Highland Technology AT180 and V180 Electrical Power Measurement Modules

AT/V180 Electrical Measurement Modules

The AT180 and V180 electrical measurement modules are intended to provide accurate acquisition of AC voltage, current, power, and energy levels. The application of modern sampled-data Digital Signal Processing (DSP) technologies to AC measurement requires careful analysis and design to ensure that the reported data is accurate and consistent with the measurements of classical electrodynamic instruments. The Series 180 modules use proprietary algorithms similar to those used in the Highland-designed C180 electrical survey meter/recorders. Hundreds of C180 meters have been used in utility company end-use load study installations where their data is "checksummed" against the standard electromechanical service meter with typical differences of less than 1%. This document explains the basic measurement techniques used in the Series 180 instruments.

BASICS

For each of its sixteen "virtual wattmeter" channels, an AT180/V180 receives two analog signal inputs from external sensors. One signal is a divided-down differential voltage signal which is proportional to the instantaneous voltage difference between the highside and lowside potential inputs of the "wattmeter," and the other is a low-voltage signal proportional to the instantaneous current flow in the associated current-carrying conductor. Using these two signals, the unit will compute the true RMS voltage, true RMS current, true power (watts), apparent power (volt-amps), integrated energy (kilowatt-hours), and apparent energy (KVA-hours) of each channel.

In brief, the signal processing sequence is as follows:

1. Both the voltage and current signals are periodically (and simultaneously) sampled and digitized, yielding numeric instantaneous voltage and current sample values.

2. The voltage samples are squared and lowpass filtered to yield a smoothed, mean-square voltage value. The square root of this filtered value is periodically computed and presented as the realtime RMS voltage value.

3. The current samples are squared and lowpass filtered to yield a smoothed, mean-square current value. The square root of this filtered value is periodically computed and presented as the realtime RMS current
value.

4. The instantaneous voltage and current samples are independently multiplied to yield a signed instantaneous power product. The power-product samples are lowpass filtered, and the output of this filter is periodically posted as the realtime true power value.

5. The realtime power value from the "power" lowpass filter is time-integrated into a kilowatt-hour (energy) register.

6. The realtime RMS voltage and current values (the output signals of steps 2 and 3 above) are multiplied and presented as the realtime apparent-power (KVA) value.

7. The realtime KVA value is time-integrated into the KVAH register.

This algorithm is accurate without regard to the signal frequency, phase angles, or waveforms.

SAMPLING AND FILTERING

The algorithm described is, on the surface, precisely identical to the processing used by continuous (analog) signal processors to deliver the same results and, in fact, is identical to the "signal processing" used by classic electrodynamic instruments. The only difference is that the input waveforms are sampled and digitally processed, rather than being continuously processed. The V180/AT180 sampling rate was chosen to ensure accurate measurement of waveforms containing frequency components up to the 50th line harmonic, up to an AC line frequency of over 100 Hz. Advanced signal processing and filtering techniques are used in the Series 180 products to deliver accurate measurements for the ranges of signals found in real AC power systems.

Although Series 180 units are normally used to measure 50 or 60 Hz AC circuits, the units are capable of accurate measurement of power circuits from DC to more than 100 Hz, if suitable current transducers are provided.

All acquired data is lowpass filtered to smooth instantaneous samples into averaged RMS volts, amps, watts, and KVAs. The digital filters used in the AT180/V180 have a normal bandwidth of 1 Hertz. The filters have a "transitional" transfer function, a compromise between good high-frequency noise rejection and clean, damped step response. For a step change of voltage or current, the realtime RMS values reported by the AT180/V180 will transition (20-80% levels) in about 400 milliseconds, and will settle cleanly without ringing or overshoot; this behavior is well suited for both measurement and closed-loop control applications, and has response similar to the dynamics of mechanical meter movements. In the selectable "fast filter" mode, filter bandwidth is increased to 4 Hz, with 0.1 second step response. Fast filter mode is useful for recording envelope transients (motor starts, for example) and for use in fast control loops, but will inherently display "noisier" data in a system connected to a real-life AC line and load.

Filtered volts, amps, watts, and other measured parameters are updated to the computer bus (ISA or VME busses for the AT180 and V180 respectively) at a rate of 32 times per second.

One advantage of digital processing is the wide dynamic range of power measurement it allows. Whereas electromechanical instruments are limited by friction, and analog-multiplier-based electronic wattmeters have inherent offset and drift errors, a digital wattmeter can use autozero and correlation techniques to eliminate offsets, and can use statistical techniques to improve ADC linearity and remove digital quantization effects. The Series 180 products demonstrate power measurement linearity of better than 1 part in 5000, and have virtually unmeasurable "zero power" offsets--below one part in 100,000. This means that a Series 180 instrument can accurately measure power levels that are a small fraction of the full-scale power of the instrumented circuit. The AT180/V180 achieves this wide dynamic range without scale or gain switching, thus avoiding the bizarre behavior that some electronic power instruments display when they are presented with low- and mid-range signal levels.

SOURCES OF ERROR

Several sources of error should be considered in applying the V180 and AT180 products. They include the following.

1. CURRENT SENSOR ERRORS. The standard current sensors provided with the Series 180 products are typically accurate to about 0.5% (amplitude) and 0.5 degrees (phase), so these sensors are the dominant error contributor in a typical system. These sensors, like most current transformers, show increased phase shift at low currents (say, below 10% of full scale). If higher accuracy is needed, special high-precision current sensors are available. Metering grade, 5-amp secondary CTs are available with ANSI ratings to 0.1% accuracy, and these may be used with precision 0.05 ohm burden resistors as current sensors for the Series 180 instruments.

2. STRAY PICKUP. Since Series 180 current inputs are usually 333.3 mV AC for full-scale current, low current levels correspond to very low signal levels. A current of 1% of full scale corresponds to only 3.333 millivolts, and a 1% error of this voltage is only 33.3 microvolts. The parts-per-million zero offset accuracy of the Series 180 modules can be degraded by microvolt-level hum pickups. To avoid microvolt-level errors, the following precautions are suggested:

A. Use torroidal (as opposed to removable link) current transformers. Removable-link CTs can pick up signals from current-carrying conductors which are outside the sense loop. If split-core sensors are
used, keep them away from other current-carrying lines or transformers.

B. Ensure that current sensor leads are tightly and uniformly twisted, and that the signal leads are not run near or parallel to current-carrying conductors or near transformers.

C. To avoid ground loops, do not ground CT leads anywhere except at the C760 termination panel.

D. Ensure that the module mainframe (PC/AT computer or VME crate) is properly grounded, and that the C760 termination panel is securely grounded to the same ground system.

E. Route the D37 cable (from the termination panel to the module) away from current-carrying conductors.

3. SENSOR OVERLOAD. Measurement errors can result from overload of current signal inputs. If a current sensor is rated at "N" amps, the instantaneous current level through the sensor should not exceed 2(N) amps. Since very distorted waveforms may have high peak-to-average ratios, it may be advisable to use a higher-rated CT to measure currents that may have high peak values.

4. CIRCULATING DC. Current transformers may lose accuracy if the measured current contains a significant DC component, such as that which might be created by a half-wave rectified load or
asymmetric SCR circuit. If DC is suspected, it should be measured with a DC ammeter, and CT specifications checked for compatibility. Note that a DC current component contributes no real power so long as the line voltage remains pure AC.





	AT/V180 Electrical Measurement Modules The AT180 and V180 electrical measurement modules are intended to provide accurate acquisition of AC voltage, current, power, and energy levels. The application of modern sampled-data Digital Signal Processing (DSP) technologies to AC measurement requires careful analysis and design to ensure that the reported data is accurate and consistent with the measurements of classical electrodynamic instruments. The Series 180 modules use proprietary algorithms similar to those used in the Highland-designed C180 electrical survey meter/recorders. Hundreds of C180 meters have been used in utility company end-use load study installations where their data is "checksummed" against the standard electromechanical service meter with typical differences of less than 1%. This document explains the basic measurement techniques used in the Series 180 instruments. BASICS For each of its sixteen "virtual wattmeter" channels, an AT180/V180 receives two analog signal inputs from external sensors. One signal is a divided-down differential voltage signal which is proportional to the instantaneous voltage difference between the highside and lowside potential inputs of the "wattmeter," and the other is a low-voltage signal proportional to the instantaneous current flow in the associated current-carrying conductor. Using these two signals, the unit will compute the true RMS voltage, true RMS current, true power (watts), apparent power (volt-amps), integrated energy (kilowatt-hours), and apparent energy (KVA-hours) of each channel. In brief, the signal processing sequence is as follows: 1. Both the voltage and current signals are periodically (and simultaneously) sampled and digitized, yielding numeric instantaneous voltage and current sample values. 2. The voltage samples are squared and lowpass filtered to yield a smoothed, mean-square voltage value. The square root of this filtered value is periodically computed and presented as the realtime RMS voltage value. 3. The current samples are squared and lowpass filtered to yield a smoothed, mean-square current value. The square root of this filtered value is periodically computed and presented as the realtime RMS current value. 4. The instantaneous voltage and current samples are independently multiplied to yield a signed instantaneous power product. The power-product samples are lowpass filtered, and the output of this filter is periodically posted as the realtime true power value. 5. The realtime power value from the "power" lowpass filter is time-integrated into a kilowatt-hour (energy) register. 6. The realtime RMS voltage and current values (the output signals of steps 2 and 3 above) are multiplied and presented as the realtime apparent-power (KVA) value. 7. The realtime KVA value is time-integrated into the KVAH register. This algorithm is accurate without regard to the signal frequency, phase angles, or waveforms. SAMPLING AND FILTERING The algorithm described is, on the surface, precisely identical to the processing used by continuous (analog) signal processors to deliver the same results and, in fact, is identical to the "signal processing" used by classic electrodynamic instruments. The only difference is that the input waveforms are sampled and digitally processed, rather than being continuously processed. The V180/AT180 sampling rate was chosen to ensure accurate measurement of waveforms containing frequency components up to the 50th line harmonic, up to an AC line frequency of over 100 Hz. Advanced signal processing and filtering techniques are used in the Series 180 products to deliver accurate measurements for the ranges of signals found in real AC power systems. Although Series 180 units are normally used to measure 50 or 60 Hz AC circuits, the units are capable of accurate measurement of power circuits from DC to more than 100 Hz, if suitable current transducers are provided. All acquired data is lowpass filtered to smooth instantaneous samples into averaged RMS volts, amps, watts, and KVAs. The digital filters used in the AT180/V180 have a normal bandwidth of 1 Hertz. The filters have a "transitional" transfer function, a compromise between good high-frequency noise rejection and clean, damped step response. For a step change of voltage or current, the realtime RMS values reported by the AT180/V180 will transition (20-80% levels) in about 400 milliseconds, and will settle cleanly without ringing or overshoot; this behavior is well suited for both measurement and closed-loop control applications, and has response similar to the dynamics of mechanical meter movements. In the selectable "fast filter" mode, filter bandwidth is increased to 4 Hz, with 0.1 second step response. Fast filter mode is useful for recording envelope transients (motor starts, for example) and for use in fast control loops, but will inherently display "noisier" data in a system connected to a real-life AC line and load. Filtered volts, amps, watts, and other measured parameters are updated to the computer bus (ISA or VME busses for the AT180 and V180 respectively) at a rate of 32 times per second. One advantage of digital processing is the wide dynamic range of power measurement it allows. Whereas electromechanical instruments are limited by friction, and analog-multiplier-based electronic wattmeters have inherent offset and drift errors, a digital wattmeter can use autozero and correlation techniques to eliminate offsets, and can use statistical techniques to improve ADC linearity and remove digital quantization effects. The Series 180 products demonstrate power measurement linearity of better than 1 part in 5000, and have virtually unmeasurable "zero power" offsets--below one part in 100,000. This means that a Series 180 instrument can accurately measure power levels that are a small fraction of the full-scale power of the instrumented circuit. The AT180/V180 achieves this wide dynamic range without scale or gain switching, thus avoiding the bizarre behavior that some electronic power instruments display when they are presented with low- and mid-range signal levels. SOURCES OF ERROR Several sources of error should be considered in applying the V180 and AT180 products. They include the following. 1. CURRENT SENSOR ERRORS. The standard current sensors provided with the Series 180 products are typically accurate to about 0.5% (amplitude) and 0.5 degrees (phase), so these sensors are the dominant error contributor in a typical system. These sensors, like most current transformers, show increased phase shift at low currents (say, below 10% of full scale). If higher accuracy is needed, special high-precision current sensors are available. Metering grade, 5-amp secondary CTs are available with ANSI ratings to 0.1% accuracy, and these may be used with precision 0.05 ohm burden resistors as current sensors for the Series 180 instruments. 2. STRAY PICKUP. Since Series 180 current inputs are usually 333.3 mV AC for full-scale current, low current levels correspond to very low signal levels. A current of 1% of full scale corresponds to only 3.333 millivolts, and a 1% error of this voltage is only 33.3 microvolts. The parts-per-million zero offset accuracy of the Series 180 modules can be degraded by microvolt-level hum pickups. To avoid microvolt-level errors, the following precautions are suggested: A. Use torroidal (as opposed to removable link) current transformers. Removable-link CTs can pick up signals from current-carrying conductors which are outside the sense loop. If split-core sensors are used, keep them away from other current-carrying lines or transformers. B. Ensure that current sensor leads are tightly and uniformly twisted, and that the signal leads are not run near or parallel to current-carrying conductors or near transformers. C. To avoid ground loops, do not ground CT leads anywhere except at the C760 termination panel. D. Ensure that the module mainframe (PC/AT computer or VME crate) is properly grounded, and that the C760 termination panel is securely grounded to the same ground system. E. Route the D37 cable (from the termination panel to the module) away from current-carrying conductors. 3. SENSOR OVERLOAD. Measurement errors can result from overload of current signal inputs. If a current sensor is rated at "N" amps, the instantaneous current level through the sensor should not exceed 2(N) amps. Since very distorted waveforms may have high peak-to-average ratios, it may be advisable to use a higher-rated CT to measure currents that may have high peak values. 4. CIRCULATING DC. Current transformers may lose accuracy if the measured current contains a significant DC component, such as that which might be created by a half-wave rectified load or asymmetric SCR circuit. If DC is suspected, it should be measured with a DC ammeter, and CT specifications checked for compatibility. Note that a DC current component contributes no real power so long as the line voltage remains pure AC.