Ok, so it’s no Godfather Part II, but this sequel still packs a punch. Dr. Bill Cardoso provides the star power for this Fireside Chat with the Xperts, as we continue to explore the important role of X-ray inspection in electronics manufacturing.
Dr. Cardoso demonstrates why he’s our MVP when it comes to understanding bottom terminated components (BTC) such as QFP, QFN, and BGAs on PCBs. Not to be missed, he shares one of his all-time favorite X-ray images of a BGA that is a one stop shop for common defects. Join us for other Fireside Chats with the Xperts here.
Dr. Bill Cardoso: Welcome to another Fireside Chat with the Xperts. Today, I’m really happy to be here with my colleague and good friend, Dr. Glen Thomas. And the topic for today is a continuation of the presentation we started last week, which is an overview on the basics of X-ray inspection. So the idea here should give you an overview of how X-ray technology works and the many applications where X-rays are utilized. So last week, Glen went over several different X-ray applications. Some of the basics and the physics of X-ray inspection, and we stopped last week at the bottom terminated component. So just to give you a quick overview, a bottom terminated component is every component where the leads or the contacts are under the component. So the bottom line is you can’t really see those solder points because they’re being hidden by the component itself.
Dr. Bill Cardoso: And got to a point in why they exist, right? Why BTCs exist? Well, as technology evolved and microchips got denser and denser and denser, engineers, component engineers, put the leads or the contacts on periphery of the component. So we’ve all seen the components. They started, they were rectangular and have leads or contacts on both sides. And as time evolved, the flat pack was developed. So you had contacts on the four sides of the component that were started to be square. There were some difficulties assembling them, but what happened is we got to a point where, as the Silicon Valley technology evolved, the density of those components increased. And as a result, the periphery of the component wasn’t enough to put all the content. So the component engineers at the time in the ’80s developed this ball grid array technology.
Dr. Bill Cardoso: So the pads ended up on the bottom of the component, and here’s a side view of that arrangement. So you have the components over right here, that’s all your silicone and all the electronic smarts are, and the interface of the world printed circuit board is done through these balls and contacts that are underneath the component. The problem as we talked about last week is that you can’t really see those pads anymore because now they’re being hidden by encapsulation of the component.
Dr. Bill Cardoso: So X-ray technology became essential in the manufacturing of electronic circuits, because only with X-ray you can see through the component and visualize these solder connections, like you can see here on the right. And with that, and why do you need to do that? Well, you need to make sure that those solder connections are there to begin with, that they are done properly. Do you have enough solder on each one of those contacts? And that those solder points are not touching each other, thus causing short-circuits. Well, we have… I would like to share with you the best and also the worst BGA or Ball Grid Array X-ray image ever, ever, ever. Why is the best and the worst? It’s the best from an X-ray perspective and from an understanding perspective, because this one X-ray image has every single problem you can find on a BGA assembly. It’s incredible.
Dr. Bill Cardoso: And it’s also the worst because it has all the defects you can imagine in one BGA assembly. So let’s walk around this image and go over what the problems are. The first one can see here is coplanarity. Coplanarity can be visualized once you have this tilted view of the BGA. So if you look straight down with an X-ray, you have this two dimensional view, and as you tilt the board, or you tilt the sensor, you have this perspective view of the BGA, and that’s what we’re looking at here. We’re going to go over more details, and we’ll call it two and a half D, which is a made up term developed by X-ray companies. But we’re going to go over what’s 2D, two and a half-D and 3D, but for now, just think that this is a tilted view of this BGA.
Dr. Bill Cardoso: And as you look at that row of balls on that corner, on that side of the BGA, you see that this ball is fairly round. And as you move towards the right side of the BGA, the ball gets elevated. You have this hour glass shape forming up here. So what this physically represents is that the BGA is like this. So that’s the PCB here, and the BGA is off, taking off from the board. So the coplanarity is a big problem because it means that you have… might end up with no contacts in one side of the BGA and short on the other side of the BGA. So what causes that? Once the BGA goes reflow profile, there is a wide range of CTE mismatches that occur during the process. CTE is coefficient of thermal expansion.
Dr. Bill Cardoso: So the CTE mismatch causes the potato chip effect. So the BGA basically can go like this if it’s a heated past it’s capabilities. So either the BGA or the board potato chip causing this effect. Or during the reflow the BGA can walk, and if not enough or a disproportional amount of solder has been placed on different pads, you might end off just by the capillarity of the solder as it wets, you might end up having this effect. Big problem. It’s definitely something that has to be inspected and looked at. Short circuits are fairly straightforward to recognize. So here you have two balls that have been a shorted together. A head-in-pillow is a difficult one to see and to find because head-in-pillow is that problem where the board, your assembly, works sometime. And sometimes it doesn’t work.
Dr. Bill Cardoso: Usually it works when it’s in your shop while you are making things. You test it and it works, you ship, and your costumer says it doesn’t work. Then you test it and guess what? It works. And because this is a problem where the ball sits on the pad, and it makes contact once in a while. It depends on… We’ve seen a situation, for example, where the test jig that was testing the board was enough to push the BGA down and make contact. As soon as released from the jig and shipped to the customer that contact open, and now the board doesn’t work. So you bring to your facility, your test on the jig and works perfectly fine because that jig is applying physical force onto the BGA to make contact. You shipped to the customer. You don’t have the jig, it doesn’t work. So it’s a very hard problem to diagnose because of this in terms of functioning, and it’s called head-in-pillow because it resembles a head on the pillow. It just rests and opens from time to time.
Dr. Bill Cardoso: Again, another defect mode that requires a tilted view and again, very hard to diagnose, very hard to find. But when you find it can signify insufficient reflow, because you weren’t able to wet the ball onto the pad. In a situation like this, since you have coplanarity, it might mean that we didn’t have enough physical contact between the pad on the board and the ball itself to properly wet and make the metallurgic contacts that you need to have a reliable solder joint. Moving on here on the side of the BGA, you have an open. An open is a good case scenario for head-in-pillow defect. What do I mean by that? It means that the open you don’t have a connection, no matter what the ball was sitting high, the pad is here, and you don’t have any connection. It’s a best case scenario from head-in-pillow from because you don’t have to worry about this on and off connection. It’s always going to be off.
Dr. Bill Cardoso: Splattered is something that happens when… usually signifies too much heat when you reflow and the bubbles that can be air or flux instead of balls, you just pop and splatter solder throughout the BGA. I mean, if you just look at this piece of solder and say, “What’s the big deal? Just a little piece of solder inside my BGA’s not a problem.” Yeah. It’s not a problem except that as you… it might lodge between the power and ground, for example, VCC and ground causing a short and causing a mess, malfunction. So you can’t… it’s not a reliable assembly and should not be allowed.
Dr. Bill Cardoso: Moving on here. Solder mask problems are not straightforward to recognize. A solder mask issue… So let’s go back. What’s a solar mask? So you start with a PCB. You have your board, and the board has copper traces and contacts, and land pads where the BGA’s going to sit on. Now you have a solder pad and you have a trace that connects to the solder pad. Now, if you just put a little solder on the solder pad, the solders going to run on the trace. And as we said earlier, you need that volume of solder so that you can keep the separation between the PCB and to BGA. So to avoid solder from running through the trace of each pad, we need to put a mask on top that’s called solder mask.
Dr. Bill Cardoso: So this mask that covers all the traces and just leaves the pads for each BGA ball. And what happens is that if the pads for some reason are… if the solder mask is defective, that solder is going to run through to the copper traces, and as a result you’re not going to have enough solder volume to keep the separation between BGA and the PCB. And as you have the discontinuing, this inconsistency of solder volume, you might end up with opens or even head-in-pillow. So that’s a big problem. It can be a defect on the PCB manufacturing itself, but it can also be excess heat in your reflow profile which caused the mask to lift. You have the peeling of the mask, and that allows solder to go like under a rug and potentially cause a lot of problems on that assembly.
Dr. Bill Cardoso: Voids are… it’s the Holy Grail of BGA inspection. That’s the very first thing you look for are voids. Too much voiding is a problem. Why? Because this… so what are voids? Voids are areas of the solder ball that are occupied by either flux or air, right? So it’s imagine a bubble of air, a bubble of a flux inside the solid object. So that bubble is, if it’s big enough, as the assembly goes through a temperature cycle, as things get hot and cold… imagine your car in the morning and then in the middle of the day. Can go from 30 Fahrenheit to 90 Fahrenheit in one day, which is a very aggressive temperature cycle.
Dr. Bill Cardoso: Well, that bubble has a coefficient temperature expansion. It moves, expands and contracts as a certain rate. The solid object around it, which is a solder ball, has a different CTE. And this guy here, this bubble is going to expand faster. So as things heat up and contract, you can imagine that now this solid object is going to be moving as well. And guess what? You don’t want to have this thing moving, because they’re going to create cracks. And eventually you can separate the ball, shear the ball away from the BGA or from the PCB.
Dr. Bill Cardoso: So voids are in some situations hard to avoid, but they should be minimized. And there are different standards. IPC 610 specifies a percentage of voiding on the BGA that you can have as a functional what the class of product you have, class one, two or three, which is a function of how reliable that assembly has to be. You can imagine that if you are manufacturing a toaster that’s going in your kitchen, the requirements are very different from a reliability standpoint then if you are making a toaster that’s going to go on an airplane or a submarine. So different electronic examples have very different reliability requirements, and voiding is a part of that reliability composition. The shape of the ball, as you can see here is a good indicator off wetting off the ball. So wetting, I don’t know if you guys… just go over. Wetting is the ability of a solder to connect and run on a specific metal. In this case, solder onto the pad of the board.
Dr. Bill Cardoso: You want to see a shape that is round or oval. That’s what we’re looking for on a BGA inspection. An odd shape like you see here in the image represent that… and as a result, you ended up moving on top of the solder mask and creating this odd shape. And finally you have a missing balls. So missing balls, as you can imagine from the name, is when the ball is just not there. And when you are attaching the BGA onto the board, the BGA has the balls already onto each one of the pads. And you can either put flux on the board or you can stencil some fresh solder onto the board. And then when you mate the BGA onto the board, you reflow the solder wet, and you connect. Situations like this often happen when the balls on the BGAs get knocked down accidentally because mechanical shock, because from picking up from the tray and putting on the board, in that process some of the balls got dislodged.
Dr. Bill Cardoso: We also have seen a counterfeit BGAs where in the process of counterfeiting the BGA, the BGA’s removed, it’s cleaned. And when new balls were attached or reflowed onto the BGA, the wetting wasn’t done properly. So the ball doesn’t really stick to the BGA and ends up falling fairly easily, causing a situation like this. So we have a whole session of this fireside chat about counterfeit detection, and we’re going to cover that at that time. But for now, keep in mind that if you see missing balls, it shouldn’t happen. Really shouldn’t happen. If it’s a brand new BGA, you shouldn’t see any missing balls. If there are missing balls, it’s very likely this BGA has been reworked at some point. By the way, feel free to leave any questions on the chat. Or if you want to raise your hand, you can get unmuted and I’ll be more than happy to address your questions.
Dr. Bill Cardoso: All right. So one of our pet peeves here at Creative Electron is that when people are designing boards, design electronics, these really should think electronics design in three dimensional way. Why? Because if you think about electronic design or when you’re designing a board, if you think about how the board is going to be inspected with an X-ray, it can save you a lot of money. Let me explain why, because as you add complexity to your… and overlap between the top and the bottom layer, as you add more complexity to the board, the complexity of the equipment inspecting the board is also going to grow. So you can go from a simple 2D X-ray system that’s going to cost you X to a two and a half D system that’s going to cost you more than two and a half X, to requiring a three-dimension, so computer tomography for a higher complexity.
Dr. Bill Cardoso: Let’s go over some of the examples where we’re going to show you how small changes on the design of the board can save you a lot of money in the inspection of the board once the board is built. So here’s a good example. It’s a 2D, two dimensional straight down X-ray of a ball grid array. And I want to point your attention to these balls here. You see this one here, that one there, this one, that one, this one. So you have several locations where decoupling capacitors, these components here are capacitors. By the way, how can you tell apart capacitors and resistors on an X-ray? Capacitors have a very dense dielectric, and as a result they are very radiopaque, which shows very dark on an X-ray image. So these dark components here, these are capacitors.
Dr. Bill Cardoso: The light ones on the other hand, resistors, or ceramic resistors are very light and not radiopaque at all. So they show basically transparent on an X-ray image, and you can tell they’re there by looking at the fillet they leave on the pad. So that’s a resistor, resistor, capacitor, capacitor and so on. So anyway, to minimize noise on the BGA, on the circuit on the BGA, you need to add a capacitor. Basically a low pass filter on your power supply to make sure the power supply has enough… too much ripple. And to minimize the inductance or the impedance between the pad on the BGA to the capacitor, what you do is you try to place the capacitor as close as possible to the balls, to the contacts.
Dr. Bill Cardoso: And what people end up doing, what designers end up doing is perfectly aligning capacitors onto the balls, which makes it very hard to expect with a 2D X-ray, because you look here, you can’t tell if this is a short or a capacitor. So it makes life very, very hard for the X-ray. Now with a 2D X-ray, you can’t make a determination if these are legit shorts or just a capacitor sitting there. So what do we have to do is to go ahead and tilt the board to get us a nice perspective, two and a half D. Now, this is a more expensive X-ray machine. You need more time. You need the right angle, better trained operator to make that determination. As you can see here, let me go back and forth between those two images. And you can see here is that people look at this, it looks like a big capacitor hiding several balls there. So as we tilt, you can see there was one short that was being hidden by that big capacitor.
Dr. Bill Cardoso: And you can actually see that some of this, as we tilt, these are legit shorts that are on capacitors. So it’s very important that if you can, what we’re suggesting here is that you shift to offset those capacitors if possible, outside the BGA, and bring traces in, or at least shift them away from the ball just enough that we can separate what’s a ball or a solder ball, and what’s a capacitor. So we looked at what a 2D inspection is straight down view with the X-ray imager. We saw what a two and a half D, which is tilting the sensor or tilting the table, the stage, to give you that perspective view, a tilted view of the image. So what’s 3D inspection. What is a computed tomography? So there’s a wide range of ways that you can collect this tomographic or 3D, three dimensional perspective on an X-ray image.
Dr. Bill Cardoso: This is a fairly straightforward one that is very intuitive to understand. So let’s go with this one for now. So imagine you have three objects. You have these blue objects, the pink and red. So you have a square, a circle, and a triangle. And this can be the ball of the BGA, the PCB, and a capacitor on the other side of the PCB, and the bottom layers of the same assembly. Now, let’s look at perspective one here. So this arrangement, the sensor and the source are going to make this movement. One is going to rotate clockwise. The source rotating clockwise, the sensor rotating counter clockwise. And in that arrangement you have many tilted views off this depth stack, the same stack of objects.
Dr. Bill Cardoso: So in perspective one, the square is on the top, the circle is in the middle, and the triangle is on the bottom here. That’s the perspective on the X-ray sensor of those three objects. As we move the source and the sensor, you can see how the square, circle and triangle separate differently. So they do this dance and they separate and come back together. And that’s a function of the projection of the source in the sensor onto that stack of samples. That’s pretty straightforward. Now, since, you know… you don’t know where those things are. You don’t know where your sample is. But you do know where your source is and your sensor. You know where your imaging training is at all times and exactly what angles and what projections your imaging at. And as a result, you can by doing some simple arithmetic on each one of those images, you can separate the top layer to the bottom layer of your board.
Dr. Bill Cardoso: And in this example, we have a QFP here on the top layer, and a resistor on the bottom layer. And that fillet of the resistor is masking a potential short between those two pads of the QFP. So by applying this 3D tomography, you separate the top and the bottom to have a clear idea of the top side and the bottom side, thus clearing this arrangement from a potential defect. So to make it very straight forward, these are the three different X-ray modalities available for inspection. 2D, which is a straight down view of your sample. Two and a half D, so that’s when your tilt and you have the perspective view. And finally, three dimensional where you get the full volume of the sample that you can move around, decompose, and provide with a larger insight, and complexity of the X-ray system grows as the number of dimensions grow. So with that we’re going to open for any questions now. We’re getting to the top of the half an hour here. Any questions?
Dr. Bill Cardoso: So we have a question here from John, which is… Okay. Yeah. So he’s asking, what is half a dimension? Well, half a dimension doesn’t exist? This is a made up term that the X-ray industry came up with several years ago to differentiate between 2D and 3D right? So you call 2D is a straight down shot of the sample, 3D is a volumetric interpretation where you have X, Y, and Z in formation. Dimensional Z information, not only density Z information on your sample. So instead of pixels, you have voxels on a 3D representation. And half a D is a tilted view. So it’s more for marketing term than anything else?
Dr. Bill Cardoso: All right, we have another question coming up here. This one’s from Dave. Does it matter where the voids are inside the ball? That’s a good question. Yes, it does matter. As you can see here, with a 2D imaging you can see what voids are, but you can’t really tell… you have a solder ball. You can’t really tell if that… imagined that this is solder ball and that a golf ball is the void. You can’t really tell if the void is on the top that interfaces with the board or at the bottom, if it interfaces with the BGA. You don’t really tell with 2D image, you’d be looking at a straight down shot. So by doing a two and a half view, you can tell that for example, this void is right on the edge and right on the interface with the board.
Dr. Bill Cardoso: So that’s a very bad place to have a void. And even if the void is below the threshold set by your quality group or IPC, whatever other standard are you going to use, even if it’s below that threshold of 20, 25%, that’s still a fail. It has to be rejected and reworked. So this is all the time we have for today. Thank you so much for participating and attending. Don’t forget next week, we have another Fireside Chat with the Xperts. Check our website to see exactly what the topic is going to be, but it’s going to be a Wednesday, 10 o’clock Pacific time. Glen. Thanks again for participating. And I’ll see you guys next week. Thank you.
In this week's Fireside Chat, we take the "X" out of Xpert. With their guest, Ross Williams of HES Solar, David Kruidhof and Dr. Bill...Read More
If data analysis is going to drive the next industrial revolution, then data fusion will be it's self-driving vehicle. Host David Kruidhof, Dr. Bill Cardoso,...Read More
The Xperts are back at it, with their first fireside chat of 2021. Who though surgical tray inspection could be so interesting, and the challenges...Read More