Optical passive device testing is an important part of the optical passive device production process. Whether it is the selection of test equipment or the construction of the test platform, it actually reflects the test concept of the device manufacturer, or the device manufacturer for precision instruments and precision testing. Awareness. Different test equipment and different test system construction methods will have an impact on the accuracy, reliability and operability of the test. This article briefly describes the testing of optical passive components and discusses the impact of different test systems on accuracy, reliability, and repeatability. ![]()
In the test system shown in Figure 1, the test light first passes through a polarization controller and then through a return loss meter. The output of the return loss meter is equivalent to the light output of the test. It should be emphasized here that since the polarization controller has 1~2dB insertion loss, the return loss meter has about 5dB insertion loss, so the output light is 6~7dB smaller than the direct light source output light. Two single-ended jumpers can be connected to the return loss meter and the power meter respectively. The welding method is used for the test reference. The welding device can also be connected to the optical path to test the insertion loss and polarization dependent loss of the device (PDL). And return loss (ORL). This method is commonly used by many device manufacturers. The advantage is very convenient. If the power meter end uses a bare fiber adapter, it only needs 5 times of fiber cutting and 2 times of fiberizing (return loss is tested by comparison method*) to complete the insertion loss. , return loss and polarization dependent loss testing. However, this test method has serious shortcomings: since the polarization controller uses the random scan Poincare spherical method to test the polarization-dependent loss, there is no need to make a test reference, so the PDL measured by the system is actually the polarization controller output to the optical power meter input. The integrated PDL value on the link between the ends. Since the passive components such as the coupler in the return loss meter and the optical port of the return loss meter APC have their own small PDL, only the APC optical port PDL value is about 0.007 dB, and the PDL addition does not hold, so the PDL test The value system error is large, and the repeatability and reliability of the test are not ideal, so this method is not a recommended method. The improved test method is shown in Figure 2. ![]()
In the test system of Fig. 2, since the test light passes through the return loss meter and then passes through the polarization controller, the polarization state of the light between the output end of the light source and the input end of the polarization controller does not change greatly, that is, the system can measure Get a more accurate DUT PDL value. However, the problem has not been solved, PDL is ok, but the return loss test is affected. We know that to test the return loss of the DUT, we need to first measure the return power of the test system itself, and then measure the return power of the system and the DUT, and subtract the DUT return power. It is easy to understand mathematically, the smaller the return power of the system, the better the accuracy, reliability and dynamic range of the DUT return loss value, and vice versa. In the second system, the system return power includes the return power of the polarization controller, so it is relatively large, which limits the reliability and dynamic range of the DUT return loss test. However, in general, as long as the return loss value Other than -60dB is not tested, the problem of this configuration is not large, so it is a fairly acceptable test method in the case where the return loss is not high. In addition to the above two test schemes, there is also a test system based on the Mueller matrix method (Fig. 3). ![]()
This test system uses an erbium-doped fiber ring-based tunable laser (EDF TLS) rather than a normal external cavity laser. This is important, as discussed later, in addition to the Mueller matrix analysis-specific polarization control. , return loss meter and optical power meter. Since the test light is required to have a stable polarization state when testing the PDL by the Mueller matrix method, a hard jumper should be connected between the tunable light source and the polarization controller and between the polarization controller and the return loss meter, so that the fiber swing pair can be eliminated. The impact of the test. Testing the PDL with the Mueller matrix method requires reference, so the impact of the test link on the PDL test can be ruled out to some extent. Therefore, the system can obtain higher PDL test accuracy and return loss and insertion loss accuracy, and the reliability of the test. The operability is very good. In this system, each test unit does not work independently, they must be integrated into one, the tunable light source keeps scanning, the power meter keeps collecting data, the test host analyzes the collected data, and finally the IL, PDL and ORL with wavelength The curve of change. This method is currently mainly used in multi-channel device testing such as DWDM devices, and is currently a very advanced test method. Among the above three test methods, the author believes that except for the last method, which is the best solution for testing DWDM multi-channel devices to achieve fast test, the other two methods are not enough, because they all emphasize convenience and neglect the precision test. Accuracy, reliability and repeatability requirements. This is also why many device manufacturers test the same products, and today's measurements and tomorrow's test results will be very different. See the coupler test assembly method in Figure 4 for the solution. ![]()
Using the configuration of Figure 4, the return loss and directionality parameters of the device, as well as the device PDL and average IL, can be derived at one time. Since the test laser source is a polarized light source, there is a PDL value system test uncertainty for the device insertion loss test. If the device itself has a large PDL, it will be a problem, so the depolarizer is used for the average loss test. The advantage of this test method is that the test is stable and accurate, basically eliminates theoretical or systematic errors, and even suppresses random errors. For example, the insertion loss is tested by passive depolarizer. The disadvantage is that three stations need to be set up. Andre Girard, a senior expert at EXFO and chairman of the ITU PMD group, has a mantra called Nothing perfect! The same is true for device testing. It is convenient to test, but the reliability of the test, the decrease in repeatability, or the reliability and accuracy of the test, but the test is relatively troublesome? Everything is between personal choices. The above is to discuss the best test of the device from the topological relationship of the test station, that is, the test station. In fact, the selection of test equipment in the test process occupies a more important position. The effects of test light sources, power meters, polarization controllers, and test systems on test accuracy, reliability, and repeatability are discussed below. 1. The light source selection test source is the excitation source of the test system. Generally, it is not required to be too high in power for testing, not for transmission. The laser source is 0 dBm, and the wide spectrum source is -10 dBm/nm, which is sufficient for the test. Also because it is used for testing, the power stability of the light source is quite important, in addition to the problem of a coherence length. In fact, any laser source has a problem of coherence length. Generally, the coherence length of a FP or DFB laser source is 1,000 meters or more, and the line width of the laser is artificially increased by about 10 meters, that is, as long as the optical path of the test system is Shorter than this length, there will be interference, the test will be inaccurate or the reliability will be reduced. There is a tunable laser based on an erbium-doped fiber ring that solves this problem. The laser has a coherence length of only 15 cm, and the device test length is generally 1 to 3 meters, so there must be no coherent influence, thus making the test The value is very stable, repeatable and reliable, and is a very suitable light source for device testing. In addition to the coherence length, the signal-to-noise ratio of the laser source is another key parameter. The ratio of the signal from the laser source to the radiated noise (S/SSE) is a key factor limiting the dynamic range of the test. If the S/SSE is only 60dB, then when the 65dB filter is tested, the filter can only filter out the spontaneous emission noise, so the test can only show 60dB, resulting in the test failure. In general, the S/SSE of a tunable laser source is 75 dB, so the S/SSE value of the source should be noted when testing large dynamic range devices. For broad-spectrum or ASE sources, spectral stability is a key parameter. Spectral stability is a more stringent and meaningful parameter than integrated power stability. It characterizes the peak-to-peak variation of a broad-spectrum source over a period of time. Maximum value. Since the broad spectrum source is typically used with wavelength selective devices such as spectrometers or wavelength meters, the integrated power stability does not make much sense for testing. 2. The power meter selects the material of the power meter detector to determine the overall performance of the power meter. Generally, there are detectors of materials such as Ge, Si, and InGaAs, in addition to a low polarization reflectivity (PDR) detector. The detector is based on the addition of InGaAs detectors to make it very insensitive to PDL, so it is well suited for PDL testing. In addition to materials, the detector area is an important parameter that determines its use. The larger the detector area, the stronger its light-receiving ability, but the sensitivity is reduced, and vice versa. Therefore, the optical power meter detectors generally used for calibration are all larger than 3 mm2, and are used for detecting small optical power such as -100 dBm. The area of ​​the optical energy detector is generally 1 mm2. Generally speaking, if the optical power meter uses a bare fiber adapter, the optical power meter detector area is required to be larger than 3 mm2, otherwise the optical fiber exiting light is difficult to be fully coupled to the detector, which greatly reduces test repeatability and reliability. In fact, even with large-area detectors, the fiber in the bare fiber adapter is very likely to touch the detector, causing the detector to age and reduce the test accuracy. Therefore, it is generally recommended to use the fusion method, so that although the fuse is added once, it is ensured. The long-term stability and reliability of the test. In addition to the above traditional detector types, there is also a wide aperture integrating sphere detector technology. The detector's detector area is equivalent to 7mm2. Due to the integrating sphere technology, it does not have the surface unevenness of the traditional large-diameter detector, the fiber alignment and the fiber tip easily reach the detector surface. The test repeatability is also Traditional detectors cannot be compared. 3. Polarization Controller Selection For the random scan Poincare ball polarization controller (PC), the scanning period, the coverage of the Poincare sphere area, the polarization of the light passing through the PC, and the optical power fluctuations due to the PC are some key parameters. The meaning of these parameters is easy to understand. I only want to focus on the impact of optical power fluctuations on the test due to PC. We know that the PDL test is actually detecting the maximum value of the optical power change through the device under test when the transmitted light polarization state (SOP) changes, so if the optical power changes due to other reasons, the test system will mistake this. It is also a PDL, which causes the PDL test to be too large. Therefore, for PC, the optical power fluctuation value will directly affect the accuracy of the test. 4. Selection of test system The so-called test system mainly refers to the joint work of two or more test tables or modules, forming a new operation interface after combination, and completing the test equipment for automatic test. The traditional system is built by a computer, using GPIB port to control several optical test instruments. Here we focus on the method of assembling the system through the module. The main idea is that the test host itself is a standard computer. The test host has 5 slots, which can be inserted into the test module to form a simple system. For the large test system, an expansion machine can also be added. Between the host and the expansion machine. Connected via data lines. In this way, there is no difference between the slot on the expansion machine and the slot on the host. The module inserted in the expansion machine and the module inserted on the host have no difference in data transmission rate. Therefore, the method of setting up the test system makes the system data The transmission speed is very fast and the operation is very convenient. The expansion machine can also cascade expansion units to form a larger system, so the scalability is very good. For example, EXFO's IQS-12004B DWDM test system will be tunable light source, fast optical power meter, Muller matrix polarization controller and wavelength calibration. The unit is organically combined, the test wavelength accuracy is 5pm, the curve of IL, ORL and PDL with wavelength can be measured with a click of the mouse, and the crosstalk matrix is ​​obtained, which also shows the system construction using the host + expansion machine. The advantages. Conclusions This paper discusses the reliability, accuracy and repeatability of the test process from the perspective of the topology of the test station and the choice of test equipment. In fact, the production process of optical devices is a very complicated science. It is not a simple sentence. It can be said that different products and processes are different. It is worthy of further study, so that it is not clear what is wrong. I am in a hurry. * The so-called comparison method test return loss refers to the standard return jumper (generally the return loss value of 14.7dB and the standard jumper certified by the international organization) to calibrate the system, the light return of the device under test is compared with Return loss value. This method of testing return loss is more convenient than the traditional method, the test value is more accurate, and it is affected by the instability of the light source, optical power meter, etc.
1
1
Longdi Light Equipment Technology Co., Ltd , http://www.gz-ledlight.com