Printed Circuit Board Design For Assembly

Printed circuit board assembly has come a long way since I worked in a factory. Here are some of the key differences: 

• The components are smaller; much smaller in many cases
• Through hole parts have become a novelty while surface mount has become the norm.
• Manufacturing has moved off shore for most consumer electronics
• The RoHS directive has put a damper on Sn63 solder for those same products

The shrinking of every type of device makes assembly and rework more difficult. Access to the pins used to be from the solder side of the board. Wave soldering and drag soldering have been largely replaced with solder paste and reflow ovens. Dedicated rework nozzles using hot air have replaced the solder wick and the solder sucker from the good old days.

Double-Sided Printed Circuit Board Assemblies

Times change. We even changed the name from solder side to secondary side to acknowledge the advent of double sided boards using surface mount technology. We still prefer one-sided boards in the factory, especially when there is a mix of surface mount and through-hole components. Ideally, it would be one or the other. 

While it is possible to reflow both sides of the board simultaneously, there is a yield concern with parts on the bottom side becoming victims of gravity. Dedicated fixtures or adhesives may be required for the secondary side. The common practice for double sided boards is to place the heavier components on one side and only the smallest ones on the other side.

Figure 1. This FPC was done in a rush and had a few mis-wired connections. We then rushed out this board with resistors on one side and test points on the other. The EE forgot to skip a few no-connect pins so the board that was supposed to correct it was also a mess.  We fixed it with blue wires; some of which have fallen off. It was a comedy of errors but we can still laugh about it nine years later! Image Credit: Author

If the component mix is within those limits, then a high temperature solder process would be used for the primary side while a lower temperature and shorter molten phase could get the job done for the secondary side - without reflowing the primary side. Even then, some solder defects are likely to occur on occasion. Single sided boards have an inherent advantage in that regard.

A/B Flip Assembly Sub-Panels

Mobile chip companies assume that a double sided board will be used so that decoupling capacitors can be placed near the assigned pins of their fine-pitch BGA devices. Furthermore, all of the components will be small and light enough to go on the secondary side. When this is the case, it is possible to use a sub-panel that has half of the boards face-up while the other half is flipped over. The benefits are as follows:

• The same pick-and-place file works no matter which side is up; one program, two sides.
• The placement takes the same amount of time for each side improving the process flow.
• The same stencil can be used either way so you only need one for both sides.

The parts mix can hinder that process. Four symmetric tooling holes are required. Note that three tooling holes are preferred for standard sub-panels so that the sub-panel cannot be put in the pick and place fixture the wrong way. For that reason, perfectly square sub-panels should be avoided. As I recall, Qualcomm has a list of parts that cannot be used with an A/B flip panel due to their excessive weight.

Figure 2. This is an example of an A/B flip sub-panel circa 2004. It is slightly wider than it is long. One thing I would add in hindsight is an arrow in the break-off area that shows how to orient the panel in the fixture. The through-hole connector would be hand-soldered as a second operation; something we’d like to avoid if possible. Image Credit: Author.

Navigating Solder Challenges in a RoHS-Compliant World

The Reduction of Hazardous Substances (RoHS) directive has reduced the use of lead, cadmium and other metals that are known to cause harm to people and the environment. The upshot is that tin-lead solder is phased out for many products. They will use it for aerospace and earthbound applications where the electronics will not end up in a landfill. The factory has to certify that they will properly dispose of it.

The benefits of tin-lead (Sn63/Pb37) solder are two-fold. One is the lower working temperature and two is that it is eutectic. That means that it solidifies over a very narrow temperature band which decreases the likelihood of certain types of solder defects including solder peaks and disturbed solder. The RoHS directive has exceptions baked in for times when nothing else will do.

We have yet to create an alternative that works as well. Tin-Silver (Sn/Ag) is close except for the fact that it has a higher process temperature. The silver is more of a trace element that helps avoid tin whiskers. Pure tin is also subject to dendritic growth in humid or high-voltage environments.

DFA Considerations for Component Placement and PCB Documentation

The assembly house is primarily concerned with putting parts on the board. Secondarily, they need to be able to remove them if necessary. Legible marking helps in both cases. The part number for the bare board usually goes on the bottom. It could potentially be done with copper or even soldermask if the board doesn’t have silkscreen on the secondary side.

The assembly number and revision go on top along with the vendor logo and datecode. At the component level, orientation is the number one reason for silkscreen. The next priority is to have reference designators followed by the component outlines. Outlining and labelling various functions of the board is a nice touch if you have the room.

The PCB library will determine the quality of the end product. A courtyard around each footprint will set the minimum area between parts. There are three common sizes for different purposes. Analog chains will want the very minimum of spacing to obtain the best performance. Those are also the least reliable footprints as they sacrifice part of the toe-fillet.

One time, I used the least material condition for tight spacing with the smallest pads on the RF section. The regular size footprints were used for everything else. The manufacturing engineer who was in charge of the assembly line was not happy with that plan. Be consistent with your sizes or you will be consistently called onto the carpet for some head-shaping. Like I wrote above, “One time” - and never again.

Sending the design out to off-shore facilities for manufacturing removes us from the heat of the assembly line. That’s when we need to really be sure our PCB is good to go. The time difference means that a day can be lost for a question to be asked and answered. By the time we get to work on Friday, they’re knee-deep into the weekend. When we come back to work on Monday, it’s already Tuesday overseas. Airtight documentation is a must even if the manufacturing team is in-house.