As the industry continues to seek lower test costs, many RF test engineers must continue to reduce measurement time. As you know, the testing of wireless network (WLAN) devices must also cater to this trend. Whether it is an automated test system for design verification or a test operation for the final product, the measurement speed of the test system has become increasingly important. However, in most cases, in addition to reducing test time and reducing test costs, the measurement accuracy and repeatability of the system cannot be affected. This technical article will address the WLAN measurement operations and illustrate several trade-offs that may affect the measurement speed. After understanding the concepts, a better practice description will be provided for the measurement speed of the test system. This technical article will explain the following elements in order: average and repeatability; EVM of complete pulse and partial pulse; composite measurement and single measurement; measurement frequency span and measurement time, and finally the relationship between CPU and measurement time. For the above related elements, this technical article will perform sample measurement operations using the NI PXIe-5663 - 6.6 GHz RF vector signal analyzer. These examples use the NI PXIe-5673 - 6.6 GHz RF vector signal generator as an excitation. And all of the examples use the NI WLAN Measurement Suite, which includes NI LabVIEW and LabWindowsTM/CVI signal generation and analysis toolkits to build measurement platforms. To learn more about how to set up a PXI WLAN test system, see Configuring a Software-Defined WLAN Test System. Although this technical article focuses on the operation of PXI RF instruments, the same basic measurement elements may be common to any RF instrument. Therefore, both PXI instruments and traditional RF instruments can be used to improve the performance.
Trade-off factor 1 – average and repeatability
Whether it is an application for automated design inspection or production testing, a common technique for improving the repeatability of measurement results is the result of average multiple measurements. However, if a large number of average values ​​are to be set to improve the repeatability of the measurement results, the measurement time will be increased. In general, the overall measurement time can be linearly adjusted by the number of average values. Therefore, if a single measurement operation takes 20ms, then the same measurement will take nearly 200ms when it is averaged 10 times.
Furthermore, since the averaging operation can offset the non-repeatable impairments, such as AddiTIve white Gaussian noise (AWGN), between multiple measurements, the repeatability can be effectively improved. To understand the impact of average operation on repeatability, you can use the NI PXIe-5673 RF vector signal generator with the NI PXIe-5663 RF vector signal analyzer to perform loopback testing. With the above device, an 802.11g Orthogonal Frequency Division Multiplexing (OFDM), -10 dBm power intensity RF signal can be generated at 2.412 GHz. Similarly, using four different signal types – BPSK (6 Mbps), QPSK (18 Mbps), 16-QAM (24 Mbps), and 64-QAM (54 Mbps), you can understand the pulse size and modulation type versus measurement time. Impact. If a 1024-bit payload is used, then each signal type will have a different number of OFDM symbols. For example, a BPSK pulse will have 343 symbols and a 64-QAM pulse will use 39 symbols. Therefore, the pulse interval of each signal type is also different, and Table 1 shows the difference in pulse widths of different types.
Table 1 Modulation method of 802.11a/g variable data transmission rate, pulse interval and number of symbols
Error vector strength (EVM) measurement operations can provide complete signal modulation quality. In the EVM measurement operation, there are two built-in methods that can show average results. For IEEE 802.11a/g pulses, the results of the measurements will cover the individual OFDM subcarriers and symbols. Expressed in the root mean square (RMS) of the EVM. From Table 1, it should be possible to see directly the number of symbols in the pulse, and if the EVM is a lower 6 Mbps (BPSK) data rate, it should be possible to generate repeatable measurement operations with more than 54 Mbps pulses. It can thus be seen that longer pulses also have more symbols. However, the above hypothesis is true only when the EVM is represented as RMS by a complete pulse (rather than a specific partial pulse). The trade-off element 2 will analyze the partial pulses to illustrate the relevant repeatability.
In the general case, we can assume that more repeatable EVM results will be produced when performing longer pulse measurements. Figure 1 shows the relationship between the average number of times and the standard deviation of the measurement. These measurements were performed using the NI PXIe-5673 RF vector signal generator and the NI PXIe-5663 RF vector signal analyzer. An average RF power of -10 dBm was used, and the intermediate frequencies of both instruments were set to 2.412 GHz.
Figure 1 Average operation reduces the standard deviation of the measured mean
Figure 1 shows that as the average number of uses per measurement operation increases, the standard deviation of 1000 EVM measurements will decrease. Please note that since the signal source used in Figure 1 is an RF vector signal generator - specifically designed to produce repeatable signals, the EVM and standard deviation in Figure 1 are significantly better than those produced by 802.11g converters. The actual situation. Therefore, the results shown in Figure 1 can be used as a criterion for repeatability. Also, please note that it is only meaningful to measure the repeatability of the measurement in terms of Absolute measurement value. In general, as long as the EVM standard of the test instrument is higher, the effect of repeatability is smaller. Table 2 shows the EVM results when the measurement operation is set to 10 times average.
Table 2 EVM and modulation type remain relatively consistent
Table 2 shows that the EVM will be consistent regardless of the modulation method. However, this also means that the user can obtain a better standard deviation by a longer pulse. Of course, you will also need to measure more symbols. For example, if 10 averaging is performed to achieve a standard deviation of 0.081 dB on the 64-QAM signal, then when measuring the complete pulse of the BPSK signal, the same standard deviation can be achieved with only 5 averaging.
In general, it takes only a long measurement time to achieve a lower standard deviation result by averaging. Table 3 illustrates this relationship with a 54 Mbps pulse. Note that the measurement times in Table 3 include gated power and EVM measurement operations.
Table 3. Measurement time increases as the average number of times increases
In Table 3, we performed a composite EVM and gated power measurement operation using a PXIe-5663 RF vector signal analyzer with a set of NI PXIe-8106 controllers. The EVM is calculated from the RMS of the complete pulse; and the mean and standard deviation are calculated over more than 1000 measurement operations. Table 3 shows the linear relationship between the measurement time and the average number of times. The NI WLAN Analysis Toolkit uses the so-called Asynchronous fetching technique, which handles the previous records as the analyzer extracts new records. Therefore, the user can perform measurement operations on multiple averaging without being limited by the linear time (Linear TIme). In addition, please note that the single-average EVM and power measurement listed in Table 3 will take 9.4ms, but if the average number of times is set to 10, the measurement operation will only take 63.6ms, which is the average consumption per time. The time is 6.3ms.
Inner Battery,Electric Bike Battery,Ebike Battery 36V 12Ah,Electric Bicycle Lithium Battery
Changxing Deli Technology Co., Ltd. , https://www.delipowers.com