Chip Component Cracking During Pick/Place

Phil
Welcome to The Board Talk. This is Jim Hall and Phil Zarrow, the assembly brothers, who by day go as consultants with ITM Consulting, but here on Board Talk to help with your assembly-process questions.

Today's question comes from D.C. He says, "Occasionally we notice chip resistor cracking during placement. Most often the cracks are in the center of 0201 and 0402 size chips.

Wow! How the heck do you see them? "Does your experience indicate this is typically a placement equipment adjustment problem, excessive force, or could it be a component design issue?"

Well, let me start off by saying there has been occasions in the past, not frequent, where components do come in cracked, so that even nominal placement or any handling can crack them. They might have been co-fired incorrectly, but how would you inspect them? You could have some poor sap in incoming looking at the reels, especially 0402s and 0201s, looking for cracks, well, hey, everybody's gotta do something.

Jim
Though you would probably see cracks of this type universally across the reel of components. So you could look at the history of the failures and determine if it's a materials problem.

Phil
We're assuming that you have inspected these components after placement, prior to reflow, and prior to singulation, so eliminating other causes of cracking, in that they are indeed, from what you see, a mechanical crack.

Jim
The first thing I would recommend with chip handling in any placement machine is to look at the nozzles. Make sure you have the right nozzles, that they're not worn, and so forth.

Worn nozzles have been a source of many problems over the history of chip placement. Haven't necessarily heard it associated with cracking, but it's a good place to check it.

Learn how your specific placement machine handles z-axis control. Does it place strictly on pressure? Does it place on absolute z-axis position relative to a zero point? How does it establish that zero point?

Whatever your machine does, make sure you understand it, and that you're programming the machine correctly. Assuming you're doing that, then I would look at the board.

Board warpage is a common problem. In this case you would expect that it would be a bowing up such that the chip in question, where the crack occurred, was actually located on the machine in the z-axis direction slightly higher.

That would be a particular problem if your machine was placing on z-position rather than pressure. A machine that places on pure pressure should be more tolerant of warpage in the board

Warpage, in the board in any of your operations, is a potential problem, so you should try to minimize it. Always use proper underneath board support in any of your machines, whether you're printing, placing, inspecting, doing ICT, dispensing - always use proper board support.

Anything you can do to minimize bowing in general would be a very valuable thing that might help you in this situation.

Excessive force can crack a chip.

Phil
Sometimes with these very small chip components it's very difficult to see, but if you're having excessive force, one of the symptoms you should see are solder balls.

As you come down in the z-axis, rather than placing the component approximately halfway or one third or two thirds of the way into the solder paste deposit, you're gonna be bottoming out and, no doubt, splattering solder paste.

We're talking about very, very small components so it can be very challenging.

Beyond that, this is Phil Zarrow and Jim Hall.

Jim
Whatever you do, don't solder like my brother.

Phil
And don't solder like my brother.

The issue of cracking chips could be related to all three of the proposed questions, probably less related to component design and more related to component quality. Placement system component definition profiles should match actual measured components. Mis-matched components/definitions will cause placement problems to some degree as noted in other discussion.

Independent studies have shown that same protocol testing of different component suppliers yields dramatically different impact force results using the same size resistors or capacitors. Capacitors are more 'delicate' than resistors.

Certainly excessive force can come from an equipment adjustment problem. Every OEM has a different spin on how their Z-axis force is measured or adjusted. Real measurements are needed to understand the Z-axis system operation.

It takes a dynamic placement Z-Force test to understand what forces are realized during the placement process. A statistically significant number of samples should be captured for each placement head to fully characterize the Z-axis performance.

Visually cracked components are indicative of something out of whack; could be related to many things. The real quality problem facing small chips like 0402's, 0201's and 01005's is micro-cracking. This is where a very small crack starts in the component body but still electrically connects the end caps. The component will final/functional test fine, but after a few weeks/months in field operation, thermo-cycling propagates the crack to full failure.

The micro-cracking comes from too much impact force. As smaller chips become more predominant in SMT manufacturing, additional quality measures need to be considered to prevent visually undetectable defects. AOI systems will not be able to see a micro-cracked component defect. What happens next?

The original letter didn't mention how the parts are distributed, but that is very important information.
Cracks are nearly always caused by excessive stress, and stress is inherently spatial. So to ID sources of cracking issues, you need to consider spatial factors. A scatterplot is a good clue generator. Is the defect rate higher in various regions of the PCBA, or higher on one side than the other?

Some sources of non-uniform failure rate distributions a scatterplot can help to uncover:
- Under board support tooling, if present at a higher rate on first side processed (overlay support locations against map of failures)
- ICT Fixtures (out of plane or larger unsupported areas)
- Proximity to mounting features in end-product (if very near mounting holes or support features in final product)
- Proximity to edges (encountering obstructions during complex installation into enclosures, or sigulation processes)

Stating that this problem occurs with caps, could we assume it does not happen on resistors? If this is the case, the machine is not the problem.

I used to work for a company that built circuit board assembly machines and we had this problem. What we found is it is not usually a machine hardware problem but a machine software problem. Resistors across different manufacturers tend to have the same thickness, where capacitor manufacturers tend to have different thicknesses. Example: if a cap is programmed for .8mm and you place a 1.2mm of the same value, length and width, then it is possible the Z-axis will travel farther than needed and crack the cap.

Check the cap thickness that is breaking and the program for that cap. A machine thickness measurement, by camera, might not catch the problem if the tolerance for thickness is set too high.