Inside Metal Additive Manufacturing
  • About
  • Blog
  • Contact

Powder bed AM machines: ask before you buy

8/12/2017

0 Comments

 
Picture
Pump made by powder bed fusion AM [1]
Have a peek under the hood of your powder bed fusion machine and here’s what you get: infra-red laser(s), powder reservoirs and overflows, a spreading mechanism, some form of atmosphere control in the build chamber and a software handling digital files and controlling the process sequence.

Fundamentally, there are no processing differences between suppliers' offers!  

For the price of a house, or a luxury car, you get a machine that will be responsible for the reputational credibility of your company and the commercial viability of your production. Yet, it rarely comes with a builders’ or after sale guarantee. 
So it’s best to make sure you have all the unbiased facts at hand to invest your cash sensibly.

​In the next 3 articles, we’ll review the points you will want to keep in mind before you invest in AM technology. Hardware, ease of use, maintenance and end-products quality are a few topics you’ll want to discuss in detail with the suppliers. In the first part, let’s have a look at the hardware itself and what you should pay attention to.

​Machine robustness and reliability, technology maturity

Technical development of AM machines is getting faster and faster. Manufacturers compete for a slice of the market by releasing new models with genuine improvements or sensational gimmicks (such as useless monitoring systems acquiring too much data for actual processing use). New products on a stringent delivery timeline may get launched straight off the R&D line. Unfortunately, proper testing might not have always been carried out satisfactorily. Don’t be one of those trusting and unsuspecting customers who end up with a prototype without being aware of it...

​Machine reliability, maturity of the equipment

​This point is naturally underestimated because you’d assume most manufacturers sell tried-and-tested products. Well, most do. But not all of them. So to be on the safe side, first things first: let’s establish how robust the hardware is and how reliable the control software performance is.

A good clue is in the maturity of the product: how long has it been out? Are current customers happy with it? And could you talk to them? And don’t confuse the official launch date of the unit with the actual release date of the usable product. Products might only just be launched as an PR coup during an exposition, in order to ‘keep up with the Joneses’ of AM. But is it actually for sale?! Does the delivery date keep being pushed back? Be aware that lapse of time between launch date and sale delivery of a robust product might unfortunately extend up to a couple of years… And if it's a new product: can you try before you buy? More dedicated centres mushroom around he world where you can do just that.

​Besides, it’s not unheard of for the company to have few versions of the ‘same’ product floating around so you want to make certain of what the hardware version is. 
​Of course, you can also agree to getting the 1st prototype and carry out beta testing, at a reduced acquisition price. Are you happy to iron out and feedback the bugs as they arise? Or will it cost you money? The downside of this option is you’ll spend months tweaking the machine before producing anything reliably, the upside is that the engineers in your team will become experts at using, trouble-shooting and making the most of the equipment. 
At any rate, and to protect yourself, wouldn’t it be sensible to put a close in the contract that states whether you can get your money back (with interests?!) or send the machine back should the production quality were not acceptable?
Picture
ref: GE gas and oil

​Processing parameters sets available

What are the processing parameters supplied with the hardware? How comprehensive are they? This is another important point. How were they developed? Using 15cm³ cubes or realistic components? Usually, manufacturers provide parameters sets that cover various materials. In the best-case scenario, these parameters can even be divided into subsets specifically tailored for various aspects such as production speed, final surface roughness aspects or mechanical performance. 
​For instance, you could have access to ‘Aluminium A357’ parameters, subdivided into ‘A357/Speed’, ‘A357/Surface’ or ‘A357/Performance’. This puts you in control of the production regime. Datasheets describing the output properties accompany these parameters. They detail output characteristics such as average surface roughness values, density and a few mechanical properties usually tested at room temperature. Here, there are a few caveats to keep in mind and you may want to enquire about the limitations of and guarantees associated with these datasheets. Are they based on or limited to specific powders characteristics (particle size, composition…)? Ask for precisions such as the difference between upskin, downskin and vertical surface roughness values (if these are important to you!). How did they measure the density (optical or Archimedes method)? How do the values vary across the build platform or across machines? Besides, can the manufacturers offer a rough production yield, albeit for a certain geometry? Note it’s actually impossible to generalise production yield (PY) for a large range of components geometries. PY still relies too much on the expertise of the engineer preparing the digital files.  But the suppliers could give you a representative geometry they use to assess this aspect. Alternatively, you could ask them to produce one of your typical (or most complex!) components to assess their standard processing parameters.

​Powder usage efficiency

That’s a big point in term of running and production costs. It is detailed here.

​Longevity

How long do the sub-components last? Do they have data to provide you? How about the laser? How often do you have to replace it? How long before the output powder decreases?

This aspect will be detailed in the 2nd part dedicated to ‘use and maintenance’.

​Effective capacity

That’s another big one. Imagine the brochure states the build chamber size is 25x25x30cm; you would expect the effective build volume to be… well 25x25x30cm right? Wrong. The effective build volume will be smaller. Why would that be? There are a few reasons: i) the laser scanning field may be setup and/or calibrated over a smaller area, ii) in some instance, the actual powder reservoir size may be too small for you to take full advantage of the maximum height of the build chamber… Yep, you read that right… Now, certain machines are designed to compensate for this drawback by allowing the operator to refill the reservoir during processing without stopping the build. Others integrate a closed loop whereby the powder found in the overflow is automatically sieved and used to refill the reservoirs during processing, without any operator intervention. In others, unfortunately, powder shortage will cause the process to stop for you to top up the powder tank in order to restart and finish the process. Needless to say, it is not the most satisfying scenario…
​
The maximum build height will also be affected by the thickness of the build substrate. Thick (~30/50mm) platforms are used to prevent warping during build when large, chunky components are melted. Standards platform are usually ~15/25mm thick depending on the machines suppliers.  Keep that in mind when you prepare the digital files and choose the best orientation, and calculate the numbers of parts you can build in one run.

​Resolution vs build speed

Features size in the XY plane

The fine features of your products portfolio will dictate the resolution you may need. The minimal resolution of PBF machines is directly derived from the laser track width in the XY plane. As a rule of thumb, theoretical lasers spot sizes oscillate in the 50/500um range depending on manufacturers. Hence, actual laser weld width, which vary with processing parameters, will scale from ~70/900um diameters upwards.
​

Z-direction resolution

PBF technology relies on successively melting layers upon layers. Each layer is a thin cross-section of the part derived from the original CAD data. In the physical world, each layer must have a finite thickness and the resulting part is an approximation of the original data. Thin layer thickness values tend to minimise staircase effect and results in a final part very similar to the original. The layer thickness also determines factors like the accuracy of the final part plus its material and mechanical properties. It also determines build speed and how much post-processing is required. The smaller the layer thickness, the lower the surface roughness value. Inversely, the larger the z-step, the faster the build rate.

Note: a few manufacturers use high precision positioning scales, recoating mechanisms or elevators. But what’s the point of a 0.010um Z-step resolution (=layer thickness) if the best surface roughness achievable varies between 5 and 50um…
​

​Atmosphere control and condensation

Atmosphere control in the build chamber and the power reservoirs is important to prevent oxidation and 02 levels variations during build. It is critical for highly reactive powders - for instance, with titanium alloys, too little AND too much O2 will ruin your mechanical properties so that you want to work in the suitable O2ppm range. 

One or two sensors are usually directly positioned in the build chamber. It may be advisable to enquire the sensing accuracy of the O2 sensors and what the control loop consists of. Do they record the O2 levels or help actively control the build environment? Does the machine come with spares readily available when one packs in? These sensors are susceptible to being coated/dusted by fine condensate that may be generated during process. In that respect, you may want to enquire about a safe cleaning procedure and frequency.

Condensation can occur during processing. It is usually noted on the inner walls of the build chamber. Obviously, this is to be avoided as the presence of H2O in the chamber could contaminate the material and ruin the properties of your components. Have a chat with your supplier to assess if they’ve noticed this and how they’ve fixed the issue.


​Processing at high temperature

Depending on your products, you may benefit from high-temperature build platforms. Building at high-temperature may minimise residual stresses in large components for some materials (eg: some nickel alloys and steels) and has avoided the apparition of distortions and cracks.

Commonly, electron beam PBF setups can reach build platform temperatures of up to 800C whereas laser based setups usually reach up to a maximum of 170/200C, yet more suppliers are working towards providing higher-temperature options. It’s worth asking how high the build platform temperature rises. If you don’t need this option now, would it be possible to upgrade at a later stage?
​

These hardware aspects cover the key points you want to discuss with suppliers before buying.  In the next part, we’ll take a look at the operability and maintenance aspects of powder bed fusion machines.
​

References:
[1]: ​Alexandros Beiker Kair, Konstantinos Sofos, Additive Manufacturing and Production of Metallic Parts in Automotive Industry, 2014
0 Comments

Your comment will be posted after it is approved.


Leave a Reply.

    We provide practical and actionable info dedicated to additive manufacturing of high-value metal components

    I want to know more about metal AM

    * indicates required

    Categories

    All
    Applications
    Background
    Certification
    Design
    Heat Treatment
    Laser Direct Deposition
    Materials
    Modelling
    NDE
    Parameters
    Post Processing
    Powder
    Production
    R&D
    Review
    SLM
    Surface Roughness
    Surface Treatment
    Technology

© Copyright 2014-2018 InsideMetalAdditiveManufacturing.com - All rights reserved | Privacy Policy | Terms of Service | Powered by TMS