Metal powder bed design guide

Recently I’ve gotten a notable uptick in emails from engineers asking whether it’ll be cost effective to use metal powder bed fusion to make an existing (but conventionally manufactured) part. The answer is that it’s complicated, and depends largely on whether or not the part fits into the design requirements for powder bed fusion. Nine times out of ten it doesn’t, but in some cases it’s possible to adapt a design to fit - or redesign from the ground up with additive in mind. In order to help other engineers begin this process, let’s review some of the part characteristics and design guidelines for powder bed fusion.



From the outset, you must understand that powder bed fusion is a near net shape process. Like casting and forging, it’s just one manufacturing process in a larger process chain; its tolerances are just too coarse to be usable as an end part.

Most machine manufacturers and job shops will specify plus or minus .005” for a feature 1” in size, and plus or minus .002” for every inch beyond that. That means that if you design a feature that’s 6” across, it’ll be toleranced at plus or minus .015”. If that sounds imprecise, it is - but talk to your supplier and ask them to look at your design before you give up hope. In many cases there may be a way to meet your requirements.

Surface finish is a bit less flexible. Laser based processes (like DMLS) usually produce finishes in the range of 250-350 µinch Ra, but that can swing coarser in areas that need support structures. Arcam’s EBM process is even more coarse - 500-700 µinch Ra is typical, and overhanging surfaces are often left with small “warts” where support structures are removed (see this video for more details).

All told, it’s much better to think of additive as a way to make something that looks *somewhat* like the finished product you’re building - but one that’ll need quite a bit of work before it reaches its final state.

Process chain

When a powder bed fusion build completes, the first step is to wait for the chamber to cool and then remove the build platform from the machine. If you’re using DMLS, the part will be completely welded to the build platform and will need to be cut off before you can proceed (EBM parts have it somewhat easier, as they can usually be snapped off of the platform manually).

Cutting the parts off the platform can be a bit messy. Many shops use wire EDM, but it’s possible to use a bandsaw as well (you’re usually cutting through mesh support structures, so the overall force isn’t particularly large). Then, the supports need to be removed from the part itself.

This process can be a bit tedious, depending on your design. A small area of freestanding supports can be broken off by hand or with pliers, but larger blocks of supports often need to be chipped away gradually - and where that’s the case, you can expect the resulting surface finish to be less than desired.

In order to improve surface finish, a wide variety of methods can be employed. Parts can obviously be ground, filed, sanded and polished as needed. In addition, a number of specialized finishes are used on high value powder bed fusion parts. Media blasting and wet abrasive blasting are both fairly common, but micro machining (see MicroTek) and isotropic superfinishing (see REM Chem) both offer unique advantages in demanding applications.

Extremely critical parts will also undergo hot isostatic pressing. HIP is a remarkable process, which is used to eliminate voids (large and small) deep inside otherwise solid metal parts. It’s common in aerospace (both for additively manufactured and for cast parts), but many applications don't require it.

Mechanical features

Mechanical features pose interesting challenges. If your part is used on its own, then generally you’ll be fine. But if it’s part of an assembly (which - let’s face it - most parts are) then you’ll have to look at any mating or assembly features on a case by case basis.

Female threaded holes, for instance, can be problematic. Powder bed fusion generally isn’t good at creating female threads itself, meaning that you’ll ultimately need to create the threads in a secondary process. In some cases it might make sense to print an undersized hole and then tap the threads afterwards, but beware: some printed metals like titanium can be *extremely* hard to cut. As a result, it may be wiser to print a solid part, then drill and tap the part afterwards.

Surfaces that’ll be used to mate to another part will also generally need to be post processed, whether by EDM, milling, turning, or grinding. These really need to be addressed on a case by case basis, but remember that most of these processes will require the part to be clamped securely in a machine - which may affect how you design the part. Adding a fixturing boss into the design could make things a lot easier in the long run, if it allows for post processing to be done more easily.

Form and shape

In general, powder bed fusion is capable of making features whose minimum dimensions are greater than .5mm or so. This will vary somewhat depending on orientation and feature type, but as a rule I try to stay above 1mm wall thicknesses unless I *really* need it to be smaller.

It’s also important to avoid long, thin shapes. Most shops will tell you that a part’s height should be no more than ~8x it’s width, but again I usually try to be even more conservative. If my design looks like something cut from bar stock, then I’m probably doing something wrong.

Large, blocky designs are also poor choices for additive. Especially in materials like titanium, the built in stresses of laser sintered parts can cause *big* problems, and those problems build up when building parts with large cross sectional areas.


As with any manufacturing process, it’s best to develop a close relationship with your suppliers and talk to them in detail about the functional requirements you’re working towards. It’s also helpful to think in terms of those functional requirements yourself - rather than the specific mechanical features that you may be inclined to design for conventional manufacturing.

It’s also worthwhile to consider a modeling tool that’s designed with additive in mind. Topology optimization and lattice design tools can have powerful advantages when working with additive, as they avoid many of the design patterns that AM often has trouble with. And more importantly, they help designers adopt a different mindset - which is, in truth, what additive requires.

A Part Number Anthology part-number-anthology-small

Part numbering. For most engineers, this two-word phrase is all it takes to conjure up especially strong feelings about what it means to be “right”, and what it means to be very, very “wrong.” We've assembled a handful of our part number greatest hits in this eBook anthology.