Meet the man who will decide if you are Gen 2 compliant

(This article originally appeared in a 2005 issue of RFIDOperations)

WHETHER YOU ARE selling or buying Gen 2 products, the readers and tags under development are about to be put to the test. RFID Operations decided to speak to the man in charge of the testing, Gaylon Morris, director of RFID programs for Met Laboratories, a company in the testing field since 1959 that was selected by EPC Global to handle this program.

Morris, 32, has been with Met Labs for almost six years, working on electrical product safety and NEBS testing. Editor Andy Kowl spoke with Morris about the plans for certifying the newest generation of RFID technology.

When do you expect testing will start? Do people have appointments already?

Morris: People do not have appointments yet. The short answer is I expect it to start in June. The long answer is we are waiting for submissions from the actual hardware manufacturers so that we can validate our test system on multiple samples as part of the test system validation. Until we get those samples to do the validation, there’s a certain period of time which may be as short as 5 or 6 weeks but most likely will be as much as 10 or 11 weeks before we’ll be able to turn on testing programs. If I got three samples today, it would be between 10 and 11 weeks from today.

How do you know that the stuff that first comes in doesn’t skew the testing procedure in case those samples may not work correctly?

Morris: We could right now, with fair confidence, test a device fully. But we would be doing it very slowly. We would be doing it by actually looking at the RF signature of the communication between the reader and the tag. In order to make that a system that can handle the volume that we’re anticipating, it has to be automated.

What we need those samples for is to first qualify those samples in a very slow method. But then we need to take those samples and validate an automated system. Otherwise it would take weeks. The Gen 2 spec has so many permutations even in its required state that it would take too long to do it manually. And be, at the least, efficient.

How many different items require testing to be validated?

Morris: Primarily we’re going to be looking at readers and tags.

I understand that applied tag performance and accreditation program would be a different form of testing. You will be testing tags, correct?

Morris: Yes. But, just because a tag is Gen 2 compliant doesn’t mean it’s the right tag to put on this box of whatever. So, there’s another whole program for basically did you pick the right tag?

Must a company belong to EPC Global in order to be tested by you?

Morris: I believe so. The testing is actually contracted and scheduled through EPC Global. So, whatever filter mechanisms they put in place, it’s completely up to them. And my experience with EPC Global has been membership is typically a filter.

Describe the environment of the lab. If I were to walk in, what would I see?

Morris: You’d see two copper screen rooms; one completely solid metal shielded room to allow us to isolate ambiance, because we may be testing multiple tags or multiple readers at one time. So you don’t want any interference. We have what are called cages, or screen rooms. The cages are basically metallic cages and the idea being that signals go in, no signals come out. Not 100% perfect but its pretty good. I know what ambiance I’m dealing with. I know what cell phones and what wireless LAN broadcasting, we know those things. The fluorescent lights are one of the biggest problems for UHF/RFID systems. And so we have looked at that and we know what we have going on with those as well.

So, now if someone comes and makes an appointment, is there a standard charge involved for every test?

Morris: I think so. I have a standard charge base to EPC Global. I know they have the ability to charge in any number of ways. I don’t know how they’re going to do that. I haven’t seen their business case model.

I see you get paid by them. That’s what you know. You have your own thing.

Morris: My contract is with them. All my services I provide are to them. I may be testing an Alien tag, but I’m doing it for EPC Global.

Is there any other obligation that a client who wants their material tested?

Morris: For hardware conformance, if you make silicon tags and we’re going to test that, I will need to be supplied some sort of test jig that has some sort of a connector, for instance an SMA connector. And I need to be able to connect my tester up to that in a conducted way. So, that is something that you would have to have.

Another thing that you would have to have if you make readers, or interrogators, is you would have to have a test mode on your software that allowed me to manipulate a reader, because I can’t test it by allowing it to just act organically. It has to go through and be structured for what we do when we’re doing our testing. So I have to be able to control it.

When we’re testing the tag silicon, we’re not testing the antenna inlay. We’re purely interested in the protocol side of the tag itself. Testing the inlay with the antenna element attached to the tag, is interoperability testing. And while we are doing that, that is a slightly different aspect of the same program.

So if I have a tag to be tested, you’re only testing the chip?

Morris: Kind of. The ecology would be, if you are a silicon manufacturer such as TI, Phillips, Alien, Impinj and a number of other ones, then we are going to perform conformance testing on the silicon chip itself to verify that it meets the requirements of the Gen 2 spec.

If you are one of the many companies who buys silicon from them, and takes that silicon and buries it into a tag that has an antenna element, then we would need to perform interoperability testing to verify that your device interoperates properly with the antenna in place, because that changes the RF aspect of the tag.

So the company that comes in just gets the silicon element tested, without the antenna? If that company also makes their own tags, then they have to get two tests done. Would that be right?

Morris: Yes. That would be right. Now depending on the difference between two or three inlays, we may be able to qualify inlays by similarity to each other. But at least some amount of inlay testing would have to be done because it changes the RF signature in the tag.

Are you testing tags with one sort of absolute testing device or are you testing against a variety of readers?

Morris: Keeping in mind that the test equipment is still under development, the current structure that’s being looked at is a test system. It may end up employing some aspects of an interrogator, but it is going to be a test system. It’s not going to be a bank of readers that we test a tag against and if it works with all of these, it’s okay. That’s the Wi-Fi concept. And it’s not the same kind of concept that EPC Global has chosen to move forward with.

So if it tests okay, then it means in an absolute, theoretical sense it’s exactly correct; and therefore it should work with anything else that’s also correct in an absolute sense.

Morris: “Should” being the operative word. And we perform the interoperability testing to verify that it does.

When you’re testing a reader, is there any relationship of the reader to either edgeware or middleware? Must it come with some?

Morris: What we will be asking is does the reader have a test mode where we’re going to interface with it with our software. In that test mode, we would be able to manipulate the reader to perform certain functions so that we can verify that it’s performing as it is supposed to according to spec. There is an edgeware, middleware type relationship that is being dealt with by a different standards group within EPC Global. And again, I expect testing for that to kick off in October.

Got it. But otherwise you’d be testing the reader not in relationship to any particular software, but in relationship to the parameters that you’ve set.

Morris: Yes. It’s kind of similar to a webpage. You don’t know what software and what background that web page was written in when you look at it – click on a link, it should work. Click on a button it should work. Same type of concept. Regardless of how the different manufacturers base their reader, it has to have certain functionality and we’re just going to verify that functionality.

It’s my understanding that Underwriters Lab buys the stuff they’re going to test “on the street” from suppliers, so that its not like the perfect item that has been gone over six times at the factory to make sure that it is absolutely perfect. How do you prevent that from happening?

Morris: You can’t.

I see. In your experience, how common are variances between readers from the same company?

Morris: The variances that we might see would be limited, I believe, like to oscillator tolerances, frequency tolerances. I know people in different fields who will take a bank of devices and test internally each one of them until they find the best one and then take that to the lab. But the variance between the best one and the worst one shouldn’t be very much.

Is there a way to test that easily? We could put a surveillance program together and that would kind of be dependent on whether or not the industry thought it was necessary. If we test a bunch of readers and then three or four months later, the readers start showing up that have been part of this tested family but aren’t working properly, then maybe surveillance would be in order.

You said that you aspire to be fully automated during tests. How many people are present?

Morris: One person – but the idea is that he or she need not be sitting there programming the reader-based step, or programming our material to work with the tag at each step. They view what’s going on and view the report for any nonconformance. And if there’s any nonconformance, they are educated and trained to the extent where they can deal with those.

The way you deal with the nonconformance is you verify your setup. You verify the equipment under test setup. You verify the test was run properly and then if all three of those things are the way they’re supposed to be, you have legitimate failure.

At that point you contact the client and you let them know, “Hey, I have a legitimate failure here. Here’s what I’ve done. Here’s what I see. Here’s what the result is. What do you think? Where do you want to go from here?” Depending on the failure we may offer suggestions. It depends on the aspect of the failure.

And you put this directly to the client, or to EPC Global?

Morris: That’s a good question. I will interface with the manufacturer of the equipment. Especially when it comes to things like nonconformance. Whether or not the final compliance report has to go through EPC Global to get to them, I don’t know the structure of that yet.

Is this all confidential? If some client doesn’t choose to have their results announced, is it up to them?

Morris: It’s up to the deal they and EPC Global put together. I will release no data except to EPC Global or the person whose equipment I’m testing. What those two people do with it, is out of my control.

Is there an acceptable failure rate in the sense that if someone gives you something and says, “Oh my goodness, I gave you the wrong one, try this other one?”

Morris: As a rule I ask that we do not retest something unless we can be shown how it is different from what we tested before. So, you bring a device in and it fails any test, even though you might be willing to pay for it, I’m hesitant to have you throw me an identical sample and say test this one, too. But if we get a different result, now we have to start picking apart your product to determine why it resulted differently. And it never ends up being the situation.

Do you get tested or inspected?

Morris: Oh yes. We are accredited by OSHA. We are accredited by the American Association for Laboratory Accreditation. We’re accredited by the Standards Council of Canada. By IECEE which is an international safety board. We’re probably accredited by 20 organizations. Verizon accredits us. There’s quite a bit of different accreditations that we go through.

Now, do any of those accreditations cover the RFID testing program? No they don’t. That is a contract obligation between us and EPC Global. We do operate the program within our accredited quality system and methodologies; but there’s no accreditation for this type of testing yet.

Do you tend to get pressure from clients or any interesting stories you have to share about that?

Morris: My favorite is its 8:30 Friday night and the customer’s representative will tell me, “If this isn’t compliant by Monday morning, I lose my job. I’ve got kids. How can you fail this?” I get that kind of stuff.
Those types of stories (about being pressured) are rampant in this industry.

Do you have any tips for companies who may be getting ready now for Gen 2 testing?

Morris: The EPC Global Hardware Action Group has a certification work group forming right now. If you’re an EPC Global member, you can opt-in. I would certainly be involved in that (if I were planning to get my products tested). Because we are going to discuss the test modes that readers need to be able to go into. We’re going to discuss the actual test cases. Somebody who’s involved in that group and has access to all that paperwork, may have a very good advantage.

A second thing I would recommend would be to allow EPC Global and MET to use your hardware to validate our test system. Because we will supply some preliminary results back to those people who give us test equipment. Those who give us equipment to verify our testing will then have some results in advance. They’ll be preliminary results but they’ll have results in advance of many of the other people in the industry. So there is definitely an advantage.

If someone has a set of the Gen 2 standards, and decides to devise their own testing, though it is not certified by MET Labs, is there any reason to think that it is less valid?

Morris: No. Let’s take Bluetooth for example. There are at least four or five validated test systems that test all the aspects of Bluetooth from different manufacturers. Some of which do it in different ways. They’re all completely valid.

Now, if somebody wants to develop a test program in alternative, there’s a very good chance that MET and EPC Global would be interested in absorbing that methodology as well and making it a legitimate methodology. I can’t think of a reason not to.

What are some of the most common problems you might anticipate?

Morris: The problems I’m going to run into from a hardware conformance standpoint is, the way the standard works, that the reader broadcasts a signal, a query to the tag, the tag responds with a random number sixteen digit bit number. And the reader has to decode that and transmit it back to the tag very quickly. Literally like 20 microseconds or something. Very small. And one of the biggest problems from the testing standpoint is, without actually having a reader, how do I decode that number sixteen and shoot it back in that short a time? That’s faster than your typical computer operates because they’re using very specific and focused technology. So that is a problem that we’re dealing with on the test development side.

On the applied tag side, the biggest problem I think that’s facing us right now, other than using the right tag, is out of sequence reads. If you’re broadcasting within a 9 meter range from your reader, and you have 15 tags on that conveyer in that range, how do you know they’re reading in order? When you get 20 reads back, you need to know which one’s the first one and which one’s the 20th. So you can shoot them off that conveyer for distribution lines. If you don’t know that, then you can’t replace the bar code.

Do you test the reader’s ability to shield or to not get duplicate reads?

Morris: Well, on the applied tag side, yes. There will be. We will be looking to determine what is the right amount of energy and what is the right setup for a tag to not be read out of sequence?

The biggest problem facing the industry, I think, is frequency allocation. It’s countries only allowing one or two or three megahertz of bandwidth. Different, completely different ranges. I mean, what’s legal in the United States isn’t legal in Asia, isn’t legal in Europe. Why is that a problem from a test standpoint? In order for me to test the device to make sure its going to operate properly in Europe or Asia or the United States, I have to test it as though it’s in Europe or Asia or the United States. And if I can’t legally test it in that way, in a lot of different locations, multiple locations within even one company, like a Proctor and Gamble, then how can I set up the supply tag lab and know that what I tag is going to read everywhere?