PCB Tech - Circuit board lead-free reflow test welding

Circuit board lead-free reflow test welding

1. The first trial production and trial welding
(1) PCB structure and reflow
First, the FR-4 sheet hardened with high Tg and Dicy is used to make two kinds of high multi-layer boards with 22 layers and 24 layers. And use two kinds of reflow furnaces for the blank board, and perform 6-9 times of simulated reflow under the two L-shaped reflow curves. Now the two board types and their reflow curves are explained below (in fact The author believes that this kind of reflow curve without heat-absorbing saddle is not suitable for general multilayer boards, and even more can not be used for lead-free reflow of thick multilayer boards):

The cooling rate is 0.81°C/sec, which is a rather slow method.

(2) Microsection analysis
After many reflows and multiple bursts, the failure analysis was carried out for the burst area. The following are the findings of the other slices.

(3) Discussion
After the first reflow above, several representative logics can be seen:

.If the reflow curve heats up too fast, it is easy to explode the plate. As for whether the temperature drop too fast is related to the exploded plate, it is still unknown. And the author feels that the reflow curve used by the above-mentioned circuit board company is actually not appropriate. This kind of straight-up and straight-down curve without saddle heat absorption is only suitable for the reflow of low-end boards and simple parts. Complex multi-layers must be equipped with saddles, or the curve of the heat absorption section of the long saddle, so that when the plate body is at a uniform temperature inside and outside, the rapid rise of the peak temperature can be performed to complete the welding.

Only 40% of the unexploded plates in the first weld were welded, and all failed after the second reflow.

The second is the performance of the human board through the reflow curve 1. Among them, the B board reflow curve 1 has the best performance.

Even though there was a 67% survival rate after 5 reflows, all failed at the 6th time.

.BGA ventral bottom porous and dense pores are easy to accumulate heat and burst.

. Those who have done lead-free tin spraying process are more likely to burst. For example, the large copper area in the inner layer is easy to burst.

.Although the plates used in this test are all Dicy hardened high Tg plates, refer to many other evidences; if Dicy hardened plates are used, even if the circuit board is made well, its bursting is not as good as PN hardened ones. less.

2. The second trial do and trial welding
The plates in the second test have been matched with different plates with Dicy and PN hardeners. From the results of this test, it can be seen that the heat resistance of the PN type is indeed better than that of Dicy. At the same time, it can be seen that there are still factors that affect the lead-free soldering and cause the board to burst: the pressing process, the baking after pressing, the water absorption of the inner layer, the water absorption of the finished board, and the degree of resin polymerization.

In the circuit board production process, the A board uses Dicy hardening and PN hardening respectively. Although two different pressing processes are also selected, it is found that it has little effect on the result. On the contrary, the baking of the blank board before welding has a direct impact on the bursting of the board. The baking condition is 125°C for a total of 24 hours. The survival rate of the plate and lead-free reflow is now sorted out.

(1) Discussion
.For FR-4 hardened by Dicy, the cracking phenomenon is almost the whole board cracking at the same time, while for PN hardened, local cracking occurs only in the porous area of the abdominal bottom.

.Dicy hardeners will all burst after two reflows regardless of whether they are baked or not before reflow. However, those that are PN hardened and baked before welding can survive 50% after four reflows.

.It shows that Dicy is easy to absorb water due to its high polarity, so it is not easy to pass the test of thermal stress. The polarity of PN is very small, the water absorption rate is extremely low, and the addition amount is more than 20% bywt. In fact, the linear nature of epoxy resin has been greatly changed, and it has the three-dimensional structural strength of phenolic resin, so it is strong. It is no longer prone to cracking.

3. The third trial do and trial welding
(1) Preparation for the test
In the third test, all the plates have been changed to hardened types, and the PCB process has been especially improved. For the sake of better lead-free reflow yield, deliberately bake all finished inner boards at 110 degree Celsius for 3 hours, and bake the outer boards at 150 degree Celsius for 4 hours after removing the slag. As for the surface treatment, the 22-layer eight plates are electroplated nickel gold instead of ENIG. A total of 6 batches of 15 boards were made this time, and the 6 boards were deliberately placed at 125°C for another 24 hours before reflowing. The other 6 boards are deliberately not baked before reflowing, as a comparison of the effects. In addition, two boards were taken from each of the two batches to perform respectively: the thermal stress test of bleaching tin, the measurement of Tg, the T260/T288 test, and the reflow curve 2 simulated reflow. In this test, it was found that the two types of PN hardened boards, which were baked before welding and unbaked, did not appear to burst after 12 simulated reflows.

(2) Discussion of results
The test results of the above six batches of boards are now sorted out and discussed as follows: For the six batches of PN hardened boards, whether or not they have been baked twice in the PCB process, they can pass 12 times of simulated reflow. Three methods were used to test the △T of Tg1 and Tg2 before the reflow. Although it was found that the △T still had a difference of 1-8 degree Celsius, the △T had indeed become much smaller after 12 reflows. It means that the degree of hardening of the original resin is very good, and the degree of hardening is directly related to the pressing process and the baking after pressing.

.Because Tg2 is still higher than Tg1, this means that the resin in the board has not yet shown signs of cracking.

.T288 test was carried out after 12 reflows, and it was found that the data obtained was not lower than the reading before welding, which can also be interpreted as evidence that the resin has not cracked yet.

.After three and six times of tin bleaching, all passed the test without any bubbling or bursting. Although there are also floating holes on the slices due to CTE mismatch, and shrinkage of the resin with volume shrinkage (not exceeding the hole length of 20 %), but these are all phenomena that are inevitably caused by strong heat. As long as there are no microcracks in the plate slices, they can generally be regarded as acceptable small defects.