Ensuring dimension accuracy of as-built metal components built using EOS M280 (1/2)

Process parameters development is a rigorous multi-step process paramount to manufacturing repeatability and reliability of production. Amongst ensuring high density, suitable surface roughness and reliable mechanical properties, dimension accuracy and precision must also be guaranteed as part of the ideal (ie no mechanical post processing involved) AM process. Here, we review the critical factors that can be used to ensure dimension accuracy. 

Various factors can be responsible for low dimension accuracy of as-built SLMed components:

1.       Large surface roughness makes it difficult to measure dimensions accuracy;
2.       Internal stresses can distort the product;
3.       Incorrect laser processing parameters or machine settings can cause discrepancy between CAD and SLMed product dimensions;
4.       Digitalisation (mostly for small features such as holes)

In addition, accuracy susceptibility varies with the type and size of features to be processed.

This is why understanding the machine and laser/material interactions is critical. 

Here we assume surface roughness is Ra<10um and that internal stresses are minimal to discuss processing parameters and machine settings needed to ensure high dimension accuracy of large components (ie small features accuracy will be discussed separately).

Processing parameters: understanding actual track width

Irradiation time and energy density are the key factors involved in creating sound metal components using Selective Laser Melting. These can be addressed indirectly using production machine SLM parameters such as Power P, scanning speed s, hatching distance h, layer thickness th,…

It is critical to point out the impact of power/speed combination on the actual track width (and on the track depth, and on the melt pool maximum temperature and solidification rate hence on the density, microstructure and mechanical properties, etc but let’s keep that for later!). In other words, a laser track can be compared to a scalpel with a variable blade thickness that varies as a function of laser power and speed (function of layer thickness – usually 20/40um for maximum output power limited to 370W).

Cross section image of the fusion zone for different laser power levels at 300, 500, 750, and 1000 W [1]

Based on the material affinity with the laser source – usually radiating in the infra-red although some manufacturers have come up with lower wavelength to accommodate highly reflective material in the IR, like copper – for instance its absorptivity and susceptibility to melting (linked to thermal conductivity as a function of temperature and physical state) the resulting melt  pool shape varies. Melt pool shape is defined as a transversal x/z cross section of a melt track melted in each x/y plane parallel to the build platform. 

Contours and offset parameters

Now let’s assume that high density processing parameters may not be suitable to ensure great surface roughness and high build rate. What this means is that different sets of parameters are required to ‘clean’ the outer surface of your component: to ‘tidy up’ the unmelted satellites stuck to resolidified walls of the components.

That’s where knowing the actual track width of the contour parameters (ie: power + scan speed scanned on the CAD contour of the component) comes into play. It is necessary to offset this contour scan by (usually) 30% to 70% of its width inwards to ensure that the overall dimensions of the components are within tolerances.

Track width (+/- surface roughness) can be measured after building  single tracks and measuring their thickness (keep in mind the width of thin wall is slightly different from the width on track built within large sample wich you can cut across and etch and examine under microscope). Or you can predict it using modelling (assumptions made will limit width value accuracy). Both techniques should give you a ballpark number good enough to use as a starting point for offsets values.

With this in mind, the parameters responsible for dimension accuracy accessible on the EOS M280 are:  offsets values, contours parameters, track-width, skywriting and surface roughness parameters. More details next week.

References
[1] Lijun Han, Frank W. Liou, Srinivas Musti - Thermal Behavior and Geometry Model of Melt Pool in Laser Material Process, Journal of Heat Transfer, SEPTEMBER 2005, Vol. 127 / pp1005-1014