Here is the part of metal 3D printing nobody puts in the brochure: the print is often the easy half. A laser powder bed fusion (LPBF) machine can build a lattice heat exchanger with hundreds of thin struts and internal channels that no mill could ever cut. Then you pull it off the plate and discover that every one of those buried surfaces is rough, and no tool can get in there to fix it.
A new paper covered by Fabbaloo on September 7, 2026 tackles exactly that problem. Published in the Chinese Journal of Mechanical Engineering, it asks whether electrochemical polishing can smooth the surfaces hidden inside complex AlSi10Mg aluminum lattices. Here is what it means for anyone who designs, buys, or prints parts.
What the study is actually about
AlSi10Mg is one of the most common aluminum alloys in LPBF: light, stiff, and good at moving heat, which is why it shows up in brackets, housings, and heat exchangers. Lattices push that further by putting material only where it is needed. The catch is that LPBF surfaces come out of the machine with partially fused powder, stair-stepping, and other roughness baked in. On a solid block you can blast, grind, or machine it away. Inside a dense lattice, you cannot reach it. And as Fabbaloo points out, that roughness is not cosmetic. It can restrict flow, trap loose powder, create stress concentrations, and make thin struts measurably different from their CAD dimensions.
Electropolishing works differently from every mechanical method. The part goes into an electrolyte bath as the anode, current flows, and material dissolves preferentially from microscopic peaks. No tool has to touch the surface, which is why researchers keep coming back to it for lattices. Earlier work on flat AlSi10Mg LPBF samples, published in Surface and Coatings Technology, reported roughness dropping from 14.90 micrometers to 1.84 micrometers with an intermittent electropolishing method. The new paper asks whether the same idea holds up inside a lattice.
Why uniformity is the real fight
Fabbaloo's read is that the hard part is not whether electropolishing works, it is whether it works evenly. Material removal depends on current density, electrolyte circulation, gas bubbles, contact, orientation, and geometry. A strut near the electrode sees very different conditions from one buried deep inside. On a chunky bracket that does not matter. On struts a few hundred microns thick, it can quietly change stiffness, open up flow passages, or knock down fatigue life while the outside looks beautiful. AlSi10Mg's silicon-rich phases and as-built porosity add another variable, so there is unlikely to be one recipe for every lattice. The takeaway: a finishing process has to prove it did not change the thing it was supposed to improve.
I designed a part with internal channels. How do I know the inside will come out clean?
You do not know until you plan for it, so tell your print shop what the internal surface has to do (carry fluid, seal, carry load) before the file is quoted, because internal surface finish is a design and process decision, not a cleanup step. This applies to plastic parts too. On a filament print, a channel printed vertically has a very different wall texture from one printed flat, and an unsupported channel ceiling sags. If it carries coolant or air, orientation, layer height, and wall count matter more than the material name on the spool.
At DC Additive Pros we ask three questions on any part with hidden geometry: what flows through it, what does it seal against, and can we get a gauge or bore scope into it afterward. If the last answer is no, we orient the critical surface toward the nozzle or split the part into two inspectable pieces. Less elegant, but it is how you get a part you can trust.
My 3D printed part is too rough where it matters. Can it be fixed after printing?
Usually yes, if the rough surface is reachable: sanding, bead blasting, machining a mating face, or a chemical smoothing process can all bring a printed surface down to a usable finish, but a surface you cannot reach has to be fixed in the design or the print orientation instead. That is the whole lesson of the AlSi10Mg study: they are electropolishing lattices because nothing else reaches.
Day to day, it looks like this. A mating face gets oriented flat or facing up so we can sand or machine it. A bore that holds a bearing gets printed slightly undersize and reamed. A cosmetic face gets the orientation that hides layer lines. For high temperature parts in PEEK, ULTEM, and PPSU, the same rules apply, and post-machining is often the cleanest route to a tight tolerance surface.
What this means for engineering buyers
If you are sourcing any part where an internal surface does real work, ask your supplier how they will finish it, how they will verify wall or strut dimensions afterward, and what the acceptance criteria are. "We electropolished it" is not a spec. "Strut diameter within X of nominal after finishing, verified by sectioning" is a spec.
Most parts through our shop are simpler than a metal lattice. A bracket, jig, housing, or drone frame can almost always be oriented so the surfaces that matter are reachable. Not sure yours can be built cleanly? Send the file to info@dcadditivepros.com and we will tell you what we would change before anything gets printed. Our build and ship page covers how we quote and turn parts around.
Frequently asked questions
Can you polish the inside of a 3D printed lattice or internal channel?
Sometimes. Electropolishing and chemical or abrasive flow methods can reach surfaces a tool cannot, but the removal is rarely uniform through a lattice, so the process has to be verified against strut and wall dimensions after finishing.
Does a rough internal surface on a 3D printed part actually matter?
Yes, if that surface does work. Roughness can restrict flow, trap powder, create stress concentrations, and change thin features from their designed size. On a purely cosmetic interior it usually does not matter.
How do I get a smooth mating face on a 3D printed part?
Orient that face flat on the bed or facing up so it can be sanded or machined after printing, and tell your print shop which faces are critical before the job is quoted.
Can DC Additive Pros print parts with internal channels or hidden geometry?
Yes. We review the file first, decide the orientation that puts critical surfaces where they can be finished and inspected, and split the part into inspectable pieces when a single print cannot be trusted. Email info@dcadditivepros.com with your file.
Sources: Fabbaloo, September 7, 2026; Surface and Coatings Technology, intermittent electrochemical polishing of LPBF AlSi10Mg.