“Will an ultrasonic cleaner destroy my board?” is the first thing anyone asks about using an ultrasonic cleaner for PCB assemblies, and the honest answer is: it can, and it’s avoidable. Electronics manufacturers have been ultrasonically cleaning boards for decades. The failures come from specific, identifiable mistakes — not from the technology being unsuitable.
Here’s what actually damages boards, and how production lines avoid it.
What ultrasonic cleaning removes from a board
The main target is flux residue. Rosin and no-clean residues left after reflow or wave soldering trap ionic contamination underneath, and that’s what causes electrochemical migration, dendrite growth and eventual leakage between conductors — usually months later, in the field.
Beyond flux: solder balls and paste residue, handling contamination, conformal coating overspray, and particulate trapped under low-standoff components. That last category is the reason ultrasonics is used at all. Under a 0.05 mm standoff, a spray nozzle delivers almost nothing and a brush delivers nothing. Cavitation reaches wherever liquid reaches, which under modern components is the only mechanism that works.
What actually damages boards — all four are avoidable
Damage is real but specific:
- Resonant vibration of long, thin structures. Wire bonds, tall unsupported leads, crystal oscillator elements and some MEMS structures can be driven into resonance and fatigue. This is the genuine physical risk, and it’s why frequency selection matters more here than in any other application.
- Liquid ingress into non-sealed components. Relays, some connectors, unsealed potentiometers, trimmers and certain displays let liquid in and never let it out. Nothing about frequency fixes this — the component is simply not washable.
- Label and marking loss. Adhesive labels and some inkjet markings don’t survive.
- Damage from the chemistry, not the ultrasonics. Aggressive alkaline attacks aluminium electrolytic seals and some plastics. Blamed on cavitation, caused by pH.
Notice what isn’t on that list: solder joints, passives, standard ICs, connectors that are sealed. Those survive properly configured ultrasonic cleaning without difficulty.
Frequency: go high, and stay high
This is the single most important decision. For assembled boards, 80 kHz should be your floor and 120 kHz is better. Higher frequency means smaller, gentler, far more numerous bubbles — enough energy to lift flux from under a component, not enough to fatigue a wire bond.
28 and 40 kHz belong on bare boards, stencils and frames, not on populated assemblies. That’s the frequency band where component damage reports come from.
For fine-pitch and high-density assemblies, our three-frequency series runs at 80 and 120 kHz, and megasonic units near 950 kHz handle sub-micron particulate on the most delicate work — semiconductor substrates and optical assemblies rather than ordinary PCBs.
Bare boards, stencils and assemblies are three different jobs
An ultrasonic cleaner for electronics covers three distinct workloads, and treating them as one is how boards get damaged.
Bare boards and frames tolerate 28–40 kHz without difficulty — there are no components to fatigue. This is ordinary industrial cleaning.
Stencils and misprinted paste are the easiest case and the most common first purchase. Solder paste comes off readily; the constraint is chemistry compatibility with the stencil, not cavitation energy. A dedicated ultrasonic PCB cleaner sized for stencil frames pays for itself quickly in a busy SMT line.
Populated assemblies are the difficult case, and everything else in this article applies to them specifically. If you buy one machine for all three, specify it for the assemblies and run the easy jobs on the same settings — not the other way round.
Chemistry: this is where boards get quietly ruined
Use a cleaner formulated for electronics assemblies. General-purpose industrial degreaser is the wrong tool, and the reason is aluminium: electrolytic capacitor cans and seals don’t tolerate high pH. Damage isn’t immediate — the capacitor degrades and fails in service, which makes it nearly impossible to trace back to the cleaning stage.
Rinse quality matters more than on any other application. The whole point is removing ionic contamination, so a final DI rinse isn’t optional — rinsing with tap water leaves behind exactly the kind of ionic residue you were trying to eliminate. If you’re measuring cleanliness with ROSE testing or ion chromatography, your DI rinse quality sets the ceiling on your result.
And drying: liquid trapped under a BGA or in a connector shell has to come out. Hot air alone struggles with low-standoff parts, which is why electronics lines use vacuum drying rather than a hot-air cabinet.
A configuration that works on populated boards
- Frequency 80–120 kHz, moderate power. Resist the urge to increase intensity when a board looks dirty; change chemistry or time instead.
- Electronics-grade chemistry at the supplier’s concentration, in a degassed bath at 40–55 °C.
- Boards racked vertically, not stacked, with space between them and nothing resting on the tank floor.
- Short cycles — typically 3–8 minutes. Long soaks add ingress risk without adding cleaning.
- Wash → rinse → DI rinse → vacuum dry, as a sequence rather than a single tank.
Then qualify it: run a sample batch, test it the way you’ll test production — ROSE, ion chromatography, or visual under magnification — and inspect the sensitive components specifically rather than the board as a whole.
How to qualify it without risking production boards
Nobody should validate a cleaning process on saleable product. The sequence that keeps the risk contained:
Start with scrap boards that carry the same component mix as production — including whichever part you’re most worried about. Run the intended cycle, then inspect that specific component under magnification rather than glancing at the board as a whole. If a crystal or a relay is the concern, power the board afterwards and check that part functions, because visual inspection will not show a fatigued bond or ingressed liquid.
Then repeat the cycle five or ten times on the same board. Single-cycle damage is rare; cumulative damage from a marginal frequency or an over-powered tank shows up on repeat exposure, and a board that survives ten cycles will comfortably survive one.
Only then move to a small production batch, with the sensitive components inspected on every board rather than sampled. That’s three days of work and it settles the question permanently.
When ultrasonic cleaning is the wrong choice
Be direct about this. If the assembly carries unsealed relays, exposed wire bonds, unencapsulated MEMS or components the manufacturer explicitly marks as not washable, ultrasonics is not your process. Spray-in-air cleaning or a manual process is the correct answer, and no frequency selection changes that.
If the board is already conformal coated, cleaning is too late; contamination is sealed under the coating.
We’d rather say this before you buy a machine than after.
Test on a real board
Every board is different — component mix, standoff heights, flux type, whether it sat for a week before cleaning. That’s not something a specification sheet resolves.
Jietai Ultrasonic has built ultrasonic cleaning equipment since 2003 from our own factory in Dongguan — over twenty years in this one industry, as a recognised high-tech enterprise — including electronics-industry lines for exactly this work. Send us a populated sample board with its actual flux residue, and tell us your cleanliness requirement. We run it, inspect the sensitive components afterwards, and report the frequency, chemistry, cycle time and drying method that worked — including if the answer is that some component on your board shouldn’t be immersed at all. The cleaning test is free, and one real board answers more than any amount of specification.