Allow me a contrary viewpoint. I've had these things on my floor and had customers get outsourced evaluations done with them.
Most expensive boat anchors I've run across.
First the good points:
The automated scanning and pattern recognition is good for sampling characterization of process shifts. That is it will capture variations in how some components or boards are wetting out, or solder printing skips. So will any conventional AOI. For RF, where solder joint geometery can affect performance, the 5DX does a better characterization of the overall joint shape.
Now for the reality check:
There is NO "laminography" involved. The terms "laminography" and "slice" that are thrown around by Agilient are the most deceptive advertising that I've ever seen. Most importantly, other than gross solder volume changes, the machine is blind and usless with BGAs. The very thing it is most often touted for. It CAN NOT discover a crack or fracture in a BGA joint. No xray machine can. But the operating principal of the 5DX makes it especially blind to BGA interfaces. Just run one in evaluation mode and translate the Z axis past the interface level. You will still see a round dot. That gets smaller the further away from the interface you get. This is the shadow of the widest part of the ball. As it unfortunately turns out the angle the system uses just about perfectly matches the form of a BGA joint. This system was developed before BGAs and this angle was fine for reducing the image density of backside components.
This is what the machine was designed to do. There used to be a little flash video on Agilent's site that showed how it worked, but they've d it. By angling the beam/detector arrangement, and spinning it, anything in the center of that angled rotation appears still. As you move away from that center (what Agilent calls the "image plane"), anything that is imaged rotates in the field of view. The further away, the larger the circle it rotates in and the blurier it gets. The image gets distributed over a larger area and therefore is less dense.
So, what you get in real life, are two cones with their bases together at the "image plane". Anything within that shape is obscured and the machine is blind to it. Including BGA solder interfaces. Remember, the BGA ball looks just about exactly like those two cones with the tips chopped off. But the machine can't see anything within those cones.
A marginally useful production screening device at a high cost. But very limited in evaluating BGA's and doing any FA.
reply »