Boosting Battery Life in IoT Devices

Earlier, the assumption was unused energy from the environment, machines, people, and so on could be used to power valuable devices and this would be done for free. The assumption was based on the convergence of four key technologies to enable mass adoption of energy harvesting—efficient voltage converters, efficient harvesting devices, low-power sensors, and low-power microcontrollers. However, it was soon realized that although energy harvesting does operate for free, the system needs investment, which is not free. That has led to the thinking that perhaps energy harvesting may not be the right technology for powering smart energy applications.

Now, with the growth of IoT devices, more sophisticated sensors, more pervasive connectivity, and secure, low-power microcontrollers, there are more devices to be powered than ever before. With most devices being small and battery powered, design engineers are facing challenges such as energy efficiency and long battery life.

In reality, it is no longer worthwhile using sensors for measuring and analyzing the energy consumption of individual light bulbs, since the cost of such a system would be more compared to the energy cost to run the lamp. In addition, there are numerous low-energy-consuming light sources available.

Development of engineering systems now place more emphasis on maximizing performance and saving energy. This is because most IoT devices spend a significant part of their life sleeping or hibernating, where the part is neither operating nor completely shut down. In this state, the device is actually drawing quiescent current, and this places the maximum impact on battery life, as it contributes to the standby power consumption of the system.

The development of nanoPower technology has led to great advancements in maximizing performance and saving energy. Newer products, with advanced analog CMOS process technology, now operate in their quiescent state with nanoampere currents that are almost immeasurable. The trick in maximizing energy-saving benefits from these products is first by duty-cycling them, and secondly by decentralizing the power-consuming architecture.

Benefits of nanoPower technology also extend to their ability to turn off circuits within the system. For instance, the nanoPower architecture may allow powering critical components such as real-time clocks and battery monitoring, while cutting off power to major consumers such as the RF circuits and the microcontroller, which can either turn off or enter their lowest power-consumption mode.

System monitoring ICs play a huge role here with their small packages and nanoamp quiescent current levels. Comparators, op amps, current sense amplifiers, and more help ensure important issues such as the voltage levels on microcontrollers are at proper levels. For instance, a nanoPower window comparator monitors the battery voltage and provides an alert if the battery voltage goes beyond allowable levels. Apart from being a valuable safety function, this also helps to extend the battery life, as the microcontroller need not operate until it has received an alarm from the comparator.

Another power-saving scheme is OR-ing the battery supply with voltage from a wall wart or an additional battery, using OR-ing diodes. These are Schottky diodes in series with the battery supply for limiting the voltage drop. For instance, MAX402000 diodes can save tens to hundreds of milliWatts of battery power when used in a smart way.

What are Harmonics and What do they do?

In the 19th century, Jean-Baptiste Joseph Fourier presented his theorem, which is known as the Fourier’s theorem. According to this theorem, a periodic function of period T can be represented as the summation of a sinusoid with the identical period T, additional sinusoids with frequency same as integral multiples of the fundamental, and a possible continuous component, provided the function has a non-zero average value in the period. The first two are known as harmonics, while the third is known as the DC component.

Of the three, the first waveform with a frequency matching the period of the original waveform is called the fundamental harmonic, while the second may have more than one component. Those with frequency equal to ‘n’ times of the fundamental are called harmonic components of order ‘n’. A conclusion drawn from the above discussion about Fourier’s theorem is a perfectly sinusoidal waveform can have only the fundamental component, and no other harmonics.

This also means an electrical system with sinusoidal current and voltage waveforms has no harmonics. However, protective devices and malfunctioning equipment in an electrical system can lead to distribution of electrical power with distortions of the voltage and current waveforms, creating harmonics. In other words, harmonics represent the components of a distorted waveform, and their presence allows analysis of any repetitive non-sinusoidal waveform from a study of the different sinusoidal waveform components.

Most non-linear loads generate harmonics. When a sinusoidal voltage encounters a load of this type, it produces a current with a non-sinusoidal waveform. It is possible to deconstruct these non-sinusoidal waveforms into harmonics. Provided the impedances present in the network are low, the distortions of the voltage resulting from the distorted current are also low, and the pollution level in the system from harmonics is below the acceptable level. Therefore, even in the presence of distorted currents, the voltage can remain sinusoidal to some extent.

Typically, the operation of many electronic devices leads to cutting the sinusoidal waveform to change its rms value or to obtain a direct current from the alternate value. In such cases, the current on the line transforms to a non-sinusoidal waveform. Several such equipment produce harmonics—welding equipment, variable speed drives, continuity groups, static converters, fluorescent lamps, personal computers, and so on.

In most cases, waveform distortion results from the bridge rectifiers present within the above equipment. Although these semiconductor devices allow the current to flow for a major duration of the whole period, they stop conducting for the balance part. This creates discontinuous waveforms with the consequent addition of numerous harmonics.

Apart from electronic equipment, transformers can also be the cause of harmonic generation and pollution. Even when a perfectly sinusoidal voltage waveform is applied to a transformer and it generates a sinusoidal magnetizing flux, the magnetic saturation of its iron core may prevent the magnetizing current from remaining sinusoidal.

The distorted magnetizing current waveform from the transformer now contains several harmonics, with the third one being of the greatest amplitude. Fortunately, compared to the rated current of the transformer, its magnetizing current is only a small fraction. As the load on the transformer increases, this percentage becomes increasingly negligible.

Meca500 – The Tiny Six-Axis Robot

Although there are plenty of robots available in the market for a myriad jobs, one of the most compact, and accurate robot is the Meca500. Launched by the Quebec based Mecademic from Montreal, the manufacturers claim it is the smallest, and most precise six-axis industrialist robot arm in the market.

According to Mecademic, users can fit Meca500 easily within an already existing equipment and consider it as an automation component, much simpler than most other industrial robots are. According to the cofounder of Mecademic, Ilian Bonev, the Meca500 is very easy to use and interfaces with the equipment through Ethernet. With a fully integrated control system within its base, users will find the Meca500 more compact than other similar offerings are in the market.

Mecademic has designed, developed, and manufactured several compact and accurate six-axis industrial robot arms on the market. Meca500 is one of their latest products, the first of a new category of small industrial robots, smaller than most others are, and ultra-compact.

The first product from Mecademic was DexTar, an affordable, dual-arm academic robot. DexTar is popular in universities in the USA, France, and Canada. Although Mecademic still produces and supports DexTar on special request, they now focus exclusively on industrial robots, delivering high precision, small robot arms. With their academic origins, Mecademic has retained the predilection of their passion for creativity and innovation, and for sharing their knowledge.

With the production of Meca500, a multipurpose industrial robot, Mecademic has stepped into Industry 4.0, and earned for itself a place in the highly automated and non-standard automated industry. With Meca500, Mecademic offers a robotic system that expands the horizons for additional possibilities of automation, as users can control the robot from their phone or tablet.

This exciting new robotic system from Mecademic, the Meca500 features an extremely small size, only half as small as the size of the smallest industrial robot presently available in the market. Meca500 is very compact, as the controller is integrated within its base and there is no teaching pendant. The precision and path accuracy of the robot is less than 5 microns, and it is capable of doing the most complex tasks with ease.

Applications for Meca500 can only be limited by the users’ imagination. For instance, present applications for the tiny robot include a wide range, such as animal microsurgery, pick and place, testing and inspection, and precision assembly.

Several industry sectors are currently using Meca500. These include entertainment, aeronautics, cosmetics, automotive, pharmaceuticals and health, watchmaking, and electronics. Users can integrate the compact robot within any environment, such as their existing production line or even as stand-alone system in their laboratories.

The new category of robots from Mecademic is already smaller, more compact, and more precise than other robots are in the market. Mecademic’s plans for the future include offering more space saving, more accurate, and easier to integrate industrial robots. They envisage this will enable new applications, new discoveries, new products, new medical treatments, and many more. Their plan is now to build a greater range of compact precision robots while becoming a leading manufacturer of industrial robots.

How Efficient are Light Emitting Diodes – LEDS?

Almost all commercial and residential establishments are moving over to light emitting diode (LED) illumination, as they are guaranteed to be more efficient compared to other forms of lighting such as incandescent and fluorescent. Unless designed with care, LEDs can suffer from premature failure due to thermal issues. Under thermal stress, LEDs can permanently lose their brightness, while degrading much quicker than the manufacturer intended. That means designers and engineers need to balance the additional cost of emitters with the thermal design for providing not only an elegant design solution, but also the long life that solid state lighting promises.

With roughly 50% of the electrical energy produced worldwide being used for lighting, and the world population growing, the only two alternatives to meet the growing needs of energy are to either generate more or to make more efficient use of what we already have. Generating more energy can take several years to plan and install power plants, but improving the efficiency of lighting can effectively mitigate the rising trend of power consumption.

Providing over 100 lumens per watt, LEDs are being increasingly used for a large selection of general applications. When converting fixture designs for incandescent bulbs to those for LEDs, engineers faced issues because of the difference of their thermal characteristics. For instance, manufacturers publish the life curves for LEDs as a function of temperature, while fixture designers do not know how to handle the information.

Incandescent bulbs were actually heaters that emitted some visible light. Nearly 90% of the light emitted by incandescent bulbs fell into the region beyond 700 nanometers—the infrared region—invisible to the human eye, but perceptible as heat. This would often cause problems in the kitchen, with waste IR light promoting premature spoilage in food illuminated by incandescent bulbs.

LEDs produce light via a different mechanism. When electrons in the LED junction cross over a forbidden energy zone called band-gap and combine with holes, they produce light because the electrons lose energy. Physicists tailor the energy by adjusting the width of the band-gap, thereby producing various frequencies of light. For instance, a white LED actually generates intense blue or Ultra Violet light, which then excites a phosphor placed in its optical path, thereby turning it into white light.

However, the process of converting electrons to light photons within the junction of the LED is not a perfect one. A vast majority of the photons created within the junction is never emitted and ultimately recombine to produce waste heat. Additionally, Stokes Shift, the phenomenon that shifts the frequency of the LED emission in the phosphor to produce white light, also generates waste heat. Waste heat from both of these mechanisms must be removed from the LED junction to prevent severe damage.

Unlike their incandescent predecessors, LEDs rarely fail catastrophically. Their slow degradation affects the photon emission mechanism, resulting in a dimming effect. Engineers use two industry end-of-life metrics for measuring the life of LEDs. One is the L70 or time taken to reach 70% of original emission, and the other is L50 or time taken to reach 50% of the original emission. The industry uses the L70 point as the useful life of an LED fixture or bulb.

What is Cabinet-Free Motion Control?

Controllers, drivers, and servomotors usually control automated platforms and machines in the automated production industry. With the evolvement of technology for machine motion, control and driving of individual machine axes is being increasingly taken over by highly intelligent electronics. Therefore, the control cabinet is assuming the central role with the rest of the system being designed around it.

With the rest of the machinery developing much more slowly, the faster evolving complex automation design and development becomes a cost-constraint for the OEM, system designers, and end users. Control cabinets need redesigning, especially with the increasing numbers of servo-driven axes. Typically, the location of the control cabinet is relatively fixed on the machine, which limits the manufacturers’ ability to modify and update the footprint of their machines.

As a solution to the above constraints, system designers are moving towards a new concept where the motion control and servo-drive mechanism is distributed rather than bound within a physical cabinet. By locating the controllers, servo drives, and power supplies nearer to the motors and axes they control, OEMs and system designers overcome several challenges arising from installation, cabling, and multiple engineering.

Initially, system designers had reoriented their designs in attempting to drive multiple machine components with a single servo motor. Although this approach had the benefit of reducing the physical number of servo motor and drives, it required a larger motor with higher power to handle the load, and several additional mechanical components for delivering the centralized power. A Cartesian motion system with a single motor for a palletizing application is an example of such a centralized approach.

By separating the servo motors on each axis, mounting them on the independent frames, and driving them separately, system engineers were able to use smaller motors, thereby reducing the overall power requirement, and developing a solution with higher efficiency.

One of the barriers to cabinet-free motion control architecture comes from PLC limitations. By limiting the axis count supported by their PLCs to 16 or 32 axes, some manufacturers force users to purchase a second PLC, which means addition of a more expensive control box with higher capacity.

For some time now, OEMs have been following a common practice of moving power supplies, servo drives, and related devices out of the control cabinet and placing them closer to each motor and its drive axis. This trend began with several leading suppliers introducing electric motors with their drives integrated into the motors’ housings. This required control electronics to be shock and vibration resistant as well as capable of withstanding the higher temperatures usually associated with environment outside the control cabinet.

Recent advances of cabinet-free components include separate ac-to-dc power supplies, independent drive units capable of mounting close to the servomotor on the machine, and power cables integrating communication capable of daisy-chaining several drive-integrated servomotors into a single circuit.

A further introduction of newer motion controllers or PLCs is helping the cabinet-free technology portfolio. These integrate the controller hardware into modules capable of mounting on the machine along with the necessary power supplies and drives. This eliminates the requirement of a control cabinet entirely.

BrailleBox with the Raspberry Pi

Reading, whether online or from the page of a book is a simple affair for those endowed with the power of sight. However, for those who are sightless, or have lost their eyesight, totally or partially, reading can be cumbersome, if not impossible. The Braille system, by allowing a changeover to the sense of touch, helps sight-impaired people to read.

Braille uses a system of raised dots that blind or those with low vision can follow with their fingertips. It is not a separate language, but rather a code for representing individual alphabets of a language. So far, the Braille system covers several languages, including Chinese, Arabic, Spanish, English, and dozens of others. Thousands of people all over the world use the Braille system of dots in their native language, providing a means of literacy for all.

The main code for reading materials in the US is the Unified English Braille, and seven other English-speaking countries use this code.

As such, Braille is useful when the material is in printed form. However, the challenge lies in reading online material. Although text-to-speech software packages are available, they are expensive and not very useful when the reader, say, wants to move back and forth while reading.

As a solution to the above problem, Joe Birch has built BrailleBox, a simple device to convert online news stories to Braille. His BrailleBox works with Android Things, News API, and the popular single board computer, the Raspberry Pi 3 or RBPi3.

Being a symbol system for people with visual impairment, the Braille system consists of letters and numbers as raised points in an array. Commercial systems are available and they produce Braille dynamically, but they are very expensive and out of reach of most people. Therefore, Joe built a low-cost alternative, the BrailleBox, which is simple to create.

Joe uses the News API as a tool that fetches jSON metadata from more than 70 news sources online. The API can integrate articles or headlines into text-based applications and websites.

The Braille system uses an array of six dots arranged in an array of three rows and two columns. Apart from representing the alphabets and numbers with various combinations of the six dots, they also represent whole words, sometimes in contraction. For instance, contracted braille includes 75 short form words and 180 different letter contractions. These help to reduce the volume of paper necessary for reproducing books in Braille.

To make the six dots for forming the Braille symbols, Joe attached wooden balls atop solenoids. He arranged the solenoids in an array of 2×3, and wired them individually to GPIO pins of an RBPi3.

Being an Android engineer, Joe controls the solenoids through Android Things, running on the RBPi3 as self-booting BrailleBox software. The reader has to push a button, which makes the program fetch a news story using the News API. As the RBPi3 deciphers the alphabets, it operates the solenoids, moving the dots.

Joe’s project is still in prototype stage, and he is yet to move all hardware inside a proper box. He also wants to add a potentiometer, preferably foot operated, so the readers can set their own reading speed.

Industrial Motors for Machine Automation

Industrial engineers use different types of motion control devices for improving the production rates and efficiencies on the floor of automated factories. Three major types of motion control devices are in demand for machine automation—stepper motors, servomotors and variable frequency drives (VFDs).

In general, stepper motors along with their drives, and controllers are widely used as they offer simple implementation, beneficial price/performance ratios, and high torque at low speeds. This motor is essentially a brushless DC version, moving in equal fixed steps during rotation, and only a single step at a time. Not requiring tuning or adjustments, stepper motors provide very high torque at speeds below 1000 RPM. They are cost-effective, as their prices are substantially lower than the cost of comparable servo systems. Since the torque they produce decreases as they speed up, it makes their operation difficult. Therefore, the work done by stepper motors becomes impractical at speeds in excess of 1000-1500 RPM.

Servomotors come with a motor, drive, a controller, and a device for positional feedback. For variable load applications, engineers prefer them to stepper motors, as they deliver high torque when rotating at speeds above 2000 RPM. Servos require adjustments and tuning, making them more complex to control compared to stepper motors. Including maintenance costs, their positional feedback arrangement can push their prices well beyond those of stepper motors.

Costing less than stepper motors or servomotors, VFD systems include an AC motor and a drive, but are unable to provide positioning. However, they can be good for applications requiring speed control on variable loads. For applications where the motor need not run continuously at full load, a VFD system can save considerable amount of energy. Another feature of VFDs is their soft-start capability, allowing a limit to high inrush currents.

In a stepper motor system, the controller regulates the position of the step, the torque generated by the motor, and the speed of the motor as it moves from one step to another. The driver operates on the control signals the controller generates by modifying and amplifying these signals to regulate the direction and magnitude of the current flowing into the motor’s windings. This way, it drive rotates the shaft of the motor to its desired position, and holds it in position with the required torque for the required time.

Controllers for stepper motors can be either open or closed loop types. Open-loop controllers are simpler, not requiring any feedback from the motor, but are less efficient. Open-loop controllers operate on the assumption the motor is always at the programmed step position and is producing the desired torque.

On the other hand, closed-loop controllers always operate with feedback based on the effective load on the motor. Therefore, the performance of the closed-loop stepper motor controller is similar that of a servo motor, and makes the operation more efficient.

Making a stepper motor rotate through each of its steps requires energizing the several windings within the motor in a specific sequence. Typically, stepper motors rotate 1.8 degrees per step, necessitating 200 steps to make a complete revolution.

What is Open Bionics?

There are people all around the world that may loose limbs for various reasons — wars, illness, and accidents being the three major ones. Artificial limbs do alleviate a part of the loss these folks experience, but often, their high cost means not all can afford a prosthetic limb. Open Bionics is a company making affordable bionic arms, making kids feel like superheroes.

A start-up tech company in the UK, Open-Bionics is changing the way people see prostheses. The 3-D printed prostheses Open Bionics makes are nearly 30-times cheaper than those available in the market are. Their biggest advantages are the myoelectric sensors that attach to the skin for detecting muscle movements. Detection of muscle movement controls the artificial hand in closing and opening fingers.

The bionic arms that Open Bionics makes are custom-built for individual children and require about 40 hours for manufacturing them. As the child grows, a revolutionary socket adjusts to the changing size. As these are small and lightweight, children as young as eight can use the bionic arms with ease.

According to the COO and co-founder of Open Bionics, Samantha Payne, they work with the NHS for creating prosthetics that are affordable and highly functional. These are meant especially for children, and come with removable covers—allowing them to choose whether they want to be Queen Elsa, or an Avenger today.

The company has a royalty-free agreement with Disney. That means they can base the removable covers on the bionic arms on characters from Star Wars, Frozen, Iron Man, and more—this can be life changing for small children, as Samantha Payne assures. For instance, Tilly Lockey, who is testing the latest model from Open Bionics, has a prototype hand themed on Deus Ex, a video game.

Open Bionics builds assistive devices offering people who use them greater freedom and independence. Moreover, as the devices are affordable, it brings bionic technology within the reach of most patients. That is why trials of bionic arms are reaching children as young as eight.

Most available prostheses do not suit young patients, as they are either way too big or very expensive. The 3-D printed bionic limbs from Open Bionics are different as they are custom-built to suit small sizes, and they are affordable. Samantha Payne feels highly satisfied seeing a young child moving their fingers individually for the first time.

Rather than making a drab skin-colored artificial limb, Open Bionics is making their arms belong to the science fiction universe. With themes from Star Wars, Disney, and Marvel, kids feel proud when wearing their prostheses. As these arms are sleek and super stylish, amputee children can identify them with their personalities and that is what makes them and the people at Open Bionics so excited.

At Open Bionics, the task begins with scanning the person’s limb using a tablet. A plan for the design of the prostheses follows, leading to a 3-D printout. The result is a low-cost, multi-grip, and lightweight bionic arm with great control. The royalty-free theme designs make the device hyper-personalized. The presence of nearly 5 million upper-limb amputees worldwide gives an estimate of the market potential for Open Bionics.

How are RS232 and RS485 Different?

When engineers need to connect electronic equipment, they resort to serial interfaces such as the RS-232 and RS-485. Although dozens of other serial data interfaces exist today, most are meant for use in specific applications. A few of them are considered universal, such as I2S, MOST, FLEX, SPI, LIN, CAN and I2C. Other high-speed serial interfaces are also used, including Thunderbolt, HDMI, FireWire, USB, and Ethernet. Despite the proliferation of interfaces, the two legacy interfaces, RS-232 and RS-485, continue to survive, used in several applications.

As a rule, serial interfaces provide a single path for data to be transmitted over a cable or wirelessly. Although some applications do use parallel buses, serial interface alone provides the only practical option for high-speed data movement today over any distance greater than several feet.

RS-232

RS-232 is one of the oldest serial interfaces, originally established in 1962, as a method of connecting a DTE or data terminal equipment such as a teletypewriter to a DCE or data communications equipment. Personal computers earlier had an RS-232 port, commonly called the serial port, to connect to a printer or other peripheral device. Embedded computer development systems still use the port today, as do many scientific instruments, and several industrial control equipment.

Officially, the standard defining the RS-232 serial interface is the EIA/TIA-232-F, with F signifying the most recent update. According to the standard, a logic 1 is defined as a voltage between -3 and -25 V, and a logic 0 as a voltage between +3 and +25 V. The logic 1 is generally termed as a mark, with logic 0 being termed as a space. Any voltage between +3 and -3 V is termed invalid and is rejected, providing a huge noise margin for the interface. The configurations of the receiver and transmitter are both single-ended and referenced to ground or 0 V.

The cable medium in RS-232 can be simple wires in parallel or a twisted pair. According to the standard, the cable length must not exceed 50 feet. However, by reducing the data rate, it is possible to use longer lengths of cable. For a 50-foot cable, the highest data rates in RS-232 are roughly 20 Kbits/s, and matched generator and load impedances are necessary for eliminating reflections and data corruption. Although earlier 25-pin connectors were used, the de-facto standard for RS-232 is the 9-pin DE-9 connector today.

RS-485

The EIA/TIA standards also define the RS-485 interface, now commonly known as TIA-485. This is not only a single device-device interface, but is a complete communication bus used for simple networking of multiple devices.

Rather than a single-ended voltage referenced to the ground, the RS-485 uses differential signaling on two lines. A logic 1 is a voltage level greater than 200 mV, while the logic 0 is a level greater than +200 mV. The maximum cable length for RS–485 is about 4000 feet or 1200 m, with typical data rates as 100 Kbits/s. However, compared to the speed of the RS-232 interface, a 20-meter cable in RS-485 can allow a maximum data rate of 5Mbits/s. Industrial control equipment using the RS-485 use the 9-pin DE-9 connector.

Raspberry Pi Controls the Cardboard Dog

This is a project for beginners using the Raspberry Pi (RBPi) single board computer. The RBPi is used to control a servo for turning the head of a cardboard dog away whenever a person is looking at it. This is to mimic a begging dog that seems ashamed of its begging nature.

This project requires the SBC RBPi, its power supply with the 5 V micro-USB cable, a USB keyboard and mouse, a display, and an HDMI cable. For storing the OS, an 8 GB micro SD card is also necessary. Another computer will be necessary to write the OS to the micro SD card and edit the files in it. The official PI camera will help to recognize the faces looking at the dog, and a micro servomotor is required to turning the head.

The RBPi will be controlling the servo through its GPIO pins. The servo has three wires that need to connect to the GPIO pins using female connectors. The camera has a ribbon cable, which goes into the port labeled camera on the RBPi. The HDMI cable goes into its port on the outside of the RBPi, and its other end goes to the HDMI-compatible TV or monitor.

Download and install the latest version of the Raspbian (with Pixel) from the official website of the RBPi. While installing the image on to the micro SD card, the process will destroy all data on the card, so be sure there is nothing of value before you begin.

Once the OS is installed on the micro SD card, insert it into the slot on the reverse side of the RBPi. If the power cord is now plugged into the RBPI socket and the power turned on, there should be some code running on the monitor screen, with the desktop showing up at the end. At this time, right click anywhere on the desktop and select “Create a New File.” Name the file Dog Turn.py, and select it to open with Python 2 IDLE.  Now open IDLE, and paste the code from here into it.

To make the code in the file to work, the RBPi will need additional Python modules to be installed. These are the libopencv-dev, python-opencv, python-dev, and you must use the sudo apt-get install command to download them.

The cardboard dog for this project uses four 9×6 inch cardboard rectangles, and two 6×6 inch squares, which form the main body. A hole at the top of the box allows the servo to go through. Another 5-inch cardboard cube forms the head, and attaches to the servo. Some cardboard legs make the dog look more realistic.

The entire electronic hardware can fit within the body of the dog. It may be necessary to use standoffs to hold the RBPi in place. The camera should look out from one of the eyeholes in the dog head. Fix it in place so that the cable has sufficient play when the servo moves the head. Simply running the python code should be enough to let the dog do its trick. To stop, turn off the power.