Why boards need to be cleaned – and what designers can do about it.
Designers typically don’t think of post-assembly cleaning. In a small lecture hall at the Lake Washington Institute of Technology this September, one expert explained to a group of designers why they should.
That expert was Dock Brown, co-chair of the IPC DFX committee and senior member of the technical staff of DFR Solutions. He presented component cleanliness experiences and models to the IPC Designers Council Cascade (Seattle) chapter.
The key, he said, to mathematical models is to balance their complexity against the answers you’re looking for. With that in mind, he shared models he’s developed, starting with his risk assessment model: reality/results vs. perception/belief (based on experience).
His medical model for device reliability (the Gold Standard) is one in which you can apply knowledge to prevention and intervention:
- Knowledge (what we know)
- Decisions (how we assess)
- Actions (how we proceed)
This model enables, foresight, hindsight and insight, he said.
Brown suggested a reliability model, as well, where mechanical and environmental stress is pitted against strength. This quantifies what reliability looks like, he said.
Next, he displayed a graph indicating environmental stress vs. product lifetime with different markets represented. They coincide with IPC Class 1 (shorter lifetime, less environmental stress, lower left quadrant of the chart); IPC Class 2 (somewhere in the middle); and IPC Class 3 (longer lifetime, higher stress environments, upper righthand quadrant).
“In the 1970s,” he said, “60% of the costs of ICs appeared in that upper right quadrant,” including, military, aircraft, etc.
“These days, geometries, stencil thicknesses, and pads are getting smaller,” which equates to “less paste applied.” But, even though the “volume of paste is 70% less, the oxide layer thickness remains the same, which requires higher activity fluxes.”
“There are four monolayers of water (a polar solvent) on each surface,” above the insulator material, between the anode and cathode. “Most measurements are performed on nonporous materials” such as metals and ceramics “but the material of interest to us is polymers.
“Electrochemical migration (ECM) is largely a nucleation phenomenon,” Brown said. When relative humidity (RH) is less than 60%, ECM is rare. When RH is greater than 80%, ECM is possible, and when there’s condensation, ECM is likely.
“You can get liquid water forming in cracks at low humidity,” he said, discussing the “small geometry effect.” For example, “this is a problem in ceramic capacitors.”
How do you get rid of this stuff? Answer: “You clean your boards."
And this is where design decisions come in, he elaborated. “It’s better to have wider spaces. Even small changes in spacing make a colossal difference.”
Brown called no-clean flux residues “sadly named.” Yet, “more and more people are using” them. “What makes ‘no-clean’ safe?” he asked. “Where is the empirical evidence? Can activation be objectively validated?”
He deferred to a ground-breaking IBM paper that suggested metals salts don’t evaporate, which can be “problematic when it comes to ECM.” IBM did tests on fluxes, and there was “no evidence of polymerization that makes flux residues benign.
“There is no such thing as an activating flux,” Brown said, calling it an “old husband’s tale.”
He said to consider doing cleanliness testing at 40°C/93% RH, instead of 85°C/85% RH, and then asked, “Where are we as an industry with cleanliness and contamination? We don’t have any validated models for margins.”
Circling back to the earlier portion of his talk, he said what the industry needs is “validated mathematical models, [but] we’re certainly not here.”
Fortunately, he said, “people are getting smarter about board design,” and “chemistry suppliers are getting better at understanding how these things work. People are cleaning no-cleans.”
And companies are designing cleaning agents to target specific soil types and products.
“In the near future, we need to be cleaner and drier: cleaner than clean and dryer than dry,” Brown concluded.
The next IPC Designers Council Cascade chapter meeting will take place Dec. 7 at the same Kirkland, WA, location: http://cascade-ipcdc.org/.
Chelsey Drysdale is senior editor at PCD&F; This email address is being protected from spambots. You need JavaScript enabled to view it..
