyback converters using hard switches. The new e-Series are a versatile set of devices for improving efficiency by reducing switching losses in the MOSFET with quasi-resonant operation.

It is also possible to use the e-Series in the continuous conduction mode or CCM in fixed frequency operations. The design offers simplicity and lowers the ripple current. Using an advanced burst mode technique, devices of the e-Series also conform to several governmental agency requirements for standby efficiency.

Fairchild uses a prefix of FSQ in the part number of these devices, and they are available for applications that can deliver up to 90 W. Depending on the requirement, it is possible to avail the series in seven different packages including DIP, TO-220F, LSOP, and others.

The devices use valley switching along with inherent frequency modulation for the quasi-resonant operation. The improves efficiency while reducing the EMI signature of the power supply. Valley switching uses the natural resonance of the primary inductance of the transformer and both circuit capacitance and parasitic capacitance for turning the MOSFET on only when the drain-to-source voltage is at its minimum. This reduces the amplitude of the current spike at turn-on typically found in hard-switched converters.

The increased efficiency from reducing the turn-on current spike also reduces the stress on the MOSFET. However, with valley switching, the power supply can operate with a variable switching frequency, changing with changes in the line and load conditions, helping to reduce the EMI the power supply generates.

What is ESD Protection?

For many years, people have realized the importance for ESD protection for electronic equipment. With increasing levels of integration, and shrinking of process geometries, damage induced by ESD presents a greater threat than ever before. With the addition of increasing demand for portable electronic gadgets, protection from ESD at the system level is more of an essential issue.

The advantage of providing systems with comprehensive ESD protection is far-reaching while being a low-cost way of preventing problems at any time in the life cycle of the product. When an ESD related problem remains hidden within a system, it often increases the chances of field failures, a reduction of system life, and other problems related to quality that is often challenging to manage and costly to resolve. Product recalls and field repairs including after-market costs can be financially crippling, while brand reputation takes a serious hit from quality issues.

One can find innumerable protection devices in the market, ranging from simple single-protection diodes to highly integrated complete companion ICs. Most ESD solutions come in a wide package portfolio, ranging from leaded to leadless packages. Furthermore, package innovations such as solderable side-pads are available that allow visual inspection, while pass-through routing packages ensure an optimal board layout together with an enhanced ESD protection as the designer can position the device closer to the interface. These are available for automotive and industrial applications.

BroadR-reach from Nexperia is a relatively new Ethernet standard targeting mainly automotive applications. Compared to regular Ethernet specifications, BroadR-reach has the advantage of not requiring any transformer. That increases the speed of signals per channel to 100 Mbit.

Typical applications for BroadR-reach are in vehicles as high-speed interface for ADAS or infotainment systems. The BroadR-reach ESD protection is meant for protecting the new 100 Mbit interface. Although single-channel protection is common, multi-channel protection is also available along with other devices.

Vehicles use Low-Voltage Differential Signaling or LVDS, DisplayPort, and High-Definition Multimedia Interface for in-vehicle networking and ultra-high-speed interfaces. Nexperia offers the PESD1LVDS for providing ESD protection for LVDS in vehicles. The device allows TMDS lines to pass-through while routing, easing the design, and minimizing parasitic influences. The PESD1LVDS is an ideal choice and popular for protecting high-speed interfaces in the automotive market, as the product is AEC-Q101 qualified.

Contactless ID functions are user-friendly due to their use of Near Field Communication or NFC standards. Typical usage includes easy contactless data transfers, ticketing in public transport, and contactless payment systems.

In most NFC designs, small contacts connect the NFC controller IC to the antenna. Therefore, these contacts provide an easy path for the entry of ESD, damaging the controller IC. Systems that use NFC essentially require external ESD protection. Nexperia offers the PESD18VF1BX and PESD24VF1Bx for minimizing the impact of impedance mismatch, loading on the antenna, and as optimal protection against ESD strikes in NFC systems.

General Purpose Input Output or GPIO also require protection against ESD, but they do not require high-frequency signaling. Three key elements are essential for the selection of ESD protection—package variant, one or multichannel, and uni- or bi-directional.

Isolated RS-485 Transceivers

A standard RS-485 transceiver sends and receives digital signals between digital equipment. They use positive and negative signals limited to 5 VDC levels. One can connect them as simple point-to-point configuration or as multi-point connections with two or more devices communicating. RS-485 transceivers allow high-speed communication in electrically noisy environments, as is usual within industrial plants.

Each of the two output lines on an RS-485 transceiver uses square waves to send serial data to another distant transceiver. A capacitive line offers high impedance to high-speed transmission, distorting the rise and fall of the signals. A capacitive line is one where the line carrying the signals is close to the signal ground.

Rather than use capacitive lines, RS-485 transceivers use a balanced line where the two output lines carry voltages of opposite polarity all the time. In balanced lines, the signal rise and fall times are much better, resulting in transmitting high-speed signals over longer distances.

Using +5 VDC and -5 VDC for each of the two output lines alternately an RS-485 transceiver can offer either non-inverting or inverting signals on its output lines. When the output is non-inverting, its polarity is the same as that at the input of the transceiver. For the inverting pin, the polarity is always opposite to that at the input.

In an industrial application, using isolated RS-485 transceivers is the normal practice, as the interconnecting cable often must pass through an environment with high voltages present. The isolation prevents any high-voltage spike inadvertently appearing on the interconnecting cable and passing on to the circuit driving the transceivers.

Analog Devices offers two types of isolate RS-485 transceivers. The ADM2867E is signal and power isolated up to 5.7 kV rms, while the ADM2561E has isolation levels up to 3 kV rms.

Both transceivers pass radiated emission testing, conforming to the requirements of EN55032 Class B standard. The tests use a double-layer PCB with two small 0402 size external ferrite beads on isolated ground and power pins.

Both devices feature integrated but isolated DC-DC converters generating low EMI. The isolation barrier offers immunity to system-level EMC standards. On the A, B, Y, and Z pins of the RS-485, a family of isolator devices offer ±15 kV air and ±12 kV contact ESD protection complying with the IEC6100-4-2 standard. Cable invert pins on the device allow users to reverse cable connections to quickly correct the connection while maintaining fail-safe performance on the receivers.

A double-layered PCB reduces the design time and material costs while providing Class B radiated Emissions. The cable invert feature reduces debug time during system install by allowing users to easily correct installation errors. With a greater than 8 mm creepage and clearance, and the IEC 6100-4-2 ESD and 5.7 kV digital isolation, the RS-485 transceivers from Analog Devices can maintain signal integrity even when the signals are passing through the harshest of environments.

Isolated RS-485 transceivers are useful for industrial automation, communication, building and infrastructure, and in aerospace and defense, mainly because of their cable invert feature, high isolation, low EMI/EMC capabilities, good surge protection, and improved ESD safeguards.

24-Bit Quad-Channel ADC

Analog Devices is offering a 24-bit Quad-Channel ADC, the AD7134, a low noise, precision type, simultaneous sampling Analog to Digital Converter that while offering exceptional functionality and performance, is also easy to use.

The AD7134 operates on the continuous-time sigma-delta or CTSD modulation scheme. This helps to remove the traditional requirement of a sampling switched capacitor circuitry preceding the sigma-delta modulator—simplifying the input driving requirement for the ADC. The device also has inherent antialiasing capability, arising out of the CTSD architecture rejecting signals around the aliasing frequency band of the ADC. Therefore, this ADC does not need the regular complex antialiasing filter.

The four independent converter channels of the AD7134 operate in parallel, and each of them has its own CTSD modulator along with a digital filtering and decimation path. Therefore, the user can sample four separate analog signals, each with a maximum input bandwidth of 391.5 kHz. The four signal measurements can also achieve tight phase matching among themselves. Therefore, the AD7134 can offer a high density of multichannel data acquisition in a small form factor, because of its simplified requirement of analog front-end, and a high level of channel integration.

ADCs normally require a complicated signal chain and an analog front-end circuitry that introduces distortion, mismatch, error, and noise at the ADC output. As the AD7134 simplifies the signal chain requirements, it also improves the system level performance of the device.

Offering excellent AC and DC performance, the bandwidth for each ADC channel of the device ranges from 0 to 391.5 kHz. Therefore, the AD7134 is an ideal choice for acquiring data with universal precision, and capable of supporting a variety of sensor types ranging from shock and vibration to pressure and temperature.

With several configuration options and features, the AD7134 offers the user the flexibility of achieving an optimal balance between power, accuracy, noise, and bandwidth for specific applications.

Analog Devices has integrated an asynchronous sample rate converter or ASRC with their AD7134 for precise control of the decimation ratio. This, in turn, allows the AD7134 to support a wide range of output data rates or ODR frequencies ranging from 0.01 kSPS to 1496 kSPS as the ODR uses interpolation and resampling techniques. Furthermore, as the adjustment resolution between the ODRs is less than 0.01 SPS, the user can vary the sampling speed granularly to achieve coherent sampling.

The user can also use multiple AD7134 devices with synchronous sampling between them using a single system clock, and this is because of the ASRC slave mode operation. The slave mode simplifies the requirement of clock distribution for a data acquisition system of medium bandwidth as each ADC no longer requires routing of low jitter, high-frequency master clock from the digital back end.

The AD1734 can perform on-board averaging between two or four of its input channels. This results in improving the dynamic range while the device maintains its bandwidth. Combining two channels improves the results by about 3 dB, while combining all the four channels offers an improvement of nearly 6 dB.

Flexible Heaters and Multiple Heating Zones

Flexible heaters are suitable for a wide range of uses that require variable heating options. While providing optimal heat transfer, they also offer the right temperature for products like foodservice, medical devices, sensors, instrument panels, and electronics.

We are accustomed to thinking of heaters in standard shapes like square, round, and rectangular. However, customized flexible heaters are available in a wide variety of shapes that have the requisite shape to wrap around specific objects like inserts and pipes. They may also have different temperature zones for applications that generate their own heat in some places while requiring heating in others.

Designers make flexible heaters from polyimide and silicone rubber, and their size depends on the resistive element necessary. For very thin flexible heaters, etched foil heaters are the most suitable, and both polyimide and silicone materials can use them. Etched foil heaters are also suitable for smaller flexible heaters, as they can be as small as 1-inch square, or as large as 18 x 24 inches for silicone rubber, and 10 x 70 inches for polyimide. For larger sizes, designers prefer wire-wound resistive elements.

If necessary, designers can make flexible heaters in odd and non-symmetrical shapes as well. However, the specific needs of the application almost always define the shape, requiring laser cutters and mechanical equipment for creating the outline of the desired shape. Recent developments ensure that internal cutouts are also possible, such as in rectangular, square, circular, and other shapes, without sacrificing the reliability and heating capacity of the flexible heater.

Etched foil elements allow quick thermal transfers and faster warmups for heaters made from both polyimide and silicone rubber. However, wire wound heaters are notably slower, and there may be a delay in heat transfer when the heater element is wire-wound. Wire wound heaters are not suitable for polyimide heaters.

Some applications do not require heating equally throughout the surface. For instance, some electronic circuitry may create its own heat, protecting itself from external or internal temperature changes. Such applications do not require additional heating. However, outside this protected zone, the rest of the circuitry may require suitable temperature control for proper operation. Engineers provide suitable heating with cut-outs along the heater material. Flexible heaters with multiple heating zones are the answer for applications that require heating but at different temperatures.

Flexible heaters made from polyimide or silicone rubber are the most suitable for such applications. Multiple zone heating is necessary if some part of the electronics requires heating at a certain temperature, while another part needs raising to a different temperature.

Making heaters with multiple heating zones requires designers to place different heating elements at suitable places, with separate controls for these zones. However, other simpler options are also available, requiring one heater and a single controller. One of the options involves using elements with variable widths.

The width of the conductor of foil etched heaters impacts the watt density in specific areas of the flexible heater. When the width of the conductor is low, its resistance increases resulting in lowering the thermal output at that zone, creating multiple heating zones.

Near-Zero-Power Sensors

The US Department of Defense has a research and development agency, the defense advanced research projects agency or DARPA, which have recently concluded a program N-ZERO—the near-zero power rf and sensor operations. One dramatic outcome of this program is sensors in the battlefield earlier running out of power in weeks or months can now keep operating for more than four years before they need a replacement of their coin batteries.

DARPA’s initiatives began about five years back, with the aim of improving IoT battery power so that sensors could operate in the field without requiring frequent battery replacement. The army deploys sensors in the field for detecting battlefield signals like sound, light, and vibrations.

According to DARPA, it is possible to improve battery lifetimes by using sensors that are idle most of the time, waking up to a triggering event, or periodically monitoring battlefield events as they happen, rather than monitoring them continuously.

In the field, troops can now gather data and intelligence from potential combat zones without venturing there personally. They also do not have to move into hazardous areas frequently for replacing dead batteries for sensors.

According to Benjamin Griffin, program manager of DARPA’s Microsystems Technology Office, unattended, untethered systems can now benefit from the N-ZERO program using idle but continuously alert sensing capabilities. Radiofrequency signatures or specific physical signals can trigger these sensors. As the sensor lifetime now extends to years, deployment of such sensing technologies can now be more cost-effective and safer in areas that lack fixed-energy infrastructure.

With progressive improvements in sensor technology, sensor capabilities are continuously expanding. As explained by Griffin, the N-ZERO program has been successful in developing sensors that wake up with infrared, acoustic, and RF signals. These near-zero-power sensors detect thermal radiation, measure sound levels, and communicate with radio frequency signals.

Even when coin-cell batteries power these sensors, their power consumption was so low the estimated battery life could extend from a few weeks to four years. However, during testing, processing, and communications of confirmed events limited the N-ZERO initiatives. Ultimately, the self-discharging of the batteries would also be another limiting factor.

The DARPA initiative has also been effective in developing an ARM M0N0 processor with ultra-low-power capabilities. The processor consumes only 10 nW when idling, and 20-60 µW/MHz when active, depending on the application. As sensors shutdown, they store the information in read-only memories (ROM), which the users can access without any power penalty, as the data resides in non-volatile memory.

As their power consumption is so low, the ARM processor can run for decades on a single set of batteries. Conventional coin batteries can have a lifetime of several years if the sensors operate in sleep mode, but not much beyond.

Sensors that operate continuously consume a lot of power, but often spend time processing useless data. In contrast, low power consumption processing options such as the ARM processor can cut costs when replacing failed batteries frequently. An example implementation for processing audio files could run continuously for more than 200 days when powered by a single LR44 coin-cell battery.

Why Low Dropout Regulators?

In this era of high-efficiency switching power supplies and voltage regulators, low dropout (LDO) regulators seem almost out of place. Contrary to popular belief, low dropout regulators are small components, simple to use, and cost-effective for obtaining an output of regulated voltage from an input of higher voltage.

For system designers, low dropout regulators offer a simple method of obtaining a voltage from a source that is very close to the output voltage. This is one major reason designers use LDO regulators widely. The second reason is LDO regulators are analog devices, and unlike switching regulators, introduce very low noise into the system.

Small LDO regulator devices such as those from Diodes Incorporated offer a variety of features such as high-power supply rejection ratio, ultra-low quiescent current, wide input voltage handling capability, physically small footprint, and high output current supply capability.

Keeping in line with other SMT components, manufacturers are making LDO regulators in smaller form factors, enabling designers to use PCB space more effectively. Designers can make better use of the newer families of LDO regulators in highly dense PCBs as these components are of very small size, and occupy the minimum space, while they offer the same high-quality performance.

Not all power supply sources offer clean and regulated outputs. LDO regulators help filter out most of the noise from unregulated power sources with their high-power supply rejection ratio specifications. By rejecting the noise from the power source, LDO regulators provide noiseless and spike-free DC power to ensure the system operates reliably.

Many systems do not require continuous power. In remote areas, where it is difficult to deliver power, engineers rely on batteries to power their equipment. LDO regulators with ultra-low quiescent current consumption are a boon, as they consume the minimum amount of power when the system is idle, resulting in a significant increase in the life of the battery.

LDO regulators can handle a wide range of input voltages, in some cases, up to as high as 40 VDC. In multi-voltage systems, which are now common-place, such LDO regulators are very cost-effective, and they make the design more robust and reliable.

Sensors and related electronics work better with clean power supplies. Noise from switching regulators can limit the sensitivity of sensors drastically, resulting in reduced coverage or misleading measurements. LDO regulators supplying clean and efficient power with high current output allow using components for sensitive measurements, without the introduction of ripple and noise. Even with their high current output, LDO regulators work with voltage differentials as low as 350 mVDC.

Automotive applications require high-temperature reliability, and LDO regulators are available that cover a wide temperature range of -40 ºC to +125 ºC. This is a necessary feature in an automobile, as many applications must work concurrently to keep the vehicle operational.

The new family of LDO regulators are ideal for portable and small consumer devices, such as smartwatches, smartphones, wearables, wireless earphones, smart homes, smart offices, and different sensor applications. The industry uses these LDO regulators for other applications such as healthcare devices, smart meters, and other devices powered by batteries.

Motor Run & Motor Start Capacitors

Electric motors exploit the interaction between two magnetic fields for rotating a shaft. The stator windings generate one of the fields, and the rotor windings provide the other. In some DC motors, permanent magnets replace one of the windings, while the commutator, whether brushed or brush-less, changes the direction of the current in the other winding to continuously alter the interaction between the two magnetic fields to allow the motor to rotate.

In three-phase AC motors, the interaction between the three incoming phases creates the rotating magnetic field in the stator windings, and this pulls the rotor along, making it rotate. The so-called single-phase AC motor is, in reality, a two-phase AC motor that is operated with a single-phase supply, with capacitors generating the second phase. These motors require two capacitors, one to start the motor, and the other, to keep it running.

A capacitor is a device to store charge. In the DC circuit, a capacitor will charge up and stay that way until allowed a path to discharge. In the AC circuit, where voltage and current change polarity regularly, the capacitor charges up to the peak voltage in one cycle, then discharges and again charges up to the negative peak in the next cycle, with the rate of charging and discharging dependent on the capacitor value and the impedance in the circuit.

Another important factor is the voltage on the capacitor does not follow the input voltage while it is charging and discharging—it lags behind. Even though the supply voltage may be at its peak, the voltage on the capacitor reaches the peak only after the capacitor charges. Likewise, as the supply voltage moves towards the negative peak, the capacitor voltage follows more slowly as the capacitor has to first discharge. This lag helps to create the second phase for the motor.

A motor starting from rest requires a high starting torque, but once it has started moving, requires a smaller running torque to keep it in rotation. That means a larger capacitor is necessary for starting the motor—providing it with a larger starting current. In fact, motors use a centrifugal relay to cut out the start capacitor from the circuit after the motor has reached a certain speed. The run capacitor, though, has to remain connected to the motor at all times.

As the run capacitor is engaged in the circuit continuously, they are oil cooled and in metal, cases to allow heat dissipation. As they face peak to peak voltages all the time, their voltage rating tends to be on the higher side—typically, 1.5 times the line voltage, although the capacitance value may be low, ranging between 5 µF and 45 µF. On most 240 V systems, run capacitors are likely to be rated 370-440 VAC, and in 480 V systems, 600 VAC capacitors are more common. Run capacitors are rated for 100% duty cycle.

Start capacitors, being of larger capacity, are physically larger as well. As the start current does not need to be very precise, start capacitors are available as 8.3 µF, 15 µF, 43 µF, 60 µF, and above. Common voltage ratings for start capacitors are 110, 125, 165, 220, 250, or 330 VAC.