Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

smd capacitors- ICT measuring

Views: 4686


smd capacitors- ICT measuring | 12 June, 2013


i have problem with ceramic capacitor 10uF ±10% 16V X7R 1206.Problem is as follows:ICT ( Teradyne 8851 ) measured value 6 ?? 8uF ( about 70% of all lot).I decide that components ( samsung ) are defective.I buy one "type&reel" capacitors "Murata" , and one "KEMET", but situation does not change.All other capacitors on the PCB are OK.I think , that may be some ICT settings are wrong???Does anyone have asimilar problem?

reply »


smd capacitors- ICT measuring | 13 June, 2013

Have you verified the measurements on the bench with an LCR meter? Chances are you are experiencing aging with these ceramic caps. We have similar issues with X5R and X7R caps. You will see the effects of aging with returns from the field as well as new production depending on the age of the parts. It's a natural phenomenon and there is little you can do about it. You must operate within the constraints of the components being used. You will need to widen your tolerance. See below for the guidelines we follow at my company.

Dielectric guidelines (Note: This is not a summation).

Y5V -- -20% of nominal value, -part tolerance, -machine tolerance X7R -- -9% of nominal value, -part tolerance, -machine tolerance X5R -- -12% of nominal value, -part tolerance, -machine tolerance

These values indicate a potential drop in capacitance of -9% for an X7R cap over 3 decade hours (10,000 hours or 14 months). The aging process occurs immediately after curing at the mfg. You also have to factor in the part tolerance and the machine tolerance for the given range of capacitance being measured. You can correct this by curing the parts at your facility (baking at elevated temp for predetermined amount of time). Another method used for de-aging is to apply a solder iron to both sides of the capacitor. This will realign the crystalline material in the dielectric and you will be good to go for a few more months. The better option depending on your application may be to just widen the tolerance of your tests to except parts that have aged.

reply »


smd capacitors- ICT measuring | 14 June, 2013

Thanks , Reese.Yes , I check the capacitors with C-meter.When measured with C-meter , all capacitors are OK - +/-10% of the value. I can not underst?nd "Note: This is not a summation).

Y5V -- -20% of nominal value, -part tolerance, -machine tolerance X7R -- -9% of nominal value, -part tolerance, -machine tolerance X5R -- -12% of nominal value, -part tolerance, -machine tolerance" If the value is 10uF, tolerance 10% , and machine tolerance 8% , what is the limit of the component?

reply »


smd capacitors- ICT measuring | 14 June, 2013

Is this a new assembly/new test fixture/program? Or is this an existing assembly that is suddenly exhibiting this failure?

Capacitors measured in circuit will not necessarily measure their correct values. Active components in circuit will be affected by the circuit that they are in.

If this is a new program, the program should be checked against a known good golden board from your customer. Run the golden board through the test, and note the values that are measured on these capacitors in the circuit. Compare this to the boards that you are building.

If this is an existing product that has suddenly changed performance, things are a little stickier. First, verify that no one has modified the settings in the tester. Operators/test engineers with access will sometimes adjust capacitor readings in circuit to maintain higher yields. Second, verify the component on the bench out of circuit against manufacturer specifications. Third, measure the component installed in the board with your LCR meter on the bench. Next, check to see if other components have changed in this circuit. And IC from a different manufacturer, for instance, could have different impedences on input/output which are active during the ICT testing of this component. Finally, if no other evidence exists, enlist customer support in testing the boards with the "failing" capacitors in their final assembly, and determine if the components are really failing in application.

Cheers, ..rob

reply »


smd capacitors- ICT measuring | 14 June, 2013

The low value will be 0.624uF. You have to factor it from the low tolerance (-10%). So:

0.9 x 0.92 (machine tolerance) x 0.91 x 0.91 x 0.91 (3 decade aging factor for X7R dielectric). We factor hours at the 3rd decade (10,000 hours), but you can choose a different decade if you wish.

That said, based on your response, it may not be aging, but a test problem. Your machine tolerance seems a bit high. We use z1850/60 machines, which are older and the tolerance for these systems are 2-3% at this capacitance.

If you are measuring 30-40% below nominal, you may have guarding issues. Can you elaborate?


reply »


smd capacitors- ICT measuring | 14 June, 2013

Chances are this is not a component failure. First, these are passive components with passive testing on the ICT. He may have guarding issues, but more information is required as has been stated. Are all capacitors with the same P/N failing in different parts of the circuit or is it in one location?

Try isolating the failing component by removing associated circuitry that could be interfering with the test.


reply »


smd capacitors- ICT measuring | 18 June, 2013

production lot is 4000pcs.About 70% of capacitors are out of tolerance ( +/-20%).There are 3pcs C=10u per PCB, in different location.Samsung capacitors wich we use are 1,5 years old - i checked this information from our supplier " Rutronik ".I have not "gold PCB" , because this is our own product.Its strange , that the same capacitor we use in 3 other product , and have not any problem ( but these products have only functional test, have not fixture for ICT). Thanks for advise and sorry for my bad English.

reply »


smd capacitors- ICT measuring | 18 June, 2013

If they're 1 1/2 years old, they are definitely aged parts. Either send them back for newer or widen your tolerances on the ICT to compensate. You are at the 3rd decade now (10,000 hours) and you will need to compensate for this. These parts are not defective, but aged and are probably within the mfg tolerances. You can de-age the parts by touching a soldering iron to both ends to test this out. You will see the part value self-correct to 10uF. You won't notice any failures on the FCT for aged parts unless the application is time dependent such as an oscillator. Most of our X7R parts are used for filtering/bypassing. The ICT is the only place that aging would be apparent. Unfortunately with these dielectrics, aging is the nature of the beast and normal.

reply »

Manufacturing Software

PCB Cleaning