Debunking Urban Legends about PCB Cleaning
By Mike Konrad, President and CEO, Aqueous Technologies Corporation
 |
| Trident batch cleaning system.
|
First, let us begin with some basic historical facts. Cleaning is not a new process. In fact, cleaning has been a staple of the electronics assembly process from the invention of the electronic circuit board. In the cleaning industry, we divide our world into two sections; pre-1989 and post-1989.
Before 1989, virtually all circuit assemblies were cleaned after reflow. Flux and other contamination were removed from the assembly prior to use. We stuffed the board full of components, cut and clinched the leads, soldered the components to the board, then removed the flux. Much of that changed in 1989.
CFCs Banned from Cleaning
Chemicals containing CFCs were determined to be harmful to the Earth's ozone layer and, thus, were restricted or eliminated from production. The most common cleaning solvents used to clean assemblies contained CFCs that presented a problem for the electronics industry. While some manufacturers converted to environmentally responsible water-based cleaning technology, another option soon presented itself. So-called "no-clean" flux was introduced. No-clean flux was designed to not be cleaned. The majority of the electronics industry embraced no-clean technology except for military, medical and a handful of other high-reliability manufacturers. Virtually overnight, the cleaning industry shrank to a fraction of its pre-1989 size. Cleaning for the majority of manufacturers was dead.
Today, buyers of cleaning/defluxing systems fall into two groups. The first group is comprised of people who were in the industry before 1989. In many cases, their knowledge of cleaning machines and processes is obsolete. The second group is comprised of "younger" people who have had no direct experience with cleaning and who frequently lack any reference point. Those in either group may find this information helpful.
 |
| Cleanliness testing system.
|
So many things have changed since cleaning was considered to be a mainstream process. Today's assemblies, due in part to miniaturization, have a much lower "contamination tolerance" threshold. While some assemblies function reliably with a specific volume of contamination, others rapidly fail. Clock speeds, component densities, assembly geographies, solder alloys, environmental influences and other factors determine how much contamination an assembly can handle without failure.
Urban Legends Debunked
Water may be used to remove water-soluble (OA) flux but not rosin. False. Water, mixed with a low concentration of a water-based defluxing chemical, removes all flux types (water soluble, rosin, no-clean).
Solvents work better than water. False. While solvent technology works well, a water-based process works better — in most cases. Some in this industry may remember the pre-1989 days when an assembly that had to be "extra" clean — i.e. for conformal coating purposes — was cleaned using a solvent, then followed up with a DI water cleaning process to improve the cleanliness of the assembly's surface. Water-based cleaning and, more importantly, water-based rinsing, typically produces superior cleanliness results compared to solvent processes. This fact was reiterated in IPC's "Phase II" test results — a comparison of solvent vs. water-based cleaning.
Conveyor vs. Batch Cleaning
Inline (conveyorized) cleaning machines are better than batch cleaning machines. True and False. The fact is that both batch and inline technologies clean equally well. No one configuration cleans better or worse than the other. One should choose batch or inline based on the volume of assemblies that need to be cleaned. Batch cleaning systems are capable of cleaning volumes ranging from low to high. Inline cleaning machines normally are associated only with high volumes. When all cost centers are accounted for, the operational cost of a batch process is approximately 10-20 percent of the cost of an inline cleaning process. As the cleaning volume increases, the cost differential lowers. The cost per assembly on an inline cleaner begins to lower when the cleaning volume approaches several thousand assemblies per day. Volumes less than that generally are better suited for batch-format cleaning processes.
Dishwashers/Glassware washers work well. False. While many batch-format cleaning machines may resemble dishwashers, dishwashers typically are not compatible with the defluxing chemicals nor are they capable of acceptable under-component penetration. Dishwashers are for dishes.
We do not need a powerful machine. The defluxing chemical lowers the surface tension, allowing adequate impingement. False. While defluxing chemicals do lower the wash solution's surface tension and allow for improved under-component penetration, the real issue is the rinse cycle. Today's defluxing results fall into two categories: excellent and disastrous. The worst thing one can do is to partially clean an assembly. Wash solution allowed to remain on an assembly, due to inadequate rinsing, is far worse than flux being allowed to remain on an assembly. Rinse is the most critical part of a successful defluxing process. Rinse water, unlike chemical-enriched wash solution, has a higher surface tension. That means the rinse water must be forced into tight spaces to effectively displace the lower surface tension wash solution. This is accomplished by forcing water through multiple nozzles that diffuse the solution into smaller water particles. The smaller the water particle, the better its ability to get under tight spaces.
IPC standards state how clean is clean. True and False. While IPC TM650 does state "how clean is clean," we cannot recommend that anyone actually embrace that value. IPC allows for the use of a resistivity of solvent extract (R.O.S.E) test to determine the cleanliness of an assembly. These testers have been in production for more than 30 years. In fact, the cleanliness standard of >10µG NaCl/in.2 was developed in the 1970s. Today's assemblies contain advanced geographies not imaginable in the 1970s. There are some that suggest that we abandon the use of R.O.S.E. testers. That would be like throwing the baby out with the bath water. Instead, one should calculate the cost of failure. If an assembly fails, will that lead to property damage, loss of life, loss of reputation? If so, one should consider selecting a lower pass/fail value. In short, the more critical the reliability, the closer to 0.0 one should strive to achieve.