Q1:A picture of the sectioned MLCC would be nice. A1: I don't think I can post pics in this (if anyone knows otherwise let me know), so I will email them to anyone who's curious. R1: You�re correct. You cannot post pix on SMTnet. Please email me the pix.
Q2: Placement Force A2: 0201s are placed on the Siemens S23s with 1 newton of force. The separation is at the metal to substrate so I doubt it is from placement force. R2: You�re correct. Placement force related cracking usually appears in the center of the MLCC, unless there are debris or something that makes the pads unequal heights.
Q3: Reflow method A3: BTU Paragon 150. 10 Zone, Nitrogen Inerted, Convection Ovens.
Q4: Where and how are you measuring Reflow recipe A4: Attaching a thermal couple to an 0201 is next to impossible. They are like grains of sand. I attach thermal couples to the laminate in the area of the 0201s. R4: Measuring the laminate in the area of the 0201 should be hotter than the component. So, if you�re keeping the laminate within the fabricator�s recommendations and getting good flow, then you�re probably fine.
Q5: Cooling rate of the Reflow recipe A5: The cooling rate of the laminate is 2-3 C/sec. The cooling rate of the air is 4-5 C/sec. I can slow it down more but according to the component manufacturer we are well within spec. R5: So if the laminate is cooling at 2-3 C/sec, then we�d expect the MLCC is cooling slower than the laminate, certainly not cooling faster than the laminate.
Q6: What is the appearance of the solder connection to the metalization A6: See pics. Solder is beautiful. Joints look great and the wetting angles are dead on. R6: If they look GREAT and there is not too much solder, then probably the problem is elsewhere.
Q7: Method used to depanel the board and distance damaged capacitors are from the board edge A7: On the first 4 builds they (lab) depanelized (for cross-section prep) with a band saw. We asked for a diamond blade on the last build and the separation was even more severe. R7: So if the lab is using a band saw, then flexure of the board during depaneling could be damaging these parts.
Q8: Distribution of the problem A8: Not sure. I�ve asked to lab to give us feedback on the distribution to see if we can identify a trend.
reply »