Category Archives: Guides

Let ezIGBT guide you to the right IGBT

Those who design with IGBTs or Insulated Gate Bipolar Transistors know how difficult it is to select the proper component for a specific application. Several factors have to be carefully weighed simultaneously before you can zero in on the right product. IGBTs are very useful in equipment handling high power, with huge currents passing through them. Failure due to a mismatch involves high expensive replacements. Therefore, engineers have to be very selective when deciding on the right component to use.

For IGBTs, just as for any other semiconductor component, the junction temperature plays a very crucial role in maintaining the proper operation and life of the component. Other parameters that the designer must take into account are the operational voltage, the continuous and maximum currents and the power that the IGBT must handle. While the designer has to rely solely on the information provided to him by the manufacturer, the datasheet of a specific component may be cumbersome and static. From the steady-state figures provided by the manufacturer, it may be difficult to estimate the performance a particular part will play when placed in a circuit.

The website of ezIGBT makes things easier for the designer. It liberates him from the paper data and allows him to model the performance of a part, given specific conditions of junction temperature, operating frequency, applied voltage, current and size of heat sink. The website provides several IGBT models and ezIGBT takes great care to maintain the accuracy of the data.

On the website of ezIGBT, the designer has to supply several operational parameters. These include the on-time voltage across the Collector and Emitter or Vceon, the forward voltage across the copak diode or Vf, on-time voltage or Von, off-time voltage or Voff and the reverse recovery voltage or Err. The website then uses Industry standard techniques to generate equations in two sets: one at 25°C and the other at the highest possible junction temperature.

The equations compute the losses and thermal conditions for the device selected, with the parameters being linearly interpolated between the two temperatures. The datasheet graphs are generally curve-fitted, and expected errors remain less than 10%, which are below the normal variations in manufacturing these devices. At present, ezIGBT is only providing calculations based on “hard switching” type of converters. Calculations for resonant operations are planned for the future.

The website provides three types of tools: Analyze, Compare and Recommend. The Analyze tool helps the designer calculate the power loss at the operating frequency, when the designer specifies the junction temperature and duty cycle. This tool also allows an understanding of how the losses are split between the IGBT and its diode. With this data, the website also estimate the size of the heat sink required.

On the Compare page, you can run the analysis calculated in the previous Analyze section, along with the additional ability to compare the performance of multiple IGBTs.

When you choose your operating conditions, ezIGBT can recommend the parts that will most closely match the requirements of your application. This is done in the Recommend section. At present, the selection depends on the incremental junction temperature rise over the case temperature. Other criteria are expected to be addressed in the future.

How to select the best inductor

A good understanding of the desired in-circuit performance is required for a proper selection of an inductor. Fortunately, this relates to the information available in data sheets supplied by the manufacturer.

Any electronic equipment needs power to operate satisfactorily. Portable instruments invariably use dc-to-dc converters for generating the required voltages from power sources such as batteries. Additionally, as equipment size reduces, so does the need for optimizing inner parts, especially the power conversion solutions; it is no longer possible to stick on to the “one size fits all” approach. For example, manufacturers are producing low profile components, which are proving to be very popular.

Inductance is the primary functional parameter that defines an inductor. This value, calculated by the converter design equations, determines the ability of the inductor to handle the desired output power and control the ripple current in a dc-dc converter. However, specifying the inductance value alone is not sufficient for defining the required inductor. You also need to specify the DCR, SRF, Isat and Imms for a complete specification.

The DCR or the DC resistance of an inductor depends on the length and diameter of the wire used in manufacturing. The power dissipated by the inductor depends on the DCR. Therefore, if the inductor were to be used in an enclosed space, a lower value for the DCR would certainly be an advantage. Please note, resistance offered and hence power dissipated at high frequencies usually depends not on DCR, but on skin effect.

Another important factor is the SRF or the Self-Resonant Frequency of the inductor. At this frequency, an inductor begins to resonate naturally, using the distributed capacitance characteristic of the winding. In a dc-dc converter, SRF might cause unwanted oscillations and lead to instability of the output. The manufacturer specifies the SRF and it depends on the nature of the construction. The designer usually selects an inductor with an SRF that does lies outside the operating frequency range of the application.

Inductors with cores are liable to saturate. When in saturation, the inductance value falls drastically, upsetting the operation of the dc-dc converter. Isat or the Saturation Current, specified by the manufacturer is therefore, an important parameter, which states the amount of current that would cause saturation and subsequently a fall in the inductance value. Designers need to select an inductor that has an Isat value larger than the maximum currents expected in the converter.

Power dissipation in an inductor is dependent on two factors – the DCR and the Irms or the RMS Current flowing through it. Manufacturers specify the maximum rms currents that the inductor can handle causing a specified temperature rise. Therefore, the designer has to select an inductor whose Irms value is larger than the maximum rms currents that the converter is expected to deliver.

Fortunately, dc-dc converters are rather forgiving towards tolerances of inductors. These applications do not require inductors to have extremely tight tolerances to achieve their outputs. Standard tolerances indicated by manufacturers are +/-20%, which most converter designs accept as suitable.

A smartphone built from Raspberry Pi: the PiPhone

You may not be looking for a new cell phone right now, but someone has just managed to transform his Raspberry Pi (RBPi) into a working cell phone. David Hunt has used only off-the-shelf components and put them together for the project. Although it is not as slick as the regular cell phones available in the market, at about $160, David has created a one-off project that certainly has no economics of scale working for it. The best part is all components of the phone can be taken apart at any time, used for some other projects and then reassembled. Can you do that with your regular cell phone?

David has called his cell phone the PiPhone, in honor of the base RBPi that powers it. The other major parts used for making the PiPhone work are a Sim900 GSM/GPRS module and an Adafruit touchscreen interface. The GSM/ GPRS module allows the cell phone to make and receive calls, while the touchscreen provides the user interface. A 2500 mAh LiPo battery powers the entire contraption. The GSM module connects to the RBPi through a UART, while the battery fits in between the TFT screen and the RBPi, allowing the PiPhone to work standalone, without wall warts or dangling wires.

The touchscreen interface has a numeric keypad displayed on the screen. After dialing the required number, you need to touch the phone icon at the bottom to make the call or hang up.

The Sim900 GSM/GPRS is an intelligent module, which oversees the entire communication process of the PiPhone, including sending the standard AT commands to the RBPi for making calls, hanging up and sending text messages or data. Towards the bottom of the PCB is the connector where you can insert your SIM card. Therefore, you can use a regular prepaid SIM card available in the local phone store.

Just below the GSM module, there is an on/off switch and an off-the-shelf standard DC-DC converter. This converts the 3.7V supplied by the battery to the 5V required by the rest of the electronics. Heat dissipation was the only problem that David faced because of sandwiching the RBPi, the TFT, battery and the GPS module together. During development, as all the components were placed apart, they remained cool to the touch even after extended periods of use. However, sandwiching prevents air from circulating within, resulting in the CPU getting a bit warm after switching the unit on for a few minutes.

The GPS card is insulated from the RBPi with a thick foam-core board, allowing no accidental electrical connections between them. David used a couple of cable-ties to hold the different parts together.

Of course, walking into a local phone shop and picking up a normal smartphone would be simpler and cheaper, but that would not be as much fun as making your own. Moreover, as said earlier, you can put the parts to other uses as well, which you cannot do with a standard phone.

David has put up all his code and instructions on the GitHub. There are links available to instructions on installing the TFT.

What is Human Machine Interface: HMI?

Accelerating quality, quantity, economic efficiencies and environmental protection are leading to an increasingly connected process flow and factory floor. In combination with decreased personnel, that has led to processing of increased amounts of information by fewer and less application-specific operators in the control center. It requires a well-designed HMI or Human Machine Interface system to decrease the gap between the production process and the operator via an intuitive visualization system, layers of detail that allow a bird’s eye view down to the minute details, and includes training material and documentation that the operator has on his fingertips.

A well-designed HMI system provides numerous benefits. Chief among them are increased safety, quantity, quality and economic efficiency. Apart from minimizing the risk of disruption in the production process, HMI systems also reduce the over downtime while allowing fewer operators to manage more information with less field-specific knowledge.

HMI provides a means of monitoring, controlling, managing and/or visualizing device processes. For example, an operator panel may allow the operator of an industrial machine to interact with the machine in a visual, graphical way. The operator can easily control the machine by using the touch screen or external buttons, as all readouts and controls readouts are graphically displayed on the screen.

HMIs can be located on the machine, in the form of simple segmented displays or LCD panels of high-resolution. They can be located in portable handheld devices that are battery operated or in centralized control rooms. Machines and process controls can use them to connect the operator with Programmable Logic Control application systems to control sensors, actuators and machines on the factory floor.

For communicating with industrial machines, the usability of the HMI system depends on the processing power of the system, its ability to render reality-like complex screens, quick responses to user inputs and the flexibility for handling several levels of operator interactions. Usually, effective communication requires the HMI to have dynamically changing graphics. This in turn, requires the system to be a high-performance type that supports various resolutions and displays of high refresh rates. For efficient communication between the operator, numerous machines and control systems, it is imperative that multiple connectivity and protocols must be supported.

Industrial automation thrives on real-time communication. Using industrial micro-controllers along with PRU-ICSS or programmable industrial communication subsystems makes it possible to support various popular, certified serial protocols, including those that are Ethernet-based. The PRU-ICSS allows HMI manufacturers easily support industrial communication protocols of multiple types on a single hardware platform. The most important advantage of this platform is that it does not require the support of external ASICs and FPGAs. This offers huge scaling in performance and the integration offers opportunities of software and design reuse.

Portable HMI solutions use several wireless connectivity solutions such as WLAN, Sub-1GHz, ZigBee and BlueTooth. This broad portfolio offers the maximum flexibility when designing for wireless. For example, the WiLink 8 solution provides high-performance BlueTooth and Wi-Fi in one module. The Sub-1GHz performance line is very popular and the most reliable in its range.

Cubli the Baron Munchausen of Robots

When you find a cube lying innocently on the table, you wouldn’t exactly expect it to jump up and start balancing itself on its edge or on its corner, will you? Probably not, if it wasn’t Cubli, the one designed at ETH Zurich. Well, it does not actually lift itself up by its own locks as Baron Munchausen was fond of doing, but Cubli the robo-cube can bring a smile on your face when you see its capers – watch its antics below.

Cubli, the 15x15x15 cm cube, has several spinning wheels and motors inside it. The contraption lets Cubli lie on one of its sides peacefully, then jump up to stand on an edge and proceed to tilt for balancing itself on one of its corners. By combining its jumping and balancing tricks, Cubli can even walk!

A closer inspection reveals three reaction wheels within the cube, one on each axis. These are able to spin at high speeds, as they have a motor each to drive them. The combined reaction of the three spinning wheels makes Cubli attain its uniquely stable postures. In fact, the positions are so stable that you can even nudge Cubli a little and it will resist falling over. Cubli can control its fall, allowing it to “walk” on a surface.

To control the speed of each of its wheels, Cubli has three motor-controllers, governed by an on-board processor and a bevy of inertial sensors. The motor-controllers rev up their individual motors based on the commands from the processor. Sensors monitor the tilt and angular velocity of Cubli at different points and feed this data to the processor to compute how fast each of its reaction wheels needs to spin to get Cubli to maintain its position.

By suddenly stopping the spinning of one of its reaction wheels, Cubli can jump up to one of its edges from its resting position and similarly, proceed to balance itself on its corner. Satellites use the same technique to keep themselves stable in orbit and Segway keeps you from falling off its scooter. However, the algorithms that Cubli uses are entirely different.

It feels funny watching Cubli walking. From its resting position, it rears itself up on its edge, hangs there for a moment and then gradually lowers itself on to the next side, repeating this sequence for its controlled walk. When Cubli is balancing on its edge or its corner, it will maintain its balance even if you tilt the surface on which it is resting.

All this is made possible because the spinning wheels act as gyroscopes. The momentum of the spinning wheels keeps the cube balanced. When one of the wheels slows down, it loses its momentum, and the center of gravity of the cube starts to pull it down. Speeding up the wheel improves the momentum and counterbalances its center of gravity, allowing Cubli to regain its position.

Another research group is considering using this amazing technology for building robots and using them for exploring other planets.

Possibilities of harvesting triboelectricity

Although we already harvest energy from many sources, new sources are always welcome. Ignoring the up-front costs, harvesting energy is often free, convenient and eases problems of several types of practical installation and replacement. Energy harvesting requires three fundamental stages: a reliable source of energy, the harvesting electronics or the converter and a load where the energy can be gainfully employed.

Apart from finding a reliable source of energy, there is one other factor before the energy goes into the converter and load. This factor is the transducer, which will allow the energy to be gathered with reasonable efficiency and reliability. Suitable transducers examples are blades for airflow, piezo-based components for vibrations or a water-driven vehicle using temperature gradient for propulsion.

Currently we use several potential sources of energy for harvesting. Prominent among them are solar radiation, temperature difference, sound, vibration, airflow and motion. One of the most common sources that have been very hard and baffling to capture is static electricity generated from friction (also referred to as Electro Static Discharge or ESD). Generally famous for the damage it causes, triboelectricity it may finally have been tamed by the work done at the Georgia Institute of Technology.

Although a precursor to generating and collecting static electricity does exist in the form of the Van De Graaff generator, the team from GIT has developed polymer materials, which are inexpensive and flexible. In addition, these materials are very good at developing a charge through rubbing and holding it until it is extracted as current flow. Static electricity is easy to generate – simply walking on carpet is good enough – and hold – that is, until you touch the doorknob – but very tricky to extract.

The team generates power from triboelectricity by sliding two materials together and then separating them to create a gap in between. Although this may be interesting, the next question could be a tricky one – how much harvestable energy does this mechanism offer?

The GIT press release cites power outputs of the order of 300W, which is of course, significant. However, there is another side to energy harvesting. This concerns power delivery to the load. Although energy collection may happen in dribbles on the generation side, when releasing energy it to the load, it has to be done in the form of power, which is the rate at which energy is expended.

The reason for this is not hard to fathom. All loads require a minimum threshold or power to function. The 300W generated may translate into barely meaningful power levels at the load. When energy levels to be harvested are on the low side, it may actually be difficult to collect enough power after accounting for acceptable losses.

The team from Georgia Tech has done something impressive and intriguing. They claim that the materials they have developed have a volume power density of more than 400KW per cubic meter and the efficiency figure is more than 50%. In addition, the team says that their material is suitable for generating energy from contact with flowing water. That is an interesting proposition, opening up new opportunities where sinks and faucets could become the hidden sources for charging mobiles and or lighting up the kitchen.

Pairing your Raspberry Pi with an SPI OLED

OLEDs are interesting because they produce very crisp and brighter displays than is possible with the regular Liquid Crystal Displays. Being made of thin films of organic molecules, OLEDs use less power than conventional Light Emitting Diodes and LCDs. Driving OLEDs from your Raspberry Pi (RBPi) can be a great project to learn about these nifty solid-state displays and the processes that drive them.

Adafruit offers a lovely little monochrome SPI OLED module with a resolution of 128×32, driven by a SSD1306 driver chip. You can refer to excellent tutorials, libraries and guides on the website if you are driving this display from an Arduino. Although other OLED modules support I2C interface, this module supports only SPI, has pixel and text drawing functions, without any geometric drawing functions.

Although very small, only about 1-inch diagonal, the display is very readable because of the very high contrast of the OLED display. It has 128×32 individual white pixels, and the controller chip can turn each one of them on or off individually. No backlights are needed, as the OLED display produces its own light. This feature enhances the contrast and reduces the power consumption as well. The SSD1306 driver chip of the display communicates over the SPI bus, and requires four to five pins from the RBPi.

The OLED and its driver require a power supply voltage of 3.3V and logic levels of 3.3V for communication. If you are using this with the 5V supply of the RBPi, a 3.3V regulator should be used to power the display. On average, the display consumes about 20mA from the 3.3V supply, but this actually depends on how much of the display is lit. OLEDs usually require a high-voltage drive for the good contrast, but since this switched capacitor charge pump is already built-in into the display, this is one of the easiest OLED modules to interface.

Apart from the SPI specifications of the module, there is also a D/C pin, which controls the Data/Command going into the module. When you pull the D/C pin HIGH (connect it to VDD), the signals present at the pin are treated as data. When you pull it LOW (connect it to VSS), the signals present at the D/C pin are treated as command and are transferred to the command register.

What this essentially means is the opcode and the argument bytes that follow it are treated as a single command, even though this is a multi-byte command. For example, the command for setting the contrast control consists of a one-byte opcode followed by a one-byte contrast value. Therefore, the D/C pin has to be pulled LOW for the entire sequence when sending this command. Similarly, when sending image data, the D/C pin has to be pulled HIGH, which will enable the data to be moved into the image memory buffer.

The project is simplified when you use the following software – Occidentalis 0.2 for boot image, WiringPi-Python for accessing the GPIO pins and Py-Spidev for accessing the python bindings to the spidev Linux kernel driver. For further details, refer here.

What is a capacitor used for?

Just as a bucket holds water, a capacitor holds charge. In fact, the world’s first capacitor was in the shape of a jar and was aptly named the Leyden jar. However, the latest capacitors do not look anywhere close to a jar. In its simplest form, a capacitor has two conductive plates separated by a dielectric. This helps maintain an electric charge between its plates. Depending on the type, different materials are used for the dielectric, such as plastic, paper, air, tantalum, polyester, ceramic, etc. The main purpose of the dielectric is to prevent the plates from touching each other.

The Leyden jar was invented in the 18th century, at the Netherlands University. It was a glass jar coated with metal on both the inside as well as the outside, with the glass effectively acting as the dielectric. The jar was topped off with a lid. A hole on the lid had a metal rod passing through it, with its other end connected to the inner coat of metal. The exposed end of the rod culminated in a metal ball. The metal ball and rod was used to charge the inner electrode of the jar electrically. Experiments in electricity used the Leyden jar for hundreds of years.

A capacitor can be used in a number of different ways, such as for storing digital data and analog signals. The telecommunication equipment industry uses variable capacitors to adjust the frequency and tuning of their communications equipment. You can measure a capacitor in terms of the voltage difference between its plates, as the two plates hold identical but opposite charge. However, unlike the battery, a capacitor does not generate electrons, and therefore, there is no current flow if the two plates are electrically connected. The electrically connected plates rearrange the charge between them, effectively neutralizing each other.

A naturally occurring phenomenon, lightning, works very similar to a capacitor. The cloud is one of the plates and the earth forms the other. Charge slowly builds-up between the cloud and the earth. When this creates more voltage than the air (the dielectric) can bear, the insulation breakdown causes a flow of charges between the two plates in the form of a bolt of lightning.

As there is only a dielectric between the two plates, a capacitor will block direct current but will allow alternating current to flow within its design parameters. If you hook up a capacitor across the terminals of a battery, there will not be any current flow after the capacitor has charged. However, alternating current or AC signal will flow through, impeded only by the reactance of the capacitor, which depends on the frequency of the signal. As the alternating current fluctuates, it causes the capacitor to charge and discharge, making it appear as if a current is flowing.

Capacitors can dump their charge at high speed, unlike batteries. That makes capacitors eminently suitable for generating a flash for photography. This technique is also used in big lasers to get very bright and instantaneous flashes. Eliminating ripples is another feather in the capacitor’s cap. The capacitor is a good candidate for evening out the voltage by filling in the troughs and absorbing the crests.

How did the diode get it’s name?

Although most diodes are made of silicon nowadays, it was not always so. Initially, there were two types – thermionic or vacuum tube and solid state or semiconductor. Both the types were developed simultaneously, but separately, in the early 1900s. Early semiconductor diodes were not as capable as their vacuum tube counterparts, which were extensively used as radio receiver detectors. Various types of these thermionic valves were in use and had different functionalities such as double-diode triodes, amplifiers, vacuum tube rectifiers and gas-filled rectifiers.

The diode gets its name from the two electrodes it has. Both the thermionic as well as the semiconductor type possess the peculiar asymmetric property of conductance, whereby a diode offers low resistance to flow of current in one direction and high resistance in the other. Similar to its vacuum counterpart, several types of semiconductor diodes exist.

The first semiconductor diode was the cat’s whisker type, made of mineral crystals such as galena and developed around 1906. However, these were not very stable and did not find much use at the time. Different materials such as selenium and germanium are also used for making these devices.

In 1873, Frederick Guthrie discovered that current flow was possible only in one direction and that was the basic principle of the thermionic diodes. Guthrie found that it was possible to discharge a positively charged electroscope when a grounded piece of white-hot metal was brought close to it. This did not happen if the electroscope was negatively charged. This gave him proof that current can flow only in one direction.

Although Thomas Edison rediscovered the same principle in 1880 and took out a patent for his discovery, it did not find much use until 20 years later. In 1900s, John Ambrose Fleming used the Edison effect to make and patent the first thermionic diode, also called the Fleming valve. He used the device as a precision radio detector.

To put it simply, a diode functions as a one-way valve. It allows electricity to flow in one direction while blocking all current flow in the reverse direction. The semiconductor diode has an anode (A, p-type or positive) and a cathode (K, n-type or negative). Since the cathode is more negatively charged compared to the anode, electric current will not flow if the cathode and anode are charged to the same or very similar voltage.

This property of the diode allowing current to flow in only one direction is utilized during rectification, when alternating current is changed to direct current. Such rectifier diodes are mostly used in low current power supplies. For turning a circuit on or off, you need a switching diode. If you are working with high-frequency signals, band-switching diodes are useful. Where a constant voltage is necessary, there are zener diodes.

Diodes are also used for various purposes such as the production of different types of analog signals, microwave frequencies and even light of various colors. When current passes through Light Emitting Diodes or LEDs, it emits light of a specific wavelength. Such diodes are used for displays, room lighting and for decoration.

Make a time clock with the Raspberry Pi

People working on projects are usually required to keep their time records up to date. However, those who are more engrossed with the technicalities of their work, find that they slip on their time keeping chores. If you fall into that category, let your Raspberry Pi (RBPi) help you out. Along with the tiny single board computer, you will additionally need one RGB LED, an OLED character display and a rotary encoder. Your time will be logged directly into a Google Docs spreadsheet.

The purpose-built time clock has several off-the-shelf components. For the RBPi, use the Occidentalis 0.2 operating system. The 128×32 OLED display uses SPI interface and this OS has the interface bundled in.

The time clock is very simple to operate. When you start it up, it will pull in a list of jobs from a specified Google Docs spreadsheet. To scroll through the list of jobs, simply rotate the knob on the rotary encoder. When you want to record time for specific project, just locate it by rotating the knob and click it. That starts the clock ticking on the spreadsheet. For logging off, simply click once again.

Use any case suitable for the RBPi or make your own. Since the HDMI, audio and video ports will not be used it is acceptable if the case does not provide access to these ports. For those who like to do things professionally, designing the carriage for the RBPi on OpenSCAD could be great fun. Printing the carriage on a Makerbot Replicator will give the required professional touch. Since the RBPi board does not have any mounting holes, you may have to put in edge-clips for holding it. Having posts at the corners of the board will make a snug fit.

Although the documentation of the RBPi mentions some ground pins as DNC or Do Not Connect, using them as extra grounds can be very convenient, just make sure you are using the proper ground pins.

Using the GPIO pin 5 of the RBPi for the push button can be of dual advantage. This keeps the pin 5 available for a Safe Mode boot, for example, when there is a recent firmware release. Simply hold the knob down when booting up and your RBPi comes up in the safe mode. Additionally, there is no chance of the pin being accidentally held low, such as could be the case if it was used for one of inputs to the quadrature encoders.

Use the software from here. The PWM library drives the RGB LED and the RgbLed class animates its color transition loops. The rotary encoder uses the RotaryEncoder.Worker class for polling through the encoder GPIOs and for keeping the application code as simple as possible.

Since the application code utilizes the CPU only to around 20%, its temperature rise is within safe range. The color-coding gives idea of the state of the machine. When no time is being logged, the LED shows purple. As time is logged against a task, the LED pulses slowly and turns blue. When rapidly flashing, you know that the spreadsheet on the Google Docs is being updated.