Issue link: https://iconnect007.uberflip.com/i/1284035
SEPTEMBER 2020 I SMT007 MAGAZINE 127 them through a reflow oven six times. Now, they want us to retest on our testers to see if we have an open that occurs. Again, this is at ambient room temperature. We were never going to find any weak micro- vias, even though they passed it through a reflow of six times. A customer might say, "I want to retest these boards, I will pay you for it." But when I look at the designs, I typically find that it was a three- or four-stacked microvia, which we highly recommend not doing (Figure 8). In most cases, if you have a defective micro- via only at reflow, then only resistance is open. In this example, you can see the six peaks. This is the resistance at each reflow cycle. The red line is a 5% threshold that we cannot exceed; if we do, it is a bad panel, and it is going to get rejected. Typically, you do not hear board shops saying this, but if we have a bad serial number, I do not want you to build with it. Rather, I want to throw it away. I do not want you to have failures. I will reject it. I will be happy that you are not getting bad products. As you can see in Figure 9, the little line that goes over the 5% red line threshold is a defect. The resistance at reflow at the peak of reflow temperature went over the 5% threshold, but only as it cooled down. You will notice that the resistance went away, and it is now connected, or self-healed, at ambient room temperature. This gives that false hope that everything is fine. But you can detect these, in many cases, only at reflow with actual reflow temperature. Traditional thermal testing with thermal shocks uses the chamber in which it normally cycles from a very cold temperature (-60°C) to 160°C. Originally, it was thought that you didn't have to check if the resistance changed after every reflow. The thinking was you just had to exercise through the six reflows. Then, you put them on the tester, and a whole mea- surement of resistance was only done in the thermal shock temp chamber (Figure 10). Figure 9: OM testing with four-wire resistance from ambient to reflow temperatures. Figure 8: Electrical testing at ambient temperatures will not find weak microvia interfaces.