Recommendation Report Measuring Gross
Contents
1. A large percent error coupled with a high price tag of 3000 means that this oscilloscope would not be the best choice for this application because of its poor cost efficiency The Agilent 34401A digital multimeter is small cheap and precise Despite the Agilent DMM s low price of 900 accuracy is the most important criterion The DMM s extremely poor accuracy means it is also not suitable for use in this application The ELVIS was able to achieve the highest accuracy with a 3 27 error for the mean and 1 26 error for RMS It is also cheaper than the Tektronix oscilloscope by 1000 Because the ELVIS would replace the existing breadboard the engineers at Burdell Electronics use and because they already have a PC at their workbenches only a small amount of additional space would be used Investing in the ELVIS will allow the engineers at Burdell Electronics to put our capacitors through rigorous testing and obtain meaningful data to ensure a reliable product recommend the ELVIS for measuring capacitor wear testing signals because it is affordable will complement existing instruments and is accurate with less than 4 error with a high precision of 4 significant figures2. The fourth most important criterion is feature count Many instruments tout lots of new automatic features that raise the price of the device substantially Extra features are often useful but many times their functionality can be replicated by the engineer by hand fairly quickly The least most important criterion is form factor It is important to keep in mind the real estate available on the bench but it is foolish to soend more money ona smaller device just so it will fit in a particular soot Adding a new shelf or rack is very often much cheaper In addition most benchtop gear is now designed to be stacked on top of one another To compare accuracy the percent error was calculated according to equation 4 below experimental theoretical 1 theoretical aoe Equation 4 Formula for calculating percent error The accuracy criterion was defined as the extent to which a given measurement agreed with the standard value for that measurement Cost effectiveness was defined as producing optimum results for the expenditure and was determined using cost potential maintenance cost feature count precision and accuracy Precision was defined as the degree to which the exactness of a quantity is expressed and was determined using the number of significant figures that the instrument could resolve Feature count was simply the numeric total of features and abilities that the device was capable of Finally the form factor was simply the physic
3. fitness of the measurement instruments of Burdell Electronics a sample waveform was analyzed by each device Equation 1 shows the mathematical representation of the waveform that was generated and analyzed 3000 3000 1 1000 e x t 5 1 e 3000 1 1000 Equation 1 Equation of investigated waveform This function closely matches the behavior of a 5v 500 uF capacitor When plotted this waveform appears as an exponentially decreasing curve as seen in Figure 1 4 0 2x107 4x10 6x107 8x104 t Figure 1 Plot of Equation 1 over an interval taken to be one period This waveform was taken to be periodic with the plot in Figure 1 serving as one period Because of the discontinuities at each end of the function it presents a nontrivial case for measuring signal characteristics using standard instrumentation The function in Figure 1 was first generated Equation 1 using Mathcad It was then turned into a discrete set of 4000 points This number was chosen because it approximated the original function to within at least 90 and gave a visually similar plot to the original continuous signal These points were converted into a waveform file that the ELVIS was capable of reading The arbitrary waveform generator in the ELVIS was then used to generate a periodic version of the plot in Figure 1 The oscilloscope multimeter and the ELVIS were each used to measure the average and RMS voltage of the waveform The oscilloscope was
4. Recommendation Report Measuring Gross Parameters of Signals and Comparing Methods of Measurement Based on Pre Selected Criteria Hunter Scott Introduction Burdell Electronics a world leader in capacitor manufacturing enforces a rigorous wear testing process ona small sample of capacitors from each batch Capacitors are very quickly charged and discharged until electrical or mechanical failure ensues The rating of the capacitors that are being tested in this evaluation are 5v 500 uF These ratings determine the maximum speed the device can charge and discharge which in this case is approximately 1 kHz By measuring parameters of the output waveform from the capacitor it is possible to not only determine when failure occurs but how the component degraded over time The parameters that are being observed are the mean and RMS voltages The waveform produced by an ideal capacitor is known and by comparing it to the signal produced by a real Capacitor it is possible to track the wear over time An instrument is needed to measure the mean and RMS voltages that fits the criteria determined for this recommendation These criteria in order of importance are accuracy cost effectiveness precision feature count and form factor The three devices that are analyzed are the Tektronix 3012B oscilloscope the Agilent 34401A digital multimeter DMM and the National Instruments Educational Laboratory Virtual Instrumentation Suite ELVIS To determine the
5. ad an error of above 45 in one case and is more expensive The Agilent DMM while cheapest has an error of above 63 in one case recommend the ELVIS for measuring capacitor wear testing signals because it is affordable will complement existing instruments and is accurate with less than 4 error with a high precision of 4 significant figures Analysis Criteria and Comparison The criteria that will be compared are accuracy precision form factor available features and cost to purchase and maintain Accuracy is the most important criteria because no matter how many features the device has how small it is or how cheap it is it is worthless and should not be purchased because it will not allow the engineer to do his job The second most important criterion is cost to purchase and maintain Cost and the other criteria can be traded back and forth until equilibrium is reached For instance in some instances an accurate instrument that is not highly precise is acceptable because it is cheaper but still does its job correctly Maintenance costs are important to consider along with initial purchase costs because they can quickly outweigh the initial purchase cost if labor or parts are expensive In this comparison the availability of repair parts and the length of the warranty are considered The third most important criterion is precision Once cost and accuracy have been decided the instrument with the best ability to do its job should be selected
6. al shape and special dimensions of the device measured in inches square inches and pounds Each criterion with its rank can be found in figure 3 Definition of Criterion Metric of Comparison rank 1 most Accuracy The extent to which a given Percentage error important measurement agrees with the standard value for that measurement Cost effectiveness Producing optimum results for Cost potential maintenance of purchase and the expenditure cost feature count precision maintenance and accuracy Precision The degree to which the Significant figures exactness of a quantity is expressed Feature Count Numeric total of features and Numeric total of features and abilities that the device is abilities that the device is capable of capable of 5 least Form Factor The physical shape and special Inches square inches and a n aan pone Figure 3 Table of criteria metric of measurement and importance of each criterion used to reach recommendation Accuracy The raw measurements made with the each device can be seen in Figure 4 Tektronix 3012B Agilent 34401A Theoretical Values oscilloscope DMM Output Average 1 81 V 0 51 V 1 451 V yY Output RMS 2 81 V 1 30 V 1 961 V 1 949 V Output Cycle Not supported Not supported Mean Output Cycle 2 80 Not supported Not supported RMS Figure 4 Results of average RMS cycle mean and cycle RMS of the input wave measured on each device compared to theoretical values To det
7. d 14 72 inches long It weighs 8 pounds and can be seen in Figure 11 Figure 11 Agilent 34401A digital multimeter The DMM has a larger footprint than the oscilloscope at 147 49 square inches However this instrument is designed to be stacked between other similar sized instruments as can be seen in Figure 11 Operation requires a single power cord in the back although test cables can also be attached to the back There are also test cable ports on the front The ELVIS is the largest of the three instruments requiring a computer and benchtop workstation The workstation can be seen in Figure 12 Figure 12 ELVIS benchtop workstation The ELVIS PC interface that allows for the control of elements on the workstation via a PCI bus can be seen in Figure 13 Figure 13 ELVIS PC software interface The ELVIS benchtop workstation is 13 50 inches wide 2 99 inches high and 11 02 inches long for a footprint of 148 77 square inches It weighs 9 pounds Unlike the other instruments the ELVIS workstation cannot have anything stacked on top of it and its dimensions are not conducive to stacking it on top of anything either Additionally a PC keyboard mouse and monitor are required to use any of the analysis tools that the device is capable of Overall the ELVIS has the largest footprint because of the need for a separate PC Power and data cables must be routed from the back of the device and test leads are connected on the front pa
8. ecision The Tektronix 3012B oscilloscope is capable of displaying 3 significant figures for average and RMS values The Agilent 34401A multimeter has the highest precision at 6 5 significant figures for these values The ELVIS can display four significant figures for average and RMS values Only the oscilloscope can measure cycle mean and cycle RMS and it is capable of displaying 3 significant figures for these data points Form Factor The form factor of the device including its size and ease of operation are important since bench space is scarce with all of the other instruments at each workstation at Burdell Electronics A table of the dimensions weight and the footprint of each device is in Figure 9 inches 3012B Agilent 34401A 14 72 10 02 147 49 a ELVIS 11 02 13 5 2 99 148 77 workstation only Figure 9 Comparison of spatial dimensions and weight of each device The Tektronix 3012B digital oscilloscope is 14 8 inches wide 6 9 inches high and 5 9 inches deep It weighs 7 pounds The oscilloscope form factor can be seen in Figure 10 Figure 10 Tektronix 3012B digital oscilloscope The oscilloscope is an average sized benchtop instrument with a footprint of 87 32 square inches It is also able to be stacked on top of other instruments To operate only a single power cable on the back along with a probe cable for each channel on the front is needed The Agilent 34401A DMM is 10 02 inches wide 4 08 inches high an
9. ermine the accuracy of each device the theoretical value of the average and RMS voltage were calculated with Mathcad The errors for each device can be found in the first column of Figure 5 oscilloscope Average Error for Output RMS 44 18 33 29 l Error for Output Cycle 28 82 N A N A Mean Error for Output Cycle 43 66 N A RMS Figure 5 Error of average RMS cycle mean and cycle RMS of the input wave measured on each device All of the values in column one of Figure 4 were obtained using the measure function on the oscilloscope Because this functionality relies on knowing the correct period of the wave a fairly large error occurred The reason for these significant errors is due to the discontinuities in the input waveform Each time the next period occurs an asymptote is encountered The waveform as the Tektronix oscilloscope sees it appears in Figure 6 Figure 6 Periodic version of input waveform plot on the Tektronix oscilloscope This pulse train like behavior makes it very difficult for the oscilloscope to determine the period In fact by using the measure function on the oscilloscope and watching the calculated period it is possible to see the value for the period quickly take on a wide range of values before settling down to the approximate real period as the middle of the wave is encountered To see this edge corruption effect with the naked eye the frequency of the wave must be artificially lowered so that the
10. g cycle mean and cycle RMS Data acquired by the ELVIS is found in column 3 of Figure 4 The error can be found in column 3 of Figure 5 The ELVIS had very low error compared to the other instruments because of its principal of operation The ELVIS uses a 68 pin M series Data Acquisition DAQ device that interfaces with a PC to read and analyze data from transducers on the benchtop workstation This setup allows the ELVIS to look at the entire set of data and apply more intensive processing than the DMM and oscilloscope because it can leverage the use of the attached PC By looking at a continuous stream of data points acquired by the transducers the ELVIS can reconstruct the wave as it is actually sensed and then do math on it without having to trade off accuracy for speed since the Agilent DMM and Tektronix oscilloscope both rely on embedded processors to analyze data There are no restrictions on wave types or shapes that the ELVIS is able to accurately analyze Data was not available on expected device error percentages from the manufacturer Like the multimeter however it is not capable of measuring cycle RMS or the cycle mean Figure 8 shows the typical setup of the ELVIS and how the workstation is connected to the PC via a PCI card 1 Desktop Computer 3 Shielded Cable to M Series Device 2 66 Pin M Seares DAG Devica 4 NI ELVIS Benchtop Workstation Figure 8 Desktop configuration for ELVIS workstation depicting the PCI interface Pr
11. he Tektronix oscilloscope can capture waveforms at 1 25 GS s while the ELVIS oscilloscope can only capture at 500 KS s The lowest percent error of any measurement taken with the Tektronix oscilloscope was 28 82 which is unacceptably high The Agilent 34401A DMM also suffers from poor accuracy with the lowest percent error of any measurement being 33 29 The Tektronix oscilloscope has a precision of three significant figures This is sufficient for the kinds of measurements conducted at Burdell Electronics The Agilent multimeter has the largest precision at 6 5 digits The ELVIS can measure the same parameters that the Agilent 34401A can measure but only with 4 significant figures The oscilloscope has a high feature count with many useful automatic measurements built in including mean RMS cycle mean cycle RMS period frequency width rise and fall time duty cycle overshoot high and low max and min peak to peak amplitude burst width delay phase area and cycle area The ELVIS offers more features than the Tektronix oscilloscope including an arbitrary waveform generator Bode analyzer digital bus reader and writer digital multimeter dynamic signal analyzer function generator impedance analyzer oscilloscope two and three wire voltage current analyzer and variable power supplies The Agilent 34401A has a small feature count compared to the ELVIS and Tektronix oscilloscope but is still capable of measuring DC voltage true RMS AC v
12. input wave can be seen to scroll across the screen The values obtained for the mean and RMS of the input signal by the Agilent DMM are found in column two of Figure 4 This digital multimeter requires a pure sine wave to correctly determine the average value of a signal Figure 7 depicts an excerpt from the user manual regarding the expected error of several major wave types Waveform Crest Factor Average C F AC RMS AC DC RMS Responding Error 1 414 Colibrated for O error 3 9 XA EE mAn 46 for C F 4 C F Figure 7 Method of measurement and expected error for sine triangle and square waves for Agilent DMM Symmetrical waves are able to be analyzed with an error of under 4 This is not the case for nonsymmetrical waves because of reasons similar to those of the oscilloscope The multimeter tries to acquire the period but often isn t able to because of large discontinuities Additionally nonsymmetrical waveforms contain DC voltages which are rejected by AC coupled measurements The Agilent DMM does however employ true RMS meaning that it can accurately determine the RMS value of a waveform even if it is not a sine wave There is a caveat though the crest factor of the wave must be small There is no public documentation on what the exact crest factor limit is however the user manual states that waveforms with large discontinuities can expect an error of around 46 Notice that this instrument does not support measurin
13. nel Feature Count The oscilloscope DMM and the ELVIS all have more features than were tested in this experiment These additional features need to be considered when discussing intent to purchase The Tektronix 3012B additional features can be seen in Figure 14 Triggering modes i Automatic measurements Mean RMS cycle mean cycle RMS period frequency width rise and fall time duty cycle overshoot high and low max and min peak to peak amplitude burst width delay phase area cycle area Figure 14 Feature list of Tektronix 3012B oscilloscope The oscilloscope has two channels with a total of 100 MHz of bandwidth and a sampling speed of 1 25 GS s It has a time base range of 10 ns to 10 s div and a maximum input voltage of 300 V with a 10x probe It has the ability to connect to an external PC over GPIB and has 9 different trigger modes Along with automatically measuring mean RMS cycle mean and cycle RMS it can measure period frequency width rise time fall time duty cycle overshoot high low max min peak to peak amplitude burst width delay phase area and cycle area The Agilent multimeter performance statistics are found in Figure 15 GPR GPIB Yes but only at 1000 Hz Supported measurements DC voltage true RMS AC voltage resistance DC current true RMS AC current frequency period continuity diodes Figure 15 Feature list of Agilent 34401A digital multimeter The 34401A can measure u
14. oltage resistance DC current true RMS AC current frequency period continuity and diodes The oscilloscope has the smallest footprint and even though it is not stackable it still only takes up 87 32 square inches of countertop The Agilent 34401A has the second smallest footprint at 147 49 square inches plus the ability to stack The large 148 77 square inch footprint of the ELVIS not including the footprint of the PC may initially seem detrimental but engineers at Burdell Electronics already have a PC at their lab benches They also normally have a breadboard for prototyping By using the existing computer as the host for the ELVIS and replacing the breadboard with the ELVIS workstation which has a breadboard built in very little additional space is taken up The Tektronix oscilloscope has a three year warranty the longest of any device and is easy to find parts for It costs 3000 The ELVIS has a one year warranty and a price of 2000 making cheaper than the oscilloscope but not the DMM Additionally parts are difficult to find for the ELVIS The Agilent 34401A costs 900 and comes with a one year warranty Given that parts are easy to find this device is expected to have a maintenance cost that is lower than the ELVIS and a slightly higher than the Tektronix oscilloscope Recommendation While the Tektronix 3012B oscilloscope offers a small form factor it lacks the ability to accurately determine the given wave characteristics
15. oscilloscope has a bandwidth of 50kHz compared to the Tektronix 3012B s 100 MHz The ELVIS oscilloscope has two channels that can each capture 500 KS s compared to the Tektronix s 1 2 GS s Cost effectiveness of purchase and maintenance A table of the cost warranty information and spare parts availability of each device is found in Figure 17 PY Cost dollars Warranty period Parts availability Tektronix 30128 3000 Easy to find Agilent 34401A 900i tear Easy to find ELVIS 2000 Hard to find Figure 17 Comparison of cost effectiveness criteria of each device The cost of the Tektronix 3012B oscilloscope is 3000 The cost of ELVIS is 2000 but this does not include the cost of the PC that is needed to actually use the device The cost of the Agilent 34401A is S900 Tektronix provides a 3 year warranty for electrical or manufacturing failures National Instruments provides a one year warranty against manufacturing failures and Agilent also proves a one year warranty In the event that an in house repair of any instruments is necessary parts are readily available for the Tektronix 3012B and Agilent 34401A however parts for the ELVIS benchtop workstation are difficult to find Conclusions The ELVIS is the most accurate instrument by a significant margin with the highest percent error of any measurement of 3 27 The ELVIS s built in oscilloscope can see waveforms but not waves as fast or detailed as the Tektronix 3012B can T
16. p to 1000 V at 6 5 digits of resolution It has 0 0015 basic DC V accuracy and 0 06 basic AC V accuracy It supports GPIB like the oscilloscope but can only output at 1000 Hz Aside from measuring DC voltage and true RMS AC voltage it can measure resistance DC current true RMS AC current frequency period continuity and diodes The ELVIS incorporates many standard benchtop instruments into a single package Details of its features are found in Figure 16 Supported functions Arbitrary waveform generator Bode analyzer digital bus reader and writer digital multimeter dynamic signal analyzer function generator impedance analyzer oscilloscope two and three wire voltage current analyzer and variable power supplies GPIB No but data acquisition is supported via LabVIEW Figure 16 Feature list of ELVIS It features an arbitrary waveform generator Bode analyzer digital bus reader and writer digital multimeter dynamic signal analyzer function generator impedance analyzer oscilloscope two and three wire voltage current analyzer and variable power supplies It also has higher level functions and integration with LabVIEW that can be used in analysis The built in digital multimeter can measure Capacitance continuity current diodes inductance resistance and voltage Compared to the Agilent multimeter it has 0 3 accuracy over an input range of 20 V DC Likewise it has a 0 3 accuracy over an input range of 14 V AC The
17. used to measure the cycle mean and cycle RMS in addition to the average and RMS An excerpt from the Tektronix 3012B user manual is shown in Figure 2 The arithmetic mean over the first cycle in the waveform The true Root Mean Square voltage over the first cycle in the waveform The arithmetic mean over the entire waveform The true Root Mean Square voltage over the entire waveform Figure 2 Definitions of each measurement parameter used to compare measurement devices The first two rows of Figure 2 give the definition of cycle mean and cycle RMS along with a small image depicting the method used to measure that data point Figure 2 also gives the definition of the mean and RMS values that the oscilloscope uses Note that the definition for RMS is a true RMS value The mathematical definition of the mean value is given in Equation 2 1 b f x dx a Equation 2 Average value of a function The mathematical definition of the RMS value is given in Equation 3 1 T aan 2 y _Uorat Equation 3 RMS value of a function The definitions of cycle mean and cycle RMS are also calculated with Equations 2 and 3 respectively however the interval over which integration is performed is a single cycle After an exhaustive analysis it was found that the ELVIS was the only instrument with acceptable accuracy along with a reasonable price and form factor It also has a very wide variety of features The Tektronix oscilloscope h
Download Pdf Manuals
Related Search