Like the star pupil that you are, I bet you are already ahead of the game-you're thinking to yourself about all the possible scenarios of what can cause bad things to happen to good readers. You might think about how the antennas could have been dropped during shipping, the cable could be kinked, or the reader placement could be causing overheating. If you have been playing with RFID long enough, you might be catching on to some of the patterns in deployments. Although that is the right way to start thinking about being ahead of the reader, what you really need is a process. The process has to focus on two critical factors of diagnosis:
The surrounding environment
The RFID system
The RFID system has everything to do with what is happening with tags, readers, and antennas. There is a logical sequence you should follow to perform a proper diagnosis, which I'll get to in a minute. The surrounding environment is the other factor that should have been looked at very closely during installation. Either may have changed after the RFID system was installed or was not properly evaluated in the first place.
What is the ideal process for diagnosis? It has to start with gathering metrics and defining the test scenarios. I always say you can't manage what you can't measure. So for the first step, find the reader site(s) that you are going to test and gather some meaningful data around how they are performing today. Depending on your time frame, you should do this over the course of a week or two, and try to get at least several hundred data points so that you will have a sample of data that is representative of the entire population. This is creating a starting level or baseline performance that can be used to assess the impact of individual changes on performance and offer insight into the most effective test sequence.
The next step after creating the baseline is to determine whether there is any potential ambient electromagnetic noise (AEN) by following the procedure for a full Faraday cycle analysis, as outlined in Chapter 3, "Site Analysis." If you see any interference in or around the frequency you are operating within, you know you have a potential cause of your problem. It's important to note that even if you are operating in the 902–928 MHz range and you spot some AEN at 900 MHz, if the wattage is sufficient, that signal outside the 902 range can bleed over and cause interferences at the low end of your useful range (for example, 902, 903, or 904 MHz).
The next step in the physical analysis is to measure the actual reader topology. You need to document and understand where the antennas are placed in relation to where the tag is to be read. Taking measurements should also include how long the cable is between the antenna and the reader, because the longer the cable, the more loss is incurred as the signal travels from the reader to the antenna. You'll also need to understand where the tag falls within the interrogation zone and how much variability there is in where the tag falls compared to the antennas.
These three steps give you critical information needed to make changes within the surrounding environment that can have a significant improvement on system performance. This is particularly true if you are dealing with a situation that includes a conveyor or shelf reader. The reason is RF localization-you would likely be looking to read in a very small area; this means that power levels are much lower and the "sweet spot" of the interrogation zone is much smaller. Therefore, measuring the specific physical area is critical, both where the tag comes through the zone and where the coverage of the antenna is.
While you are doing an analysis of the environment, you can start to look at issues with the actual equipment as well. The first place to begin is with the reader. One of the biggest areas of variation are the reader ports. Sometimes the connections are not welded securely, other times the reader quality check was not performed thoroughly, and other times there may just be damage during use. You'll be the first one to determine how consistent the power output of the reader is across each port. You will need a very accurate spectrum analyzer hooked up to each port, and set the reader to broadcast at full power. With your spectrum analyzer, track and record the power out of each port and see what kind of variation you have.
The next step is to investigate the cable leading to the antenna. Cable that connects RFID readers to antennas comes in all different shapes and sizes. There are several factors that can cause loss over antenna cables:
Amount of insulation on the cable (thickness)
Length of the cable
Number of connections on the cable
Radius of bends as the cable is routed from reader to antenna
Type of connector used
A basic investigation of the cabling can reveal some interesting information. Make sure you take the time to look for cuts, scrapes, bends, and connections because those things all induce loss or noise.
It may be more helpful to look at troubleshooting as a flow of closely related steps. If you look at Figure 8.1, you'll see the two specific areas (the RFID system and the physical environment) and the steps that you need to take to investigate each one in greater detail.
Figure 8.1: The process flow of troubleshooting an RFID system
As you can see from the flow chart, it takes about a dozen tests to do a good job evaluating a poor RFID system. Now I'll take you through the details of each one of those tests and teach you specifically how to conduct each one from start to finish.
This is the easiest of all the tests because you should already be familiar with site surveys, full Faraday cycle analysis, and path loss contour mapping. If you're not, you need to go back and review Chapter 3 to learn how to set up a spectrum analyzer for the proper protocol.
If you want more detail, you can purchase a copy of my other book, RFID For Dummies (Wiley Publishing, 2005).
A simplified version of a full Faraday cycle analysis would entail setting up a spectrum analyzer to investigate what is happening near the area under investigation. If you do it initially with the reader functioning, you should set your spectrum analyzer to Max Hold and watch the signal as it frequency hops pseudo-randomly through the entire band. In just a few minutes, the entire 902–928 MHz band should be filled with the exact same signal power. This is a great chance to see whether the reader is broadcasting outside the 902–928 MHz band and to make sure that each channel is outputting the same power level. After you've investigated that, shut off the reader.
If you set up the spectrum analyzer while the power light is on the outside of the reader and you don't see anything, you may have already found your issue. The first thing you should try to do if the power light is on is to ping the reader. You can do that by connecting the reader to a laptop via an Ethernet cable or a serial cable. To connect using a serial cable you need a cable with DB9 connectors and you have to know the COM port number on the laptop that allows you to configure the COM port. It is very important that you know what baud rate settings are required to connect with the reader, so check the reader manual before using the serial port.
To ping a reader, you need to know the reader's IP address and have an Ethernet cable (a crossover cable or use a switch with a regular Ethernet cable) and a laptop or desktop computer. It's as easy as plugging the computer into the reader and clicking the Start button, followed by Run, and then typing cmd. This gives you a command window where you will likely see C:\Documents and Settings\Administrator>_. Just type the word ping followed by a space and the reader's IP address. If you get a response indicating "request timed out," the reader is not communicating with the outside world.
As you by now realize, troubleshooting is slightly different from the initial site survey because you'll want to make sure that you are capturing your reader's RF output as well. The specifics of the reader are going to be addressed later in the "Reader Settings" section, but you need to know that the reader is functioning first. First set up a spectrum analyzer and signal generator as you would to create a path loss contour map (PLCM). The difference when using the PLCM for troubleshooting is that the antennas are already set up and mounted where they will be. So rather than have the outside antennas on a tripod, they will be attached to the dock door, conveyor, or other location-in other words, you'll be using the actual antenna from the interrogation zones.
The process to use PLCM for troubleshooting is as follows:
Turn on the spectrum analyzer as you would to create a normal PLCM, with the center frequency set at 915 MHz if you're testing UHF, or 13.56 MHz if you're testing high frequency (HF). Set the span appropriately.
Verify that there is no external noise generating a signal to the spectrum analyzer.
Unplug the first reader antenna from the reader (which should be turned off) and plug it into the spectrum analyzer.
Set up a signal generator as directed in the PLCM method and attach it to a ground plane at the same location you would ideally like to read a tag.
Set the signal generator to broadcast a signal at the desired frequency-UHF or HF. If it's UHF, you will want to test across the frequency band, at the very least at 902, 915, and 928 MHz.
Leaving the signal generator to broadcast at a constant power level, test all the antennas attached to the reader as they are mounted.
Gather all the input data across the full frequency span for each antenna that is used in the interrogation zone and input the average power into an Excel spreadsheet.
Armed with the performance data from each individual antenna, you can look at how they are performing relative to each other and determine whether there are any glaring issues-usually this will be seen as nulls or null spots within the interrogation zone.
The next step is to leave the spectrum analyzer set up in that same location and test through the reader's frequency band again. This will quickly identify whether other readers are on and operating at a powerful-enough level to interfere with the interrogation zone under test. You may pick up other readers or other devices. Figure 8.2 shows a screen from a spectrum analyzer showing several eruptions of RF interference that need investigating. If you do pick up interference locally, your poor read rates could be caused by standing wave or multi-path issues. (See "Surfing a Standing Wave.")
Figure 8.2: Spectrum analyzer showing points of potential interference
Because radio frequency communication travels over electric or magnetic waves, it has a consistent, predictable waveform, or pattern. For instance, a UHF wavelength at 915 MHz is about 33 cm. That means that from start to finish, a full wave takes just over a foot. A standing wave is caused when a wave that starts in one direction is heading toward a wave from the opposite direction, and those waves started 15.6 cm out of phase from each other-a half a wavelength apart. This is also referred to as a multi-path issue. When the waves eventually meet, they cancel each other out and there no longer is any RF energy to read a tag.
Standing waves most often happen when multiple readers are set up in close proximity to each other, because there are so many RF waves being produced at usually a very high power. The second cause of standing waves is a highly reflective environment, so that waves might bounce off metallic surfaces and come back to create standing waves. Combine the two causes-multiple readers set up in a metallic environment-and you've got RFID Armageddon!
The antenna is to the RFID network what the tires are to a car. If the reader has the best engine, is running high-octane gas, and has bald tires, you'll never get the benefit of using that great engine because the tires will keep slipping when you hit the gas. An antenna has the same issues-you might have a great reader and ideal configuration, but if your antenna is poorly tuned, you may never get good results. The only way to do a specific and isolated analysis on an antenna is to use a spectrum analyzer equipped with a voltage standing wave ratio (VSWR) bridge. The VSWR bridge will allow you to tune an antenna properly for whatever resonant frequency you are looking to read across. You can also compare the performance across the antenna. The best way to do this is to save the VSWR measurements, like the ones in Figure 8.3.
Figure 8.3: Antenna performance
The deeper the trough, the more power will be available to power the tag and get a response back. Look at Figure 8.3, a set of real-world test results from an ODIN Technologies deployment team that was brought in to fix an RFID system that was poorly performing in a manufacturing line. Scenario A was the original result, which is tuned close to the resonant frequency (in this case 13.56 MHz) but is not very deep, so the read performance is mediocre at best. Scenario B is another faulty antenna that has a much deeper response (the trough goes all the way down to the bottom of the screen), but the resonant frequency is significantly off of 13.56 MHz and the result is no read rates. Finally, scenario C is an ODIN-tuned antenna that has a resonant frequency spot on to 13.56 MHz and a nice deep trough. This resulted in 100 percent read rates in the manufacturing line.
HF is very different from UHF in terms of working in the field environment. HF works on the principle of inductive coupling. The bad news is that most HF antennas straight out of the box are tuned to read in free air at that 13.56 MHz. The resonant frequency can easily be changed, usually by just turning a screw on the antenna. This allows you to account for any detuning of the antenna that might happen because of a close proximity to steel, aluminum, or other metal surfaces.
Cables can contribute as much to poor performance as the antenna. The industry typically ships out LMR-240 with a reader. If you are running to the other side of a dock door, using LMR-240 across a span of 25 or 30 feet can mean that the power getting to your antenna might be cut in half, particularly if there are connectors and elbow attachments. Consider shortening the length of the run or using a lower-loss antenna cable such as LMR-400. Another cause of loss in signal is using connectors that are not soldered to the coax cable. Make sure that for maximum efficiency soldered connections are what connect the core and the connector.
The good news from a troubleshooting perspective is that tags can have a huge impact on the performance of the RFID network, and can be easily measured using the physics of science-not trial and error. The bad news is, if you don't have control over what tags come into the RFID network that you are troubleshooting, the tag can be your biggest headache. Aside from the obvious problems of a tag being damaged because of how it was applied or how it was handled before storage, the biggest issues with tags are where and at what orientation they are put on an object.
A number of factors impact a tag's performance. Figure 8.4 summarizes these influences.
Figure 8.4: Performance factors of a tag
The best way to determine whether the tags are the cause of a problem with read rates is to get some high-quality tags, test for an average-performing tag, and run a known "average" performing tag through the interrogation zone on empty cardboard boxes with the tags exactly parallel to the interrogation zone antenna. If you consistently get read rates that are above 99 percent, your tag is unlikely to be the problem.
A common problem is that tags are stored improperly or put onto the cases with too much force. The tags might have been written to without issue, but then been damaged enough to stop working once applied. It's always a good idea to test actual tags that are in the supply chain to see how effectively they are being read.
Just to refresh your memory on testing your tags, take a look at Chapter 4, "Tags." Make sure that you use a sample of 10–12 tags that are indicative of a larger population of tags to find one that is an average tag. This is why it's so important to test with average tags-tags that require the same minimum effective power to get a response. Using a tool such as ODIN's EasyTag is the best way to determine-by using science, not trial and error-what the proper location and the optimal tag is for a specific product. Tools such as EasyTag also give users a measurement of minimum effective power (MEP), which is the de facto standard for tag testing, so that different tags can be measured using MEP as a way of comparing performance with a similar metric no matter where the tags are being tested.
Electrostatic discharge (ESD) can, although rarely, damage RFID tags. ESD occurs when, because of an excess of electric charge, electric current suddenly flows on a path from where it is stored on an electrically insulated object to an object at a different electrical potential such as a ground-which is what an RFID tag can act like.
The last thing to investigate on the tags is the data stored on the tag. If the tags are coming off a printer/encoder or being prewritten, it is possible they have the wrong data or the incorrect data construct. Some tags can be "put to sleep" by using special commands or even made entirely inoperable. It is possible that bad tags aren't really bad but have been written with an improper command.
Before you even start to troubleshoot an individual reader, you need to take five simple steps:
Reboot the reader.
Verify that the reader has the latest available firmware on it.
Verify that the reader is communicating to the outside world.
Visually inspect the cables and connections.
Verify that the antenna configuration is in the proper mode (either multi-static or bi-static).
If you complete these five simple steps and still have not identified a simple solution, you have to put on your physics hat and start using your science training and RFID tools such as a spectrum analyzer and signal generator to find the problems.
The granddaddy of reader settings is ERP. Different readers have different power settings, but the biggest single mistake that novice RFID technicians make is to set the reader on full power. This will work if there is only a single reader set up in an area without much interference, but introduce another reader or two and some reflective material, and full power is a recipe for disaster. The problem with using full power is that you are likely to get ghost reads and to pick up tags that are coming in through other dock doors, or worse yet that are sitting on a spool inside an RFID printer. At ODIN Technologies we have deployed hundreds of readers across the globe, using HF, UHF, and other technologies, and we have found that very seldom is putting a reader on full power the best way to start setting up a reader.
Start out the investigation of your reader by checking the power setting. Use the operating manual to determine what the power levels are and how much control your reader gives you. Then set the reader at its lowest level and begin your testing. To test the right setting, you can use a very scientific method with the spectrum analyzer and ¼-wave dipole antenna in the middle of the interrogation zone and track the power being broadcast at each channel in the frequency of the reader. This should result in a very even power level on your spectrum analyzer showing that each channel puts out the same power level.
The next step to investigate is the reader temperature. Two scenarios need to be investigated. First, is the reader being left on for too long, causing the temperature to get too hot inside the reader? Second, is the reader turning on only for the read event, via an I/O, so it is not effectively warmed up? Either issue can cause poor read results.
After the obvious power settings and reader temperature reviews, you'll need to break out the operating manual and look at what features you have available to you for tuning the reader. Depending on the reader, you may have preset configurations that would be best used for a conveyor or a dock door, which can be changed for better performance. The better readers will have many parameters that can be changed one at a time to see whether that improves performance. These settings can include parameters such as protocol setting; reading Gen 1 Class 1, Class 0, Gen 2 Class 1, or all three; retry rates; antenna sequencing; listen-before-talk; dense reader mode; and many other specific tuning parameters.
If you've exhausted all of these possibilities and still are not getting adequate results, it's time to call the reader manufacturers and find out whether there are any known problems or changes required. It is not uncommon for your problem to be the same one faced by many other people, and changing firmware or performing a specialized test may reveal whether the problems you are having are something the reader manufacturer has seen before.
The wrong firmware can make a good reader fail miserably. Firmware is the reader's driving software. It tells components such as the antennas, the DSP chip, the processor chip, and the communication devices how to behave, when to act, and what to do. Firmware is the nerve center of an RFID reader.
The physical layout works hand in hand with the business process. If the reader is set up properly and reading when you go through the testing protocol outlined in the preceding section, you might have a problem with the business process.
The physical layout and business process can interrupt the read success in various ways. For example, a forklift may pick up a case of goods with the RFID tags or pallet tag oriented against the forklift blades and body. The metal against the tag would prevent any RF energy from getting into or backscattered back from the tags on that side of the pallet. The two simple ways to solve a problem like this are to train the end users not to slide the pallet directly up against the forklift, or to put in place a spacer so that the skid cannot slide directly against the forklift. Either is a simple solution, but the problem can be identified only by actually observing the process that happens during the normal day-to-day business operations.