Reliability Testing in Automotive and Digital Factories
Barry Matties spoke with Gen3’s Andrew Naisbitt about what readers should know about SIR and CAF testing, what to consider if you’re thinking of bringing it in-house, and how reliability testing fits into automotive and the digital factory.
Barry Matties: Andy, why don’t you start with a little bit about Gen3 and what you do.
Andrew Naisbitt: We manufacture equipment that tests reliability in electronics, so we focus on aspects such as solderability testing, surface insulation resistance (SIR) testing, and conductive anodic filament (CAF) testing as well as our ionic contamination testers. We also have our Gensonic stencil cleaning center that has proved to be very popular in America.
Matties: You’re selling the equipment, and a lot of the services would be your largest customers. Do you see OEMs?
Naisbitt: It’s quite mixed. We have a wide variety of customers and we have 41 distributors looking after our stuff around the world. We sell into all sorts of different industries and we’re finding ourselves a lot with some of the large board manufacturers. Reliability in testing is so critical in the industry, and we do a lot to help with that as well.
Matties: Reliability is a huge and growing issue as we see with autonomous vehicles, aerospace advancements, etc. If you were to tell somebody what they really need to know about reliability testing, what would that be?
Naisbitt: There’s quite a significant amount. The critical nature is always going to be in the circuits and how reliable the circuits are. You need to know if it is going to last. The test methods that we build for are for hypercritical stuff used by companies from aviation to automotive. There’s a lot of work on standards being done right now and changes being made; we need to focus our time and attention and know where the industry is going. There’s going to be more of a need for SIR and CAF testing with high-voltage testing, especially for automotive. You should ensure that those tests are done in the right way to enable you to have a far more accurate understanding of how reliable your products are that you’re building on the boards.
Matties: If someone thinks they need reliability testing, to set it up and bring it in-house, what do they need to know?
Naisbitt: First, you need to validate your products, so make sure you have a benchmark. With J-STD-001, there has always been a number of 1.56 micrograms of sodium chloride per centimeter squared on your board, so that has been the pass-fail limit. If you’re looking into the reliability, that limit is now gone. The pass-fail limit has been rendered obsolete in the standards. You need to make sure that you validate your process in terms of what numbers are you going to get from your boards or what are you going to do with the terms of the reliability. Then, you need to use process control tools to enable you to reliably reproduce the results that you get. There are two essential steps: validation, and then making sure that your production line is using the data that you get from the validation.
Matties: As companies consider digital factories, and they begin to come online, how does testing fit into that?
Naisbitt: We have a lot of customers who ask us the same sort of questions. The pass-fail limit number I mentioned has always been on the shop floor. What we do has traditionally always been in the labs, and we provide a lot of data. There are all sorts of graphs. It’s quite analytical and very different from having a red light/green light situation. We’re finding more and more and looking at what we’re developing with our products moving forward. New tests might come out, but it’s going to take a long time. The industry would appreciate another yes/no question on the shop floor, but whether or not that’s a realistic scenario, I don’t know because the testing doesn’t really do that at this moment in time.
Matties: When you consider real-time testing as the board is produced, it’s a good or bad kind of thing, in simple terms. When you look at reliability testing currently, do they test every single board or a panel from a lot?
Naisbitt: With SIR and CAF testing, they test coupons, which are representative of their final product. But the future of it will need to be production boards as well. As I said, there are different test methods that we’re developing. It’s an exciting time. I can’t say too much in terms of what we’re looking to move toward but based on where we think the industry is going with electric vehicles, etc., testing will change and need to become quicker and easier. Trying to balance science and what people want can be tricky.
Matties: When you look at autonomous vehicles, reliability is critical because it’s not just one passenger that could be injured; it’s potentially groups of people.
Matties: So, it better be right. I would think that the OEMs in such situations would want every board tested if possible, in a cost-effective way.
Naisbitt: From my point of view, it’s both yes and no. We’re talking hundreds of thousands of millions of boards being produced.
Matties: That’s where I’m looking to the future. If you’re looking at an inline situation and it was available, what would you say?
Naisbitt: There are so many different tests and ways to measure reliability. We only do a small fraction of the ones out there, and the processes that companies have at the moment, the likelihood is there’s nothing wrong with them. But with more development going into smaller and smaller components, you have to look at your testing, including how to do it and where to improve. And it’s the same for us as the manufacturer of those pieces of equipment and the tests and types of testing; you have to move with the time. In the future, maybe you will test the board, but maybe there will be a test that makes life even easier than it necessarily is right now.
Matties: That’s what you’re working toward.
Naisbitt: We have some good ideas floating around.
Matties: With the digital factory, lot sizes of one, and recipe-driven manufacturing, that makes sense. But I don’t know all the economics behind it from an investment cycle time process.
Naisbitt: It’s such a huge topic, and we’re making significant industry advancements. It’s a big change to go from fossil fuels to electric vehicles, and they’re looking at developing some crazy things with airplanes, so who knows where we will go. But there’s so much education that needs to be done, and people need to learn it’s not as simple as some might think.
Matties: It’s interesting that you have mentioned electric cars. I was recently at the Porsche museum in Stuttgart.
Naisbitt: Picking up your car? (Laughs)
Matties: It was nearly ready (laugh). And they had the electric motor that Porsche built in 1900 as part of a hybrid vehicle on display. When we start talking about electrics, a lot of people think this is a new technology, but it was built in 1900 by a 24-year-old. We’re refining it, but there’s a lot of history behind this.
Naisbitt: Like with anything, there are ideas, and then there’s having it on the mass scale that the automotive industry is at, which are two very different things. Even with batteries, lithium is a finite resource, and as far as I’m aware, lithium is the main factor of what is used.
Matties: Talking about autonomous vehicles includes bringing in electronics and drive systems—the computer aspect.
Naisbitt: And once they’re switched on, they’re never switched off. Life cycles now are going from 15,000 hours for the traditional lifespan of a fossil fuel car, and the warranty period that the manufacturers demand is seven years, up to 150,000 hours of a lifespan expected last year and 14 years warranty with electric vehicles. That is such a huge jump, so the traditional testing methods are going to change because we’re talking about much longer reliability. As I said, the electric car is always going to be on in some capacity.
Matties: The pressure for reliability is only going to grow.
Naisbitt: As you correctly stated, these manufacturers need to have an ROI on the billions of pounds that have been poured into it. But you can’t make people buy electric cars either, so do people really want to go down that route?
Matties: The green aspects are still there, but they’re doing a great job of making electric vehicles go from zero to sixty miles per hour in two or three seconds.
Naisbitt: I know; that’s fun. And there are millions and millions of cars on the road around the world.
Matties: And they are operating in extreme conditions.
Naisbitt: Yes, and this may be a bit extreme, but potentially 30–40°C to -20°C.
Matties: Or waking up in Arizona when it’s 80°F out and having the temperature to 120°F, for instance, but cars go even higher.
Naisbitt: And if you live in Alaska, your car is going to be subjected to negative temperatures for the majority of the year, which is a huge challenge.
Matties: I was on a flight coming back from Heathrow. They have GPS and metrics. The air temperature when we took off was 23°C, and when we were airborne, it was around -52°C.
Naisbitt: The good thing is that you know it’s coming. With aviation, you know it was going to get cold when you’re in the air and warmer when you land. It’s not as variable because you know it will happen, but on the roads, anything can happen; there are so many more variables and events that can occur. But then when it comes to driverless technology, how do you manage the switch between driverless and non-driverless cars. Do you introduce a law banning all non-driverless cars one day? I highly doubt it.
The pace of development in the industry and world is incredible. The cars from the ‘50s to the ‘80s were pretty basic, and now we’re entering the stage of autonomous vehicles. People have grown up on the classic petrol cars, which moved to diesel and is now moving to hybrids, all in a very short timeframe.
Matties: Is there anything that we haven’t talked about that you feel we should include in this interview?
Naisbitt: Again, at Gen3, we’re doing a lot of development work to evolve with the industry. At productronica, we released the new software for our AutoSIR2+ and AutoCAF2+, which are real step changes and improvements. A big thing with our customers being able to interpret the data and get the reports from the measurements that they’re taking and our new software is going to help that.
Matties: Does it organize the data in a user interface?
Naisbitt: Yes. We’ve got a whole new user interface. It takes the data in a whole new way. One of our distributors says that their customer wants to do four months of testing, which is 4,500 hours. We have new report features that enable you to quickly access, read, and interpret your data and easily save it to whatever you want, such as CSV files, Excel, Word, etc.
Matties: How do you decide what your development projects are going to be?
Naisbitt: We’re fortunate enough to have a lot of close ties with some large companies combined with standards and our knowledge. These companies are pushing that development and want to get ahead of the game; we have good relationships with them. Since we are a manufacturer as well, we can build on the ideas they have.
Matties: You’re getting direct input on a unique level from the standards point of view too.
Naisbitt: Exactly. We get input from different angles, including internally, with our engineers who are constantly improving the products as well. There’s lots of work on that side as well.
Matties: Thank you, Andrew.
Naisbitt: Thank you very much.
Visit I-007eBooks to download your copy of Gen3's new micro eBook today:
The Printed Circuit Assembler's Guide to...Process Validation