• Scale up productivity by scaling up the laser
    spot size of a LPBF-System

Authors: Dominik Wolf & Michael Jan Galba

Unleash the potential of your today not profitable AM-applications due to too low cost per part requirements.

 

Within serial production only two things count:

  1. Manufacturing the part at the same cost as conventionally or lower
  2. and at the same quality, with the same reliability as traditional production methods.

It is as easy as that, when both are not given the likeliness an application makes it into serial production is very low. Besides the direct economic benefit, with AM you would gain enormous freedom in design and flexibility of your production chain e.g. produce spare parts fast on demand without big storage capacities. But one of the biggest influencing factors on cost per part is still the productivity of the AM-System.

My colleagues and I from the innovation team at EOS developed and investigated a new light engine for 3D metal printing, which showed the potential to reduce the building time of up to a factor of 5x.

 

Enhance your toolset to gain productivity

The current standard additive manufacturing process is a very precise process for high quality demands but has adjacent its limitations regarding big volume parts.

Compare it to painting a large room with a small brush. You would reach any corner accurately, but it would be exhausting and cumbersome to paint all the plain surfaces. Obviously, a paint roller would get the job done in no time. Transferred to the world of additive manufacturing it is a logical consequence to extend the set of the existing tools. In case of laser powder bed fusion, it is the shape of the laser beam interacting with the powder, which is a main influence factor on the melting process.

My colleague Anoush from EOS Innovation Managment wrote a comprehensive overview of the multiple possibilities to manipulate the laser beam and its potential value for the additive manufacturing of metal parts. Producing laser spots of any size and shape in an AM-machine is no longer an engineer’s dream. It is already possible to today, which will soon be explained by another team member Markus. 

A more straight forward solution, already fitting the needs of an industrial production environment today, is to scale up the size of the Gaussian laser spot diameter. With said changes it is possible to apply more energy at the same time into the material to speed up the melting process and to make use of the continuously increasing power of state-of-the-art laser systems.

But besides the big spot you would still want your precision spot to realize sharp edges and fine details. The tool set of choice needs to be flexible to pick the right spot for the different application demands whilst withstanding the high laser power needed for the productive processes.

Our innovation team at EOS recently developed together with a partner a new cutting-edge optical system, which can form multiple kilowatts of laser power in a continuous range of spot sizes. (Here installed on an M280 lab system)

The challenge was to preserve the intended spot diameter and intensity distribution over the whole working range of the laser. A deviation would cause process fluctuations and in a consequence material defects in the produced part. But due to ultra-low absorption optical components the influence of multiple kW laser power on the intensity distribution in the working area was reduced by a factor of 4. Additionally we are working on a innovative and from the scanner independent inline measurement system to monitor the current performance of the optical train. With this approach it would not only be possible to erase the last residues by feedback loop control, but it would be also a huge advantage to certify the production system for high end industries. E.g. in aerospace the standard "SAE AMS 7003 Laser Powder Bed Fusion Process" requires the user to monitor the optical path of his AM-System throughout the production process to ensure the necessary quality level.

To reach the desired maximum flexibility for process and application development the process software was modified - this allows a direct allocation of parameter specific spot sizes. Now it is possible to set individual spots for different parameter sets and even make adjustments for every exposure type e.g. contour, infill, downskin, upskin etc. to optimize the energy input for every part region.

But how does this new light engine perform on an AM-machine? To find out the potential for the AM-process, we installed the system together with a 1kW-laser on an EOS M280 lab machine, which fits perfect for fast and flexible process experiments. Next step was to conduct several experiments including building density cubes in 316L with a wide variation of the key process variables like spot size, laser power, scan speed and hatch distance. The layer thickness was set to 80µm. After analysis of these samples in our EOS lab, we got an overview about the achievable increase of build rate and the resulting material density. In the high-quality range of today’s standard 316L parameter it was possible to get a volume rate increase of factor 3x and still getting fully dense material. Taking the quality level of various casting processes as a reference it was even possible to boost the factor to 4-5x. And these are certainly not the final results, as there a lot of ideas and improvment potentials to push the boundaries even further. We'd like to highlight, that it is important not to involve too many options at the same time. Otherwise the allocation of cause and effect would become extremly tricky. EOS does have at the back of the hand very innovative exposure patterns as well as optimizations and finetuinings which are today not available as commercial products. There's more to come soon and please stay tuned.

Stepping from the lab into the real world

Test specimens are necessary to evaluate the material properties, but the part shapes are very simple and don’t reflect the spectrum of intended application designs. Different surface angles and wall thickness can have a heavy influence on the buildability of unproven parameter sets. To benchmark the new technology a demonstrator representing a small imitation of a drill head was built with one of the most promising parameters at maximum productivity.

Those drill heads are used to produce wellbores to access water, thermal energy and other resources underground. It consists of a steel bit attached with cutters made of extremely hard tungsten carbide. The bit is a very bulky part and therefor fits well to test a high volume rate parameter.

The first building trial was already successful and the waiting time to unpack the final part was very short as expected. That’s the fun part about developing a high productive production system. You don’t have to be very patient to get feedback on your test results!

The surface of our demo part is already quite satisfying and has still a great potential for optimization due to a lot of unused tuning options for up- and downskin parameters.

The build of a drill bit of the size of 250mm filling the whole building platform of an M290 would normally take >400h which equals approximately 17days uninterrupted printing time, for one part! And just as a side note, we do have customers today with applications in this job run times... . Using the new parameters, the build job would already end after 82h which represents a time reduction of 80%. This would at least cut the cost by half compared to a standard setup.

 

Combination of productivity and fine details

The recoating process went so smooth that we already took the next challenge to try a segmented building process. To unite the resolution for fine detail sections and high productivity in bulkier areas it is necessary to combine the use of different powder layer thicknesses in one part. In this procedure the CAD-file is split in the specific segments and applied with different parameters in the process software.

The test part for the segmentation approach is a small imitation of an impeller which is used e.g. in water pumps. For the realization of the thin blades, the standard parameter with a layer thickness of 40µm was chosen, while the bulky segment was again built with the highly productive parameter developed with a doubled layer thickness of 80µm.

Conclusion

After starting with a first feasibility study on 316L we can state that there is the huge potential in using a bigger spot size combined with high laser power to speed up the additive manufacturing process by a factor of up to 5x and to reduce cost per part significantly. Here we used a M280 for pre-development purposes. Moving to industrial serial production, the EOS M400, EOS M404 or M304 would be the machine of choice. Of course those machines had been taken into consideration during the development phase of the new optical system and therefore it is a small step for implementation. Besides of the benefit of automation there would be an additional building time reduction by double recoating coming with these production machines and in case of the multi-head systems another multiplier of 4 on the build rate. So in a nutshell, the technology is easily scalable to quattro head machines. Having this in mind you'd be resulting in a productivity gain of a total factor of 12-20 in direct comparison of a single head 400W system!

Thanks to all the colleagues involved contributing to that great success!

At EOS we strive for perfection in helping our customers with best in class available technology. The EOS Innovation Management team has the task to develop the products for tomorrow and the day after tomorrow. The maturity level after the handover to series product development lies typically within a technology readiness level from 3 to 5. This said, the technology is today commercially not available for the wider market but can potentially be made accessible through a cooperation agreement.