Cleaning blind holes and internal passages is where most parts cleaning processes quietly fail. So why does ultrasonic cleaning of blind holes so often leave residue behind when the outside of the part comes out spotless?
The short answer: immersion alone doesn’t clean what liquid and sound energy can’t reach. A component comes out of the tank looking spotless, goes into assembly, and the coolant passage or the tapped hole still holds swarf, cutting fluid or lapping compound. Nobody sees it until something seizes downstream. Here is what actually reaches those features.
Why cavitation stops at the entrance to a hole
Ultrasonic cleaning works because cavitation bubbles form and collapse against the surface. That needs two things at the surface: liquid, and sound energy reaching it. A blind hole breaks both.
When a dry part goes into the tank, air is trapped at the bottom of every blind feature. Liquid cannot fully displace it because the air has nowhere to escape — the hole has one opening, and it’s occupied by the liquid trying to get in. The pocket stays. Below that air, no cleaning happens at all.
Sound energy attenuates too. It travels well through open liquid but drops off sharply once it has to turn a corner or travel down a narrow channel. On a long, narrow passage, the far end receives a fraction of what the outer surface does. The part is clean where you can see it and dirty where you can’t — which is the worst possible combination, because visual inspection passes it.
The features that cause the most trouble
- Through-coolant passages in tooling and spindles. Long, narrow, often with internal transitions, and packed with a slurry of coolant and fines that has dried in place.
- Tapped blind holes. Thread roots trap chips, and the geometry holds air stubbornly.
- Cross-drilled intersections in hydraulic and valve bodies. Burrs at the intersection catch swarf, and the junction is exactly where flow is weakest.
- Deep bores after honing. Honing residue clings to the surface finish that was the point of honing.
- Assembled modules with internal gaps. The gap fills with rinse water that later reappears as a stain.
What actually clears them
No single technique solves this. Every method that works does one of two things: gets the air out, or gets liquid moving through the feature instead of past it.
1. Orientation and fixturing — the cheapest fix
Blind holes must not point downward. Fixture the part so openings face up or sideways, and trapped air can rise out as the part submerges. This costs nothing and is the single most common unrecognised cause of failed cleaning. If results vary between operators, this is usually why — one of them is loading the basket differently.
2. Vertical agitation and lifting
Moving the basket up and down through the bath forces liquid exchange in and out of features. Our multi-tank lines and lifting machines do this on a programmed cycle rather than by hand, so every basket gets the same movement. Static immersion relies on diffusion, which is far too slow for a deep passage.
3. Directed spray through the feature
For coolant passages and cylinder bores, immersion is the wrong tool on its own — you need flow through the channel. This is why several of our automotive line builds combine ultrasonic tanks with a spray stage where nozzles are aimed into the bore. The hydraulic barrel cleaning machine we built works exactly this way: bore-directed flushing clears honing residue that immersion leaves behind.
4. Degassing the bath — not optional here
A freshly filled tank holds dissolved air that absorbs cavitation energy. For open surfaces you might not notice. For blind features, where energy is already marginal, it’s the difference between clearing the hole and not. Run a degas cycle before the first load and after any refill.
5. Vacuum drying for what remains
Liquid that gets into a blind feature has to come out again. Hot air will not pull rinse water from a deep pocket in any reasonable time, and what stays behind evaporates later, leaving mineral residue exactly where you can’t inspect. Vacuum drying lowers the boiling point so trapped liquid leaves as vapour. Our seven-tank vacuum systems for display components were built for this problem specifically.
Does a bigger machine fix it?
This is the question we get most often, and the answer is usually no. Doubling tank power raises cavitation intensity across the whole bath, but it doesn’t move air out of a blind pocket and it doesn’t push liquid down a narrow channel. What it does do is increase the risk of surface damage on the outside of the part — the area that was already clean.
Where extra power genuinely helps is on large parts in big tanks, when the energy simply isn’t reaching the far end of the basket. That’s a coverage problem, not a geometry problem, and adding immersible transducer plates on additional faces usually solves it more cheaply than replacing the machine.
For internal features, spend the money on movement instead: agitation, spray, and drying. A modest tank with programmed lifting and a spray station will out-clean a much more powerful static tank on any part with holes in it.
Chemistry matters more here than on open surfaces
On an open face, a mediocre detergent still gets replenished constantly by bath circulation. Inside a blind hole, the small volume of liquid that does get in becomes saturated quickly and then stops working. That’s why parts with internal features often respond better to a chemistry change than to more ultrasonic power.
Two practical points. Lower viscosity liquid enters narrow features more readily, which is one reason hydrocarbon cleaning sometimes outperforms water-based systems on deep bores despite lower cavitation energy. And temperature helps up to a point — warmer liquid is thinner and dissolves oil faster, but past the peak it degrades cavitation, so hotter is not automatically better.
If you’re running a benchtop machine and struggling with tapped holes, changing chemistry and orientation is nearly always the cheaper experiment to run first.
How to check whether it actually worked
Visual inspection cannot verify a blind hole; that’s the whole problem. Practical checks that do work:
- Solvent flush and filter. Flush the passage with clean solvent, catch it on a filter membrane, inspect the membrane. This is the closest thing to a real answer.
- Borescope. Slower, but shows where residue actually sits — which tells you whether it’s a fixturing problem or an energy problem.
- Weight before and after. Crude, but on a batch it detects gross failure.
- Sacrificial section. If a part can be cut, cutting one from a batch settles arguments quickly.
The process order that works
For parts with significant internal features, the sequence that produces consistent results is: orient the part so air escapes → degassed ultrasonic bath with vertical agitation → directed spray through the passages → rinse → vacuum or forced-air drying appropriate to the geometry.
Skip any of the four and the failure shows up at assembly rather than at inspection. That’s the expensive place to find it.
Send the part rather than the drawing
Internal features are the case where specifications are least useful and a sample test is most useful. A drawing shows the geometry; it doesn’t show how the residue behaves, whether it has dried, or whether your existing process is already close.
Jietai Ultrasonic has built cleaning lines since 2003, and forty of the custom systems we’ve delivered exist because a standard tank couldn’t reach something. Send us the part with its actual soil — we’ll run it, section or borescope it where that’s the only way to tell, and report back the frequency, agitation, spray configuration and drying method that cleared it. The cleaning test is free, and if immersion alone won’t do it, we’ll say so rather than sell you a tank that can’t.