Fullscreen HTML5Med Res (??MB)HTML4

Pt.1: By JOHN CLARKE Control your power costs with the: ENERGY METER Have you recovered from the shock of receiving your last power bill? Have you resolved to reduce your electricity usage? This Energy Meter lets you accurately monitor energy usage for individual appliances and even figures out what it costs to run them. I F YOU WANT to save power and reduce costs, you need to know how much power each appliance uses over a period of time. Most appliances don’t run all the time, so you need to know the power they use while they are actually running and how 30  Silicon Chip much they use over the longer term. The easiest way to determine that is to use an electronic power meter and this new “Energy Meter” fits the bill nicely. It displays the measured power in Watts, the elapsed time and the total energy usage in kWh. In ad- dition, it can show the energy cost in dollars and cents. As a bonus, it also includes comprehensive brownout protection. One obvious use for this unit it to show refrigerator running costs over a set period of time, so that you can quickly determine the effect of different thermostat settings. Alternatively, it could be used to show the difference in energy consumption between the summer months and the winter months. If you have a solar power installation, this unit will prove invaluable. It will quickly allow you to determine which appliances are the most “power hungry”, so that you can adjust your energy usage patterns to suit the capacity of the installation. And there are siliconchip.com.au lots of other uses – for example, the unit could be used to determine the cost of pumping water, the running costs of an aquarium or even the cost of keeping your TV set on standby power, so that it can be switched on via the remote control. Standby power The cost of standby power is something that most people never think about. However, there are lots of appliances in your home that continuously consume power 24 hours a day, even when they are supposedly switched off. These appliances include TV sets, VCRs, DVD players, hifi equipment and cable and satellite TV receivers. They remain on standby so that they are ready to “power up” in response to a command from the remote control. Then there are those devices that are powered via a plugpack supply. These devices include modems, some printers, portable CD players and battery chargers (eg, for mobile telephones). However, simply switching these devices off when not in use is not the complete answer because their plugpacks continue to draw current – unless, of course, they are switched off at the wall socket. Some high-power appliances also continue to draw current when they are not being used. For example, most microwave ovens have a digital clock which operates continuously and the same applies to some ovens. Typically, the standby power usage for each of these appliances is about 2W. What else? Well, let’s not forget computers. Then there are those appliances which must always be on, otherwise there’s no point having them. These include cordless telephones, digital alarm clocks, burglar alarms and garage door openers. Do a quick audit of your house – you will be quite surprised at how many appliances you have that are either permanently powered or operating on standby power. By using the Energy Meter, you can quickly monitor these devices and find out which are the energy wasters. Perhaps when you learn the results, you will be persuaded to turn some of these devices off at the wall or even do away with them altogether! that, when it’s not being used to check energy consumption, the unit can be used to provide brownout protection for a selected appliance. Basically, a brownout occurs when the mains voltage goes low (ie, much lower than the nominal 240VAC) due to a supply fault. This can cause problems because motor-driven appliances (eg, washing machines, airconditioners, dryers, refrigerators, freezers and pumps) can be damaged by a low mains supply. If the supply voltage is low, the motor can fail to start (or stall if it’s already running) and that in turn can cause the windings to overheat and burn out. In operation, the SILICON CHIP Energy meter can switch off power to an appliance during a brownout and restore power when the power is returned to normal. The power can either be restored immediately the brownout condition ends or after a delay of 1824 minutes. This delay feature is ideal for use with refrigeration equipment, as it allows the refrigerant to settle if the brownout occurred during the cooling cycle. Using the Energy Meter As shown in the photos, the SILICON CHIP Energy Meter is housed in a rugged plastic box with a clear lid. This plastic case is important because the internal circuitry operates at mains Main Features • • • • • • • • • • Displays power in Watts Displays energy usage in kWh Displays measurement period in hours Displays energy cost in dollars and cents Brownout detection and power switching LCD module shows several readings simultaneously Calibration for power, offset and phase Adjustment of cents/kWh for cost reading Adjustment of brownout voltage threshold, calibration, hysteresis & duration. Optional delayed return of power after brownout is restored to normal voltage potential. Two 10A mains leads are fitted to the unit – one to supply power from the mains and the other to supply power to the appliance. The unit is easy to use: simply plug it into the mains and plug the appliance into the output socket. The unit is easy to build, with all parts mounted on two PC boards. Pt.2 next month has the assembly details. Brownout protection A bonus feature of the SILICON CHIP Energy Meter is the inclusion of brownout protection. This means siliconchip.com.au July 2004  31 Specifications • • • • • • • • • • • • • • • • • • • Wattage resolution ......................................................................... 0.01W • • Zero Offset adjustment .................................... 0.12% of reading per step Maximum wattage reading ....................................................... 3750.00W Kilowatt hour resolution ................................................. 1Wh (0.001kWh) Maximum kWh reading ..................................................... 99999.999kWh Cost/kWh resolution .....................................................................0.1 cent Maximum cost/kWh reading ...................................................... $9999.99 Cost/kWh setting from ...........................................................0-25.5 cents Timer resolution............................................................... 0.1h (6 minutes) Maximum timer value .................................................................. 9999.9h Timer accuracy (uncalibrated) typically ........................................ ±0.07% Maximum load current .................................................... 10A (15A surge) Reading linearity ............................................. 0.1% over a 1000:1 range Frequency range of measurement ......................................40Hz to 1kHz Battery current drain during back-up ............................................... 10mA Accuracy .............................Depends on calibration (error can be <0.5%) Accuracy drift with temperature ............................................... 0.002%/°C Brownout voltage detection accuracy after calibration ...................... ±2% Brownout return delay ........................................................18-24 minutes Wattage calibration adjustment ................... 0.0244% of reading per step (±2048 steps) Current monitoring resistance .......... 1% tolerance, 20ppm/°C coefficient An LCD display is visible through the lid of the case and the only exposed parts are four mains-rated switches. These switches are used to set the display modes, reset values and (initially) to set the calibration values. In use, the Energy Meter is simply connected in-line between the mains supply and the appliance to be monitored. The LCD shows two lines of information and this information includes: (1) the elapsed time; (2) the power consumption in watts; (3) brownout indication; and (4) the energy consumption in kWh (kilowatthours). The elapsed time is shown on the top, lefthand section of the display and is simply the time duration over which the energy has been measured. This is shown in 0.1 hour increments from 0.1h (ie, 6 minutes) up to 9999.9h. That latter figure is equal to just over 416 days or 1 year and 51 days, which should be more than enough for any application! 32  Silicon Chip After it reaches this maximum elapsed time, the unit automatically begins counting from 0.0h again. Alternatively, the timer can be reset to 0.0h at any time by pressing the Clear switch. The power consumption figure (watts) is displayed to the right of the elapsed time and is updated approximately once every 11 seconds. This has a resolution of 0.01W, with a maximum practical reading of 3750.00W (ie, equal to the power drawn by a 15A load with a 250V supply). A 10A load will give a reading of about 2400W, depending on supply voltage. Immediately beneath this figure is the total energy consumption (in kWh) since the measurement started. This has a range from 0.000kWh to 99999.999kWh, with a resolution of 1Wh. The maximum value represents over 4.5 years of energy consumption for an appliance drawing 2500W continuously. This reading can be reset to 0.000kWh by pressing the Clear switch. In this case, the switch must be held closed for about four seconds before the RESET is indicated on the display. Finally, brownout indication is shown in the lower lefthand section of the display. It displays “SAG” if the mains level drops below the selected voltage for a set time, with the unit also switching off the power to the connected appliance. Alternatively, under normal power conditions (ie, no brownout), the SAG display is blanked and power is supplied to the appliance. Function switch Pressing the Function switch on the front panel changes the display reading, so that the energy reading is shown in terms of cost instead of kWh. Once again, this reading can be reset to $0.00 by pressing the Clear switch. The maximum reading is $9999.99 but this is unlikely to ever be reached. Pressing the Function switch again toggles the energy reading to kWh again. Holding down the Function button switches the Energy Meter into its calibration modes. There are eight adjustment modes available here and these can be cycled through by holding the button down or selected in sequence with each press of the Function switch. We’ll take a closer look at the various calibration modes in Pt.2 next month. Making power measurements OK, now that we’ve looked at the main functions of the Energy Meter, let’s see how we go about making power measurements. In operation, the Energy Meter measures the true power drawn by the load. It is not affected by the shape of the waveform, provided that the harmonics do not extend above 1kHz and the level does not overrange. In a DC (direct current) system, the power can be determined by measuring the applied voltage (V) and the current (I) through the load and then multiplying the two values together (ie, P = IV). Similarly, for AC (alternating current) supplies (eg, 240V mains), the instantaneous power delivered to a load is obtained by multiplying the instantaneous current and voltage values together. However, that’s not the end of the story when it comes to siliconchip.com.au average power consumption, as we shall see. Fig.1 shows a typical situation where the current and voltage waveforms are both sinewaves and are in phase with each other (ie, they both pass through zero at the same time). In this case, the instantaneous power waveform is always positive and remains above zero. That’s because when we multiply the positive-going voltage and current signals, we get a positive result. Similarly, we also get a positive value when we multiply the negative-going voltage and current signals together. The average (or real) power is represented by the dotted line and can be obtained by filtering the signal to obtain the DC component. In the case of in-phase voltage and current waveforms, it can also be obtained by measuring both the voltage and the current with a meter and multiplying the two values together. For example, the voltage shown in Fig.1 is a 240V RMS AC waveform and this has a peak value of 339V. The current shown is 10A RMS with a peak value of 14.4A. Multiplying the two RMS values together gives 2400W, which is the average power in the load. Note that, in this case, the power value is the same whether we average the instantaneous power signal or multiply the RMS values of the voltage and current. Multimeters are calibrated to measure the RMS value of a sinewave, so if a sinewave has a peak value of 339V, the meter will read the voltage as 240V (ie, 0.7071 of the peak value). For non-sinusoidal waveforms, only a “true RMS” meter will give the correct voltage and current readings. RMS is shorthand for “root mean square”, which describes how the value is mathematically calculated. In practice, the RMS value is equivalent to the corresponding DC value. This means, for example, that if we apply 1A RMS to a 1Ω load, the power dissipation will be 1W – exactly the same as if we had applied a 1A DC current to the load. The waveforms in Fig.1 are typical of a load that is purely resistive, where the current is exactly in phase with the voltage. Such loads include electric light bulbs and electric radiators. By contrast, capacitive and inductive loads result in out-of-phase voltage and current waveforms. If the siliconchip.com.au Fig.1: this graph shows the voltage (V) and current (I) waveforms in phase with each other. Note that the instantaneous power is always positive for this case. load is capacitive, the current will lead the voltage. Alternatively, if the load is inductive, the current will lag the voltage. Inductive loads include motors and fluorescent lamps. The amount that the current leads or lags the voltage is called the power factor – it is equal to 1 when the current and voltage are in phase, reducing to 0 by the time the current is 90° out of phase with the voltage. Calculating the power factor is easy – it’s simply the cosine of the phase angle (ie, cosφ). Lagging current Fig.2 shows the resulting waveforms when the current lags the voltage by 45°. In this case, the resulting instantaneous power curve has a proportion of its total below the zero line. This effectively lowers the average power, since we have to subtract the negative portion of the curve from the positive portion. And that’s where the problems start. If we now measure the voltage (240V) and current (10A) using a multimeter and then multiply these values together, we will obtain 2400W just as before when the two waveforms were in phase. Clearly, this figure is no longer correct and the true power is, in fact, much lower, at 1697W. This discrepancy arises because the power factor wasn’t considered. To correct for this, we have to multiply our figure of 2400W by the power factor (ie, cos45° = 0.7071). So the true power is 2400 x 0.7071 = 1697W. These calculations become even more interesting when the current leads or lags the voltage by 90° as shown in Fig.3 – ie, we have a power factor of 0. In this case, the voltage and current waveforms still measure 240V July 2004  33 Fig.2: here’s what happens when the current lags the voltage by 45°. In this case, the resulting instantaneous power curve has a proportion of its total below the zero line, effectively lowering the average power. and 10A respectively when using a multimeter but the power dissipation is now zero. This is because the same amount of instantaneous power is both above and below the zero line. This means that even though there is 10A of current flowing, it does not deliver power to the load! Alternatively, we can use our formula to calculate the true power dissipation in the load. In this case, we get 240 x 10 x Cos90° = 0 (ie, cos90° = 0). So once again, we get a power dissipation of 0W, despite the fact that the current is 10A and we have 240V applied to the load. Other waveform shapes such as produced by phase control circuits, where the waveform is “chopped”, present even more difficulties when it comes to making power measurements. However, the SILICON CHIP Energy Meter overcomes these problems 34  Silicon Chip by averaging the instantaneous power signal over a set interval (11s) to obtain the true power. The result is an accurate power measurement which takes into account the phase angle and the shapes of the voltage and current waveforms. Converting the measured power dissipation (Watts) into energy consumption (kWh) is straightforward. This is simply the average power used by the appliance over a 1-hour period. So if an appliance draws 1000W continuously for an hour, its energy consumption will be 1000Wh, or 1kWh. Specialised IC The SILICON CHIP Energy Meter is based on a special “Active Energy Metering IC” from Analog Devices, designated the ADE7756AN. Fig.4 shows the main internal circuit blocks of this IC and also shows how it has been connected to the mains, to make voltage and current measurements. As can be imagined, the internal operation of this IC is quite complicated and it has a host of features, some of which are not used in this design. If you want to find out more about this IC, you can download a complete data sheet (as a pdf file) from www. analog.com. Most of the features and adjustments available in the ADE7756AN IC are accessed via a serial interface. This communications interface allows various registers to be accessed and altered and also allows them to receive processed data. As shown on Fig.4, there are two input channels – one to monitor the voltage and the other for the current. Amplifier 1 (Amp1) is used to monitor the load current but it doesn’t do this directly. Instead, it monitors the voltage developed by passing the load current through a 0.01Ω resistor (R1). The maximum dissipation within this resistor at 10A is 1W, which gives an expected 30°C temperature rise above ambient. For this reason, we have specified a low-temperature coefficient resistor to minimise resistance changes as the temperature rises. In operation, Amp1 can be set for a gain of 1, 2, 4, 8 or 16 and for a full-scale output of 1, 0.5 or 0.25V. These values are set by writing to the appropriate registers within the IC via the serial communication lines. In this circuit, the gain is set at 1 and the full-scale output at 250mV. The 250mV range was chosen to suit the 100mV RMS (141.4mV peak) that’s developed across resistor R1 when 10A is flowing through the load (which is in series). It also allows sufficient headroom for a 15A current to be measured – equivalent to 150mV RMS across R1, or 212mV peak. Amp2 is similar to Amp1 except that its full-scale output voltage is fixed at 1V. Only the gain can be set and in this case, we have set the gain of 4. As shown, the Active input from the mains is divided down using a 2.2MΩ and 1kΩ resistive divider. This divided output is at 113.5mV RMS (161mV peak) for a 250V input and this is then fed directly to Amp2. As a result, the signal level at the output will be 454mV RMS, or 644mV peak, well within the 1V full-scale output capability of this stage. The circuit is even capable of casiliconchip.com.au tering for situations where the mains voltage reaches 280V RMS (396V peak). In this case, the voltage from the resistive divider will be 180mV peak, which gives 720mV peak at the amplifier’s output. Both Amp1 and Amp2 have provision to zero the offset voltage at their output (this is the voltage that appears at the output when the amplifier’s inputs are both at ground or 0V). Of course, an ideal amplifier would have an output offset of 0V but that doesn’t happen in practice. In this application, however, we don’t have to worry about trimming out the offset voltages because a highpass filter is included in the signal chain (following multiplier 1). This filter prevents the offsets from affecting the power reading but note that offset adjustment would be required to accurately measure DC power in other circuit applications. A/D converters The output signals from the amplifier stages are converted to digital values using separate (internal) analog-todigital converters (ADC1 & ADC2). For those interested in the specifications of this conversion, the sampling rate is 894kHz and the resolution is 20 bits. An analog low-pass filter at the front of each ADC rolls off signals above 10kHz, to prevent errors in the conversion process which might otherwise occur if high-frequency signals were allowed to pass into the ADC. The output of each ADC is then Fig.3: it gets even more interesting when the current lags (or leads) the voltage waveform by 90°. In this case, the voltage and current waveforms still measure 240V and 10A respectively but the average power dissipation is now zero. This is because the same amount of instantaneous power is both above and below the zero line. Fig.4: this block diagram shows the main components of the ADE7756AN Active Energy Metering IC and shows how it is connected to the 240VAC mains supply. Two internal op amp circuits monitor the current (Amp 1) and voltage (Amp 2) signals and the sampled values are then fed to separate analog-to-digital converters. siliconchip.com.au July 2004  35 36  Silicon Chip siliconchip.com.au Fig.5: the circuit uses a PIC microcontroller to process the data from the ADE7756AN Active Energy Metering IC and to drive the LCD module. digitally filtered with a low-pass filter to remove noise. This filter does not affect 40Hz to 1kHz signals but rolls off frequencies above about 2kHz. Next, ADC1’s output is applied to a multiplier. This stage alters the digital value fed into it according to a “gain adjust” value that’s applied to the multiplier’s second input. This gain adjust value can be changed by writing to this register and in our circuit, it’s used to calibrate the wattage reading to its correct value. A High-Pass Filter (HPF) stage is then used to process the adjusted signal from the multiplier. This removes any DC offsets in the digital value and applies the resulting signal to one input of Multiplier 2. ADC2 operates in a similar manner to ADC1 and also includes a low-pass filter (LPF) stage. Another LPF stage then rolls off the signal at frequencies above about 156Hz. This effectively removes any extraneous high-frequency components in the signal before it is fed to the SAG detection circuit. This detection circuit monitors the voltage level and outputs a SAG signal if the voltage drops below the level set in the SAG register. As well as going to the LPF stage, the signal from ADC2 is also fed to a phase compensation circuit (Phase Adjust). This stage can change the signal phase relative to the signal from ADC1 and is included to compensate for any phase differences which may be caused by any current and voltage-measuring transducers (not applicable here). Immediately following this stage, the signal is applied to the second input of Multiplier 2. This effectively multiplies the current and voltage signals to derive the instantaneous power value. This is then filtered using another low-pass filter, to produce a relatively steady value, although it does allow some ripple in the output since it does not completely attenuate AC signals and only rolls off signals above 10Hz. The resulting power value is then mixed in the Offset Comparator with an offset adjustment, to give a zero reading when there is no current flowing through R1. Its output is stored in the Waveform Register, the contents of which are continuously added to the Active Energy register at an 894kHz rate. Finally, the data in the Active Energy Register can be retrieved via the siliconchip.com.au WARNING! This circuit is directly connected to the 240VAC mains. As such, all parts may operate at mains potential and contact with any part of the circuit could prove FATAL. This includes the back-up battery and all wiring to the display PC board. To ensure safety, this circuit MUST NOT be operated unless it is fully enclosed in a plastic case. Do not connect this device to the mains with the lid of the case removed. DO NOT TOUCH any part of the circuit unless the power cord is unplugged from the mains socket. This is not a project for the inexperienced. Do not attempt to build it unless you know exactly what you are doing and are completely familiar with mains wiring practices and construction techniques. Serial Data Interface. Note that the values retrieved from this register will vary, because of the ripple allowed through the LPF at the output of Multiplier 2. However, these variations are less noticeable if the period between each retrieval is made as long as possible, so that any ripple can be integrated out over time. For this reason, we have selected a retrieval interval of about 11 seconds and this removes most of the variation. That’s about the maximum practical limit, as a longer period could cause the register to overrange when high powers are being measured. Circuit details OK, so the way in which the ADE7756AN chip works is rather complicated. Fortunately, we don’t have to worry too much about this, since the complicated stuff is all locked up inside the chip. Refer now to Fig.5 for the full circuit details. Apart from the ADE7756AN chip (IC1), there’s just one other IC in the circuit – a PIC16F628A microcontroller (IC2). This microcontroller processes the data from IC1 and drives the LCD display module. And that’s just about all there is to it – apart from the power supply circuitry and a few other bits and pieces. IC1 operates at 3.58MHz as set by crystal X1 and this frequency determines all the other operating rates, such as ADC sampling and the phase variation. In addition, the device operates from a single +5V supply rail, although its inputs at pins 4, 5, 6 & 7 can go below the 0V level. In operation, the sampled current and voltage waveforms are applied to the balanced inputs of the internal amplifiers – ie, to V1+ and V1- for Amp1 (current) and to V2+ and V2- for Amp2 (voltage). These balanced inputs are provided so that any common mode (ie, noise) signals at the inputs are cancelled out. However, in order to do this, both inputs to each amplifier must have the same input impedance and signal path. So, for the voltage signal, both inputs of Amp2 are connected to a 2.2MΩ and 1kΩ voltage divider and these in turn are connected across the Active and Neutral lines. Similarly, the current monitoring inputs are both connected to series 0.01Ω and 1kΩ resistors but note that only one of these (ie, R1) carries the load current. This resistor is rated at 3W, while the non-load current carrying resistor (R2) simply consists of a short length of fine-gauge copper wire. R2 is necessary to mimic the noise picked up by R1. All inputs are filtered to remove high-frequency hash above about 4.8kHz by connecting 33nF capacitors to ground (ie, from pins 4, 5, 6 & 7). Note that the whole circuit is referenced to the mains Neutral, with the 0V rail for both IC1 and IC2 connected to this line. However, because the circuit is connected directly to the mains, it must be treated as live and dangerous (as can happen if Active and Neutral are transposed in the house wiring – eg, the power point is wired incorrectly). IC1’s reference voltage at pin 9 is filtered using parallel-connected 100µF and 100nF capacitors. This provides a stable reference voltage for the ADCs and is typically 2.4V. However, variations between individual ICs could result in a reference voltage that’s 8% above or below this value but this is taken care of by the calibration procedure. July 2004  37 Fig.6: the top trace in this scope shot is the voltage that appears on pin 7 of IC1 (TP2). This is the sampled mains voltage from the 2.2MΩ and 1kΩ resistive divider. The lower trace is the current waveform at pin 4 of IC1, resulting from a 4.3A load. This produces a 43.45mV RMS signal across the 0.01Ω current sensing resistor (R1). Fig.7: this scope shot, captured at the output of the Energy Meter, shows the operation of the brownout feature. In this case, the brownout protection is set to switch off below 203V RMS (288V peak) and power is restored only when the voltage increases by the hysteresis level (35V RMS or 50V peak) – ie, to 238V RMS. WARNING: these two scope waveforms are shown to explain the operation of the circuit. DO NOT attempt to monitor these waveforms yourself – it is too dangerous. The SAG output appears at pin 13 and is normally held high via a 1kΩ pull-up resistor. This, in turn, holds Mosfet Q1 on and so relay RLY1 is also normally on (assuming link LK1 is in position). Conversely, when a power brownout occurs, the SAG output goes low and Mosfet Q1 and RLY1 both turn off. The SAG output from IC1 also drives RA1 (pin 18) of IC2 and this does two things. First, it “instructs” the microcontroller to send the SAG indication data to the LCD display when a brownout is detected. Second, it allows IC2 to provide the optional delayed turn-on feature after a brownout via RB0 and LK2 (ie, LK2 used instead of LK1). When the SAG output goes low, RB0 also immediately goes low and turns off Q1 as before. However, when the brownout ends, RB0 remains low and only goes high again after an 1824 minute delay to switch on Q1 and RLY1 and thus restore power to the appliance. Note that the relay contacts are used to break the power to the load by opening the Active connection. When there is no brownout, the relay is energised and the supply is connected to the load. 38  Silicon Chip IC1 also connects to IC2 via its serial interface and these lines are labelled Data In, Data Out, Serial Clock and Chip Select (pins 20, 19, 18 & 17, respectively). In operation, IC2 uses these lines to program the registers within IC1 and to retrieve the monitored power data. Microcontroller IC2 also drives the LCD module using data lines RB7-RB4. These lines also connect respectively to switch S4 (direct) and to switches S3-S1 via diodes D3-D5. These diodes are necessary to prevent the data lines from being shorted together if more than one switch is pressed at the same time. In operation, IC2 can determine if a switch is closed (ie, pressed) by first setting its RB7-RB4 data lines high and then checking the RB3 input which connects to the commoned side of the switches. If none of the switches is pressed, the RB3 input will be held low via the associated 10kΩ resistor to ground. Conversely, if a switch is pressed, the RB3 input will be pulled high via that switch (and its associated diode, if present). The microcontroller then determines which switch is closed by setting all data lines low again and then setting each data line high (and then low again) in sequence. The closed switch is the one that produces a high at RB3. IC2’s RA2 & RA0 outputs (pins 1 & 17) control the register select (RS) and enable (EN-bar) inputs on the LCD module, to ensure that the data is correctly displayed. Trimpot VR1 adjusts the LCD’s contrast by setting the voltage applied to pin 3 of the module. A 4MHz crystal (X2) sets IC2’s clock frequency. This crystal determines the accuracy of the 0.1hr timer and the watt-hour calibration. However, frequency adjustment has not been included since the crystal’s untrimmed accuracy is better than the accuracy provided by IC1 for the wattage reading. Power supply Power for the circuit is derived from the mains via transformer T1. Its 12.6V AC secondary output is rectified using bridge rectifier BR1 and the resulting DC rail filtered using a 1000µF capacitor. This rail is then fed through rectifier diode D1, filtered using a 100µF capacitor and fed to 3-terminal regulator REG1. REG1 provides a stable +5V rail for IC1, IC2 and the LCD module. Note, however, that this +5V rail must also be regarded as being at mains potential (as must all other parts in this circuit, including the back-up battery). It might have a low DC voltage but it can also be sitting at 240VAC! Note also that we have specified a siliconchip.com.au low dropout regulator here and this has been done for two reasons. First, it allows the +5V rail to be maintained for as long as possible when the mains supply falls – important for maintaining the supply during a brownout. Second, this regulator was designed for automotive use and is capable of suppressing transient voltages of up to 60V at its input. This latter feature is useful for mains supply circuits, where there are likely to be transients during lightning storms. In addition, a Metal Oxide Varistor (MOV) connected between Active and Neutral at the mains input has been included to suppress transient voltages above the normal mains supply. The supply rail for relay RLY1 is derived from the output of the bridge rectifier (BR1). This rail is fed to the relay via a 68Ω 1W resistor, which reduces the voltage to about 12V. Diode D6 protects Mosfet Q1 from damage by quenching any back-EMF voltage spikes that are generated when RLY1 turns off. Back-up battery An optional 9V back-up battery has also been included in the power supply and this is connected to REG1’s input via diode D2. This back-up power is useful if the energy consumption of an appliance is to be measured over a long period of time (eg, weeks or months), since it maintains the active energy register values and allows the timer to continue counting if there is a blackout. You can use either a standard battery or a rechargeable nicad battery to provide back-up power. If a nicad battery is used, resistor (R3) is installed to provide trickle charging from the output of D1. Most applications will not require battery back-up, since you will just want to measure the energy consumption over a relatively short period. In this case, the accumulated energy reading will be lost when the mains power is switched off. However, all the settings (ie, the SAG parameters, offset and power calibration, cost per kWh and phase, etc) are retained when the mains power is off, as these are stored in a permanent memory. That’s all we have space for this month. Next month, we will give the complete construction and calibration SC details. siliconchip.com.au Parts List 1 PC board, code 04107041, 138 x 115mm 1 display PC board, code 04107042, 132 x 71mm 1 front panel label, 138 x 115mm 1 sealed ABS box with clear lid, 165 x 125 x 75mm (Altronics H0328 or equivalent) 1 12.6V 7VA mains transformer (Altronics M2853L) (T1) 1 12V SPDT 30A 250VAC relay (Altronics S4211) (RLY1) 1 LCD module (DSE Z 4170, Altronics Z 7000A, Jaycar QP 5515) 1 S20K 275VAC Metal Oxide Varistor (MOV) 1 3.58MHz crystal (X1) 1 4MHz crystal (X2) 1 18-pin DIL socket (for IC2) 1 M205 safety fuse holder (F1) (Jaycar SZ-2028 or equivalent) 1 M205 10A fast blow fuse 1 2-metre or 3-metre mains extension cord 2 cordgrip grommets for 6mm diameter cable 4 mains-rated pushbutton momentary-close switches (Jaycar SP 0702)(S1-S4) 1 4-way 0.1-inch pitch pin header 1 6-way 0.1-inch pitch pin header 1 4-way 0.1-inch header plug 1 6-way 0.1-inch header plug 4 stick-on rubber feet 1 9V battery (optional – see text) 1 connector plug & lead for 9V battery (optional, see text) 1 U-shaped bracket to suit 9V battery (optional, see text) 1 M3 x 6mm screw (optional) 1 M3 metal nut (optional) 6 M3 x 10mm Nylon countersunk screws 2 M2 x 9mm Nylon screws 4 M2 Nylon nuts 6 M3 x 12mm tapped Nylon spacers 7 M3 x 6mm screws 1 M3 x 12mm screw 5 M3 metal nuts 5 M3 star washers 1 14-way single in-line pin header (for Altronics and DSE LCD module); or 1 7-way dual in-line header (for Jaycar LCD Module) 1 3-way single in-line header 1 shorting plug for header 1 3mm diameter solder lug 3 6.4mm insulated spade connectors 2 2.8mm spade connectors 1 100mm length of 4-way rainbow cable 1 100mm length of 6-way rainbow cable 1 40mm length of 0.2mm enamelled copper wire 1 400mm length of 0.7mm tinned copper wire 1 150mm length of hookup wire 1 50mm length of 16mm diameter heatshrink tubing 1 50mm length of 2.5mm diameter heatshrink tubing 1 50mm length of 6mm diameter heatshrink tubing 5 50mm long cable ties 12 PC stakes Semiconductors 1 ADE7756AN Active Energy Metering IC (IC1) 1 PIC16F628A-20P programmed with wattmetr.hex (IC2) 1 LM2940CT-5 low dropout 5V regulator (REG1) 1 STP30NE06L logic Mosfet (Q1) 1 W04 1.2A bridge rectifier (BR1) 3 1N4004 1A diodes (D1,D2,D6) 3 1N914, 1N4148 diodes (D3-D5) Capacitors 1 1000µF 25V PC electrolytic 1 100µF 25V PC electrolytic 4 100µF 16V PC electrolytic 1 10µF 16V PC electrolytic 3 100nF MKT polyester 4 33nF MKT polyester 1 1nF MKT polyester 4 33pF NPO ceramic Resistors (0.25W 1%) 2 2.2MΩ 1W 400V 1 10kΩ 5 1kΩ 1 680Ω 0.5W (install only if backup battery is rechargeable) 1 68Ω 1W 1 10Ω 1 .01Ω 3W resistor (Welwyn OAR-3 0R01) (Farnell 3274718) (R1) 1 10kΩ horizontal trimpot (code 103) (VR1) July 2004  39