Ac Power Logger Using Mcp39009 Engineering Essay

Ac Power Logger Using Mcp39009 Engineering Essay In this project we are going to record power consumption using MCP3909 by using AVR 5A microcontroller. This MCP 3909 is an energy metering IC with SPI interface and active power pulse output. Where the MCP3909 used in two different phases that can be operated at a time like 1.Output through active pulse power and 2. Waveform obtained as output through SPI interface. For the output real pulse power, the device gives frequency output proportional to instantaneous power. For the waveform output, it gathers data from the current and voltage channel and both are 16 bit second order delta sigma ADC . Through out this project time I learnt to do coding in C programming to get pulse output that shows the consumption of power. This program is developed to use with microcontroller Easy AVR 5A type of AT Mega 16. This program is executed by using AVR Studio by running it through AVR Flash and the output pulse wave form is gathered at AT Mega 16 board and we can check using oscilloscope. Here to have connection between AVR board AT Mega 16 and CPU we use USB which acts like supply and we use USART to transfer data between each other. As in my project I have MCP3909 as separate board I should give correct connection between MCP 3909 and AT Mega 16 correct supply connections. As I am doing project in embedded microcontroller with C programming and output I get is pulse waveform so for better performance and other reasons I took Easy AVR AT Mega 16 development board with 8 KHz frequency. My project is AC power logger using MCP3909 where MCP 3909 is an energy metering IC with SPI interface and active power pulse output. Where the MCP3909 used in two different methods where they can be operated at a time like 1.Output pulse power and 2. Waveform we get as output through SPI interface. For the active real power pulse output, the device gives output frequency which is proportional to instantaneous power. For the waveform output, it gathers data from the current and voltage channel and both are 16 bit second order delta sigma ADC which over samples input signal at frequency equal to MCLK/4 and allows large range of input signals. For channel 0, the increase in the current at channel 0 is done through programmable gain amplifier increase.. As I get pulse count at output when we use oscilloscope to get exact count and good performance I use Easy AVR 5A AT Mega 16 microcontroller. This AT Mega 16 microcontroller allows assembly language and C language programming but I did coding for counting the power consumption using MCP3909 in C language. In this first I gave USB and USART connection between CPU and AT Mega 16 board, here I gave MCP3909 connection with AT Mega 16 and gave current and voltage channel input to MC3909 and I get pulse output this is with out SPI pin not in use but still they should be connected. But for SPI connection program I will give voltage and current with 50hz frequency and take output voltage and current values from the hyper terminal connection and calculate the output values. In this whole project time I began to understand the SPI and USART connection to write program in C language and also understood how the MCP3909 works . The energy meter is a device which is used for electrical measuring, it is used to record electrical energy consumed in specific period of time in terms units Every house, small factory, business establishments, shops, offices etc need at minimum one energy meter to register power consumption. The one who supplies electricity raises bills based on readings shown in energy meter. The one who produces electricity sale the electricity to the electricity boards and board will sale this to costumer. So the data generated by the energy meter is the base to raise bill by the power supplier. This energy meter products are available in single and three phases at different current ratings as per customer requirements. This energy meters are basically electrical and mechanical components. The design of energy meter depends upon which rating of voltage and current meter has to work. In this project the energy metering IC that we take is MCP3909 which is used for supporting IEC 62053 which is standard international meter. It gets the output frequency which is proportional to real power as input so as to access the ADC channel and output of multiplier data. The delta sigma which is 16bit ADC is used to allow large range of currents using the design. The exact or appropriate energy IC is available in the industry which is highly reliable and which has 24-lead SSOP output pin. In this functional as we can see it shows ch0 and ch1 channels where these are inputs and given from the current and voltage transformers .The gains g0 and g1 are also given as input and the dual functionality pin is connected for SPI connections and also for f0, f1,f2. The outputs we take from the active power DTF conversion as HFout and other outputs Fout0 and Fout1 are obtained from stepper motor output drive for active power. The above one is the general block diagram of MCP3909, but the diagram that I use here shows the diagram with the connections , this shows how the MCP 3909 is internally connected to oscillators ,jumpers and for the output using its 24pins. It shows that for the jumper j7 it connects internally to fout 0/1, hfout. For the jumper j2 it shows the connections to spi communications connected and the jumper j6 to ch0 and ch1 channels. The digital voltage acts as digital circurity in MCP3909 where it is the one where we get digital power supply. This pin requires appropriate by pass capacitors and should be maintained to 5V. In both the input channel levels this pin acts as HPF and where it controls the flow of signals. The logic 1 will activate both the filters for removing the DC offset from the system and this logic 0 will disable both the filters so due to this they allow DC voltage. This is analog circuit pin which is used to give analog power supply with in MCP 3909 an this pin requires exact bypass capacitor which gives ramp signal with rising and falling edges and it must be maintained at 5V. This pins are used for current measurements and where they initially take analog voltage as input and convert to current and this will have PGA for small input signal. The linear and the region where it characteristics of this channel are dependent on PGA gain. It relates to maximum voltage of 470Mv/G and the voltage range changes from 1 to 6 V with respect to Agnd. . This pins are used for voltage measurement and this pins initially take difference analog voltage input. The linear and specific behaviour of this voltage channel is maximum at 660mV with absolute voltage 1V Here for the internal 2.4V reference the output is the reference in/out and with temperature coefficient of 15ppm/c. Here by applying the voltage to this pin from the specified range we can use external reference and these reference in/out pin uses bypass capacitor to AGND even when using internal reference.. This is the analog ground where all the ADC,PGA,POR and band gap reference are connected to ground and this is analog circuit. To have noise signal to be cancelled this pin should be connected to same ground as Dgnd with star connection. This is the normal ground connection where SINC filters, multipliers, HPF, LPF, digital to frequency convertor and oscillator; this is used as internal circuit connection. To have accurate and noise to be cancelled this digital ground should be grounded same as analog ground with star connection. The output pins that are connected to MCP3909 are frequency outputs that give us real power and the signal that we get when connected to oscilloscope is pulse where this pulse period is directly proportional to power and Fc constant. This pins helps us to activate the electro mechanical counters and also two phase stepper motor. The high frequency output supplies instantaneous real power information and out put is periodic pulse and where it is directly proportional to measured l power and HFC constant obtained by F0,F1,F2 logic gates and the output that obtained is the fastest output frequency. These oscillators will provide sine waveform with clock source and these oscillators are mainly used to give clock signal for master in the device. The clock frequency is given as 3.57MHz and this clock frequency value should range from 1to 4 MHz with out any error. In this to convert the signal from digital domain which has wide dynamic range we use PGA to do this function which is common thing done in wireless communication. To load normally input signal from analog to digital we need to increase the amplitude so to do this we use PGA. High resolution sigma to delta ADCs all have Programmable Gain Amplifier at input to the sigma to delta modulator is given as shown below The PGA on the ADs chip offer eight input ranges to ADC with 2.5 voltage reference, the eight reference voltages are 2.56V,1.28V,640mv,160mv,40mv.if reference voltage is doubled to 5v then full scale input for each range is halved. So the actual signal range for any PGA settings are given as, VREF*1.024/2(7-RN) where RN value is 111 when ref voltage is 2.5v The main use of PGA is that the noise in terms of micro voltage decreases when the gain increased. In effect the input signal is gained up but the noise is not gained up, so there is an improvement in signal to noise ratio. The pins reference sampling rate is given at 524 khz and capacitor value is fixed so there is no variation in reference current and any gain error that has due to resistance on reference input is also fixed. If reference current changes on sigma to delta the gain error that occurs also varies and the benefit of using ADC will be lost. The PGA for the ADC offers benefits of high resolution and low noise at high gain , but without the disadvantages of requiring regular calibration every time the range is changed. A buffered input and new reference sampling scheme avoid many of the problems associated with previous multirange ADC. All the delta sigma ADCs , registers, filters, multipliers are controlled by reset of master clear and this pin is also used to change their serial interface and behaviour or functionality. The logic 0 controls the ADC and registers in reset condition. The only one that uses power during master clear is oscillator. The microcontroller manufactures produce other design products so that they are related to their own design and in this we require another output pin. This condition or situation is correct for small design products where they have eight pins are fewer than that. This microcontroller has two output pins, one input pin , RAM flash and ADC module .For programming the microcontroller mode you need MCLR and supply pins(VCC and GND). To run or make the coding active we need mainly power supply and MCLR,microcontroller must see the difference between normal and program mode. Here the MCLR takes 12V to enter program mode and it takes external reset or input pin to enter into normal mode. The microcontroller design uses one pin for analog input and has other three outputs and it also requires an additional output, so for this reason the circuit uses MCLR pin as output. To make MCLR pin to act as output the microcontroller uses weak pull-ups. Analog to Digital Conversion (ADC) is the process of sampling continuous analog signal and converting the signal into quantitized representation of signal in digital domain and all the ADC architectures will convert analog signal into digital representation. The conventional ADC process takes input signal x(t) into sequence of digital codes x(n) at sampling rate fs=1/t, where T represents sampling interval this sampling function is equivalent to modulating input signal by set of carrier signals with frequencies 0,fs,2fs †¦. The sample signal is expressed as summation of original signal component and frequency, here the frequency modulated by integer multiple of sampling frequency. The signal component at frequency in input signal cannot be properly sampled and such signals get folded in base band signal creating in this non-linear is referred to as aliasing . Anti-aliasing filters are therefore required to prevent aliasing. Many A/D converters have successive or continuous approximation register and flash converters operate at nyquist rate fn. These converters sample analog signal at sample frequency equal to twice maximum frequency of input signal. Sigma Delta AD converters do not digitize the incoming analog signal into digital sample of n-bit precision at nyquist rate, sigma delta ADC samples the analog signal by an sample ratio N resulting Fn One of the advantages of sigma-delta ADC over nyquist ADC is the relaxation of the requirements for the anti aliasing filter. The requirement of anti-aliasing filter for nyquist rate ADC require sharp pass band (fs) to stop band (fn) The sigma delta ADC contains simple analog circuits like voltage reference, comparator, integrator,summing circuit and switch and in this the digital circuit consists of digital signal processing which acts as filter. Now consider technique of oversampling in frequency domain when converting to dc signal it has quantization error up to  ½ LSB and this sampled data has quantization noise. If ADC is less than perfect or exact value its noise is greater than quantization noise so due to this its resolution will be small than n bits and its actual resolution is given by The sampling rate is chosen as Kfs then quantization noise is q/sqrt12 due to this noise will spread at bandwidth dc to Kfs/2 , So to reduce noise we use digital low pass filter at output with out disturbing the wanted signal. K is referred as sampling ratio and this sampling relaxes requirement on the analog antialiasing filter. Here the data rate is less than the sampling rate and to satisfy nyquist criteria and this is done by using low pass filter to reduce the bandwidth, this process can be done by giving Mth result to output with neglecting the remainder and this process is known as decimation by factor M. This M can have any value such that output data rate is greater than twice the bandwidth. If we use oversampling to improve resolution then the oversampling must be factor of 22N to get N bit resolution increase, the sigma delta converters does not require any large oversampling because it limits to pass band signal and shapes the quantization noise to fall outside the pass band as shown in figure. Here we have 1-bit comparator (ADC) when we use it integrator output, then sum the input voltage with output of 1-bit DAC which we get from ADC output.. The digital low pass filter and decimator at digital output are added to get sigma delta ADC and after this signal is given to modulator where it modifies quantization noise by making it to lie above pass band filter ,so due to this the ENOB is larger than the expected sampling ratio. The sigma delta ADC operation is like the input given as Vin which is dc and the integrator consistently move up and down at node A and here output of comparator is given to 1-bit DAC and summing point at node B. This negative feedback value will force the average dc voltage at node B to be equal to Vin. The output voltage from the DAC is controlled in the 1-bit data stream of the comparator output. After that when the input signal rises at Vref, the number of ones at the serial bit stream also increase and due to this there is decrease in zeroes and in the same way as the signal of Vref goes negative the serial bit stream at one decrease and at zero it increases. Here it shows that average value at voltage as input is in serial bit stream which comes from comparator and decimator and filter allow stream and give output. The data from the 1-bit ADC is not worth full when the given input value is single sample interval, so when we have more number of samples that are averaged will provide correct value. The sigma delta can not give detailed values in the time domain because of the single bit data output, so when the single input is near positive side it shows more ones than zero and in the same way when the input signal is near to negative it shows more number of zeroes than ones and if it is in midscale then it shows equal number of zeroes and ones. The below figure shows the output of integrator for two conditions where the first one is for input zero near the midscale so decode them pass output samples through low pass filter that averages every four samples this shows the bipolar zero. So from this we can say that if more number of samples are averaged more dynamic range is obtained. The sigma delta ADC can also be seen as synchronous voltage to frequency converter with the counter. If the number of ones in the output data stream is counted from the samples then the counter output will give digital value of output, this method applies only when have dc or for slow changing input signal. The 2N clock cycles are counted to achieve N-bit resolution and there by for getting effective sampling rate. Here noise shaping is explained in frequency domain by using sigma delta modulator. In this the integrator which is present in the modulator represents an analog low pass filter with transfer function H(f)=1/f and this transfer function shows that the amplitude not directly proportional to frequency. The one bit quantize gives quantization noise Q and it is given to output sum block. If we have input signal X and output signal Y the value that comes out of summing point is X-Y and after that it is multiplied by the transfer function and this is given as, From the equation if we see that if f=0 the output Y reaches X with no noise , and at high frequency the amplitude of the signal reaches zero and noise value reaches Q. So due to this the analog filter has signal effect on low pass and high pass effect on noise Q. This filter does noise shaping at given frequency in delta sigma model and higher order filter gives more attenuation in sigma delta modulators but some precautions should be taken. We get good quantization noise and best ENOB for given sample when we have more integrator and summing points in sigma delta modulator. This figure is giving the relationship between order of sigma delta modulator and oversampling amount to reach SNR. If oversampling is taken 60 then the second order capable of giving SNR of 80db and also gives ENOB value as 13 and in this we have filter to reduce noise and decimator to decide degree. This carries 13 bit outside but if you want to use additional bits, these added bits that carry signals has no useful value and buried in quantization noise unless the post filtering is used. The resolution can be increased from the 1-bit system by increasing the oversampling ratio or by higher order modulator. In the other method for the waveform output we give current and voltage as input which are 16bit and then given to second order sigma delta ADC where it oversamples input at frequency equal to MCLK/4 and with this it allows for wide range of input signals. The input current channel (channel 0) usable range is increased with the programmable gain amplifier and this is linked with block diagram of MCP3909 and gives in detail of its signal processing blocks. To cancel the system offset on both the channels we use to high pass filter and from output of filter we get voltage and current, so when calculating power we should not get any offset. As this signals are not having DC offset so the averaging technique is used to give active power output. The power signal at we get after filtering is active power output it is DC component and for averaging technique use sine and non-sine waveform after this the ADC takes real power to give output pulse where the frequency is directly proportional to real power. The frequency present at FOUT 0, FOUT1 outputs are used to drive counters and stepper motor which shows power consumed. Every pulse from F0, F1, F2 settings are used to give fixed amount of energy , the HFOUT has less integration and high frequency to represent power signal and due to less time it helps the user to get values fastly under steady condition. . For the current and voltage transducers the MCP3909 analog inputs are connected and each pin has specifications like it should pass from 5kV to 500V contact charge. The differential input is given for both the channels to reduce noise and absolute voltage should be kept at 1V related to AGND so this can do error measurement. The common mode signal is taken to respect both last condition and input voltage difference range and for good common mode ration to should be referred to ground. The current channel has PGA gain to measure small signal with out other signal. The maximum differential voltage we have at channel0 is 470mV/Gain. The maximum voltage fro channel1 is 660mV. For channel 0 gain selection is given as, This MCP3909 has internally POR to check supply voltage AVdd and this check when the systems power is on or off. This POR has built in hysteresis and timer to check potential ripple and noise on power supply. For this the threshold voltage is typically set to 4V. The MCP3909 is kept in reset state if the supply voltage falls less than threshold voltage and hysteresis value is 200mV to prevent glitches. Once the power is on the internal timer stop sending the pulse with MCLK=3.58MHz there by preventing potential metastability. For calculating the active power the MCP3909 use digital filter which is first order IIR filter where we can extract real power (DC component) from the power signal. Since the input power signal has harmonic content. We get ripples from the filter output at line of frequency when the filter is not ideal. To reduce the noise for line frequency at 50Hz we use cut off frequency as input clock (MCLK=3.58 MHz). The rejection of frequency component will be more than 20db. In this at the frequency converter the output of filter is stored and then it is helpful to compare threshold for Fout0/1 and HFout and each time threshold is crossed we get pulse. The Fout0/1 require more energy to get output pulse than HFout , like integration period and as this acts as filter the output ripple or noise is minimum. The threshold or transfer function of HFout and Fout0/1 are different to each other. The threshold energy or transfer function are different to each other , the Fout0/1 output frequencies are quite low in order to allow integration. In this synchronous serial transmission clock is shared between sender and receiver or the sender gives timing signal so that the receiver knows when to read next bit of data. In this serial transmission if we do not have data to send then fill character is sent instead of data so to keep transmission continually, these synchronous communication is efficient because in these we have only data transmission between sender and receiver.. for example the synchronous transmission is used between printer and fixed device where data is sent in one set of wire and clock is sent in different wire. This RS 232 is asynchronous serial communication method which is used for computers and others, it is called as asynchronous because there is no synchronizing clock present like which is in SPI where it is serial protocol, the serial protocol is such that it automatically synchronize itself. We can use RS 232 to easily create data link between boards and standard PC, you can makes data loggers that read analog value from ADC and give it to PC this is done by writing program that shows data with using graphs. In serial communication the byte is sent or transmitted one bit at time but in parallel communication the whole data like byte (8 bit) transmitted at a time. So for that we use parallel communication to send data in shorten distance like between graphic card and CPU and these parallel can have say many wires as possible , but serial communication uses one wire to transfer data so it is used for long distance. In series the logic level changes with the bit being transmitted (0 or 1) and to know which is start bit and end bit in byte we need to add synchronize line and note the value of data line when the clock line is high but this is the way the serial buses like SPI work . UART is not having clock because it is asynchronous but start bit and stop bit are used to synchronize the incoming data. When the word is in transmission start bit is added at start of each word and this tells the receiver that data is read to sent and forces the receiver clock to be synchronous with clock of transmitter. These two must not have same frequency drift but can have same clock. After the start bit is sent each bit in word are given least significant bit (LSB) and each bit from transmitter is sent with same time and receiver is in half way to check that bit is one or zero. The sender will not know when receiver looks at the bits but sender knows when the clock says to begin to send next bit of word. When the complete data word is sent the transmitter adds parity bit and at the receiver uses this parity bit for error checking and at last the one stop bit is sent by the transmitter, if the receiver does not receive the stop bit when it is supposed to be the UART thinks the entire word to be garbled and reports framing error to host when data word is read. This framing error occurs because the sender and receiver clocks are not running at same speed. Whether the data is sent or not the UART automatically discard start, stop and parity bit and if another word is coming the start bit for new word comes as soon as the stop bit for existing word been sent RS 232 In this it has two data line like RX and TX , where TX is the wire where data is sent out to other device and RX is the line in which other device put data it needs to send. We know that high = 5v and low =0v for MCU boards and this RS 232 has high=12v and low=-12v. So to make RS 232 to interface with MCU which understands 0 to 5 volts we use MAX232. As RS232 has no clock line for synchronization perfect timing is needed so transmissions are carried out in certain speed which is bits per second and number of bits transmitted per second is know as baud rate. Some standard rates are 1200, 2400, 4800, 9600, 19200, etc. RS 232 Level conversion As seen above the RS232 signals differ from signals in MCU, this level converter will convert RS 232 signals from -12 to 12 volts from PC to signal 0 to 5 volts to fed to MCU. It is good to check the operation so we use converter to see its working nature, so for this we need Hyper-terminal windows software which is used to open COM port and to send and receive textual data. For testing we need to connect output RX/TX together so data written to COM ports to enter our circuit and converted to MCU board signal level. After this understand the USART of AVR Microcontroller and write code to activate USART to send and receive data, like other microcontrollers AVR also has main hardware for serial communication this is called USART. In this USART hardware you need to write data to one of registers. Clock generation. This generator generates the base clock for transmitter and receiver, this USART supports four modes of clock operation 1. Normal Asynchronous, 2. Double speed asynchronous, 3. Master synchronous and 4.slave synchronous mode. The UMSEL bit in the UCSRC (control and status register) is the one that selects between synchronous and asynchronous operation. Double speed(asynchronous mode) is controlled by U2X found in UCSRA register. When UMSEL=1 the data direction register for the XCK controls weather the clock source is internal (master mode) or external (slave mode) and this is shown in block diagram Txclk- transmitter clock(internal signal) Rxclk receiver clock(internal signal) Xcki used for synchronous slave operation Xcko used for synchronous master operation Fcso- system clock Baud rate generator The USART Baud rate register and down counter are connected as programmable prescaler or baud rate generator. The down counter which is running at the system clock(fosc) is loaded with UBRR value each time the counter has counted down to zero and clock is generated each time counter reaches zero and the clock generated is the baud rate generator clock output = fosc/(UBRR+1). The transmitter divides the baud rate generator clock output by 2,8.16 depending on the mode and this baud rate generator is directly used by receiver clock and data recovery units. The baud rate generator equations are given as, Operating mode Calculating baud rate Calculating UBRRvalue Asynchronous normal mode Baud= fosc/(UBRR+1)16 UBRR= fosc/16baud   1 Asynchronous double speed mode Baud= fosc/(UBRR+1)8 UBRR= fosc/8baud   1 Synchronous master mode Baud= fosc/(UBRR+1)2 UBRR= fosc/2baud   1 External clock The synchronous mode operation is done by using external clock and external clock input from XCK is sampled by synchronous register to reduce change in stability and the output from synchronous register must pass through edge detector before it is used by transmitter and receiver. This process includes two CPU clock period delay and its frequency is given as FXCK USART of AVR The USART of AVR is connected to CPU by these six registers UDR- USART Data Register: basically this is not one but two register , when you read it data is stored in receiver buffer and when you write it gives to transmitter buffer. UCSRA: USART Control and Status Register: as it name says it stores some status about USART and there are some of this kind like UCSRB and UCSRC. UBRRH and UBRRL: This is USART baud rate register, it i s16 bit wide so UBRRH is high byte and UBRRL is low byte . To write programs with using USART you need to study about each register, the seen behind using USART is same with other internal peripheral. now we will describe each registers clearly This bit is set when USART completed receiving byte from host and program should read from UDR and this flag bit is set when unread data is present in receiver buffer and gets cleared when receiver buffer is empty. If the receiver is disabled, the receiver buffer is flushed and the RXC will completely zero. Bit 6- TXC: transmit complete This bit is set 1 when USART has completed transmitting byte to host and program can write new data to USART through UDR. The transmit flag bit is cleared automatically when TXC interrupt is executed. Bit 5 UDRE USART Data Register Empty The UDRE flag first tells us that the transmit buffer (UDR) is ready to