Will Ultrasonic Welding Damage Chip Capacitors or Other E-components?

Hello dgankey

I am presuming that it is always the same chip capacitor which is damaged, if not advise please.

The ultrasonic vibrations have probably caused resonance cracking inside the chip capacitor.

It may be the particular frequency that your ultrasonic welder operates at which causes the resonance.

If it is only the one part = The chip capacitor which is failing, then perhaps a size change of the chip capacitor, or a maker change of that chip capacitor will avoid the resonance, and enable the production of your circuit board/housing problems.

There is an alternative: Change the frequency of your ultrasonic welder, not easily and reliably done, and may result in another different circuit board component/s failing, instead of the chip capacitor failures you have experienced.

I would be inclined to use a chip capacitor of larger μμF, which would result in a lower natural physical resonance - But if the chip capacitor is used as part of a timing circuit, you need to keep it the same electrical size.

Do you need to actually bake the assembly, as there are room-temperature encapsulating resins, which may not add any thermal stresses to your board and components.

Advise further with

Kind Regards....

Did you check to see if parts failure was before encapsulation?

There are potting resins that cure without external heating but might be to slow for your production.

Which other 10% parts failed?? does the chip capacitor cause the failure of these parts?? How are the parts failed??

Encapsulation with a strong resin is known to cause failures, too. So you have to do step-by-step evaluation to sort it out.

thanks.

All ceramic parts are exposed to failure when ultrasonic pulses of "lethal magnitude" are applied. Especially quartz clocks may be damaged - mostly not complete but they change their resonant frequency.

it seems it's not like this case.

The problem is not in parts which fail in test, they are easily weeded out, but what about parts which have been damaged, work in test but display "early life failures" when installed! What will that cost your company to fix.

I would personally say that ultrasonic welding and electronic components of the type(s) you are using is maybe not a good idea at all......

A few other points that you should check on before blaming anything:-

1) were the PCAs fully tested before being placed in the plastic box, if not, your process may be OK, but 10% of these caps were already defective when they were mounted!

2) Did you test the PCA after welding but before encapsulation? That would point to the Ultrasonic process if the PCAS were tested before the Ultrasonic process...

3) If the failure is after the Epoxy process, then you would just need to change to a cooler Epoxy....

Till the process runs failure free, you must check the PCAs at each stage. But even once the process is working correctly, I do feel that the PCAs should be tested before being placed in the plastic boxes at all times......simple QC.

You are right. Now we are meeting this problem - "early life failures" .

you may find my #11 reply. The most difficult case is that we just found the failure at final function test.

Rhabe, you`re absolutely right with your resonant frequency - we once made a finite element simulation of bond wires with the result that those bonds were pulled with forces of up to 150 grams (!!) if they only had the right length.

Good post. I rated it as such too...

Qn.

Why are you welding the enclosure?

Why not go and "O-ring" the enclosure with a tounge-groove arrangement, and some clips that latch when the thing is put together?

Have you looked into inductive heating?

What about if the mounting you have used is too rigid and its twisting the PCB?

Is the unit being tested before welding, at the point when the unit is assembled, and ready for welding?

This last one, the cracking problem could be due to assembly of the device, where the PCB is stressed.

Have you done a FEA (Finite Element Analysis) on the assembly? done a test for the stresses during manufacture, testing, assembling, welding, thermal stresses, resonance tests.

Imagine spending thousands of dollars to find out that the problem wasn't what you thought it was.

My job here, is to design test systems for the PCB industry, and FEA is a important part of the design process, and the chip Caps are generally the problem.

You could end up redesigning the board to rotate those caps to a location where they don't crack, they could be too close to the mounting locations, or support locations.

One customer had a unit that needed TCIL testing (inline test cell) there were 4 units on a biscuit 100x3oz probes under the board, and a large plastic bracket on the top side, so you couldn't press the board fully (counteracting the probe forces) they had cracking capacitors, some found their way to the customer as the cracking generally happens in the test transfer stage, not in the testing stage itself

Before suspecting the ultrasonic welding, you should do a controlled study where you do a pre-test of the circuit, then test after performing each of the processing steps.

I have never had a problem with chip capacitors or any other surface mount components when using ultrasonic welding. Won't say it can't happen but didn't see it in the 11 years I was involved with producing a product that required that type of welding and we used alot of ceramic in the design. From the ceramic capacitive sensor to the ceramic hybrid board that held many ceramic components. Being an OEM supplier to the automotive industry, we do extensive vibrational testing with our product and have had solder joints and welds fatigue at certain harmonics but never had chip capacitors or any other SMT components fail due to harmonics.

You might want to check if the heat curing is causing an existing problem with the batch of chip capacitors to worsen. I once had a problem with resistors failing and found that we were introducing micro-fractures in the resistors during processing. We were damaging them during lead-forming, then after potting the sensor into the outer housing and performing a 1 hour cure, we were getting about a 10% failure rate. After rerunning the passing product through the oven again I had another 10% fail, so I ran them a 3rd time and even more failed.

In this case, the resistors were leaded and it was the alignment of the dies that caused the damage, but the symptoms are similar. A controlled test group should isolate the problem.

Our USA factory have nearly the same process but they don't have this problem.

Now we are changing the welding fixture.

Thank you very much. I am late.

12 days ago, Chinese's day - Spring Festival.

I will try my best to reply your comments. But this tough problem is not fixed yet.

let you know the process: Assemble PCBA in housing -> Ultrasonic Welding -> function test (temporarily added) -> Epoxy Encapsulation ->Assemble Terminals -> function test. At final function test, we even found the intermittent fail part. Per our analysis, we found the outside lead of some capacitor were broken and resulted in failure. But more parts we can not analysis because the epoxy encapsulation, the whole PCBA has been encapsulation. The worse case is that the reason of resulting in function fail is not just only one, although most of the case we knew is caused by capacitor. In the mean time we are in the process of reworking the ultrasonic weld tooling to provide a lighter clamping pressure

-

Hello again dgankey

Trust you enjoyed your Spring Festival.

Have you also considered encapsulating in a flexible compound, like a silicone rubber material, instead of your existing epoxy system?

It may give reliability, with damping of mechanical vibrations, and no thermal fatigue of connections.

You could try some, and see if that works for you.

Kind Regards....