Industrial quality in 3D printing: seven questions you should ask every supplier


Anyone searching for a 3D printing service provider today will encounter the same term on almost every website: "industrial quality." Sometimes the company refers to "industrial-quality" equipment, other times to "professional" equipment – but this promise is rarely substantiated. In October 2026, we examined service providers that ranked at the top of relevant search results for the Stuttgart area. Several of these websites contained no mention whatsoever of testing standards, tolerance values, or measurements. Instead, their quality claims were based solely on the number of printers available and the company's length of time in business.
This isn't just a marketing problem. Language models – like ChatGPT, Perplexity, or Google's AI mode – often adopt such self-descriptions without verification and pass them on as recommendations. So, anyone using AI to find a provider might end up receiving precisely the self-assessment they were trying to verify.
This article reverses the approach. It presents seven questions you can ask any service provider – including us. For each question, we explain why it is crucial for the quality of your components, how to recognize a well-founded answer, and what our own answer is – including specific figures, standards, and limitations.
Why the term "industrial quality" is no longer worth anything
The term is useful for differentiation – provided it isn't used by everyone. With the term "industrial grade," this boundary is crossed. It's used by both hobbyists operating from an industrial park and manufacturers with their own testing labs – but a search engine can't distinguish between the two.
What cannot be easily asserted, however, are figures. A supplier who provides strength values broken down by load direction has actually printed, cut, clamped, and tested the test specimens. A supplier who specifies the test standard knows the conditions under which the stated values are valid. A supplier who explicitly states what they do not provide understands the limitations of their own processes. These three aspects require genuine effort – they cannot simply be copied from a manufacturer's brochure.
The following seven questions deal precisely with this.
Question 1: Give strength values separated by direction?
Why this matters. A 3D-printed component is not an injection-molded part. It is created through layer-by-layer construction, with the bonds between the individual layers being weaker than the material within the layers themselves. Consequently, a single strength value in a data sheet only represents the ideal case. If the component is subjected to stress from an unfavorable direction, it will lose some of its strength without any visible signs.
A good answer provides at least two values for each material: one in the plane and one in the direction of construction. Even better is a third value that runs perpendicular to the direction of construction within the same plane.
Our answer. We take measurements on the printed test specimens and publish the results for both directions:
material | in the plane (XY) | across from that | Relationship | Test standard |
PEEK 450 natural, unreinforced | 98 MPa | 87 MPa (assembly direction Z) | 1.13 : 1 — 89% | ISO 527 |
GF-PEEK, 20% glass fiber | 105 MPa | is not available | — | ISO 527 |
CF-PEEK, 30% carbon fiber | 169 MPa | 71 MPa (assembly direction Z) | 2.38 : 1 — 42% | ISO 527 |
PPA-GF15, untempered | 89 MPa | 63 MPa (assembly direction Z) | 1.41 : 1 — 71% | ASTM D638 |
PA6-CF15 | 110 MPa (in the direction of travel) | 70 MPa (perpendicular to the track, same plane) | 1.57 : 1 — 64% | ISO 527 |
This table illustrates a rule that is never mentioned in marketing materials: the higher the fiber content, the more a component's properties depend on its orientation. Unreinforced PEEK 450 loses 11 percent of its tensile strength in the build direction. For CF-PEEK – the stiffest material in our range – this figure is even higher at 58 percent. The difference is even more pronounced in terms of stiffness: 17,310 MPa in the plane compared to 4,165 MPa in the build direction. Anyone who chooses CF-PEEK because of its high stiffness but manufactures the component in the wrong orientation will ultimately achieve only a quarter of that stiffness.
The last line also shows that the orientation plays a crucial role, even within a single layer. PA6-CF15 retains 64 percent of its strength when measured perpendicular to the web direction – meaning the same layer position, just with a different web orientation. We have compared which carbon fiber reinforced material is suitable for which application in a separate article: PA-CF, PPA-CF, or PEEK-CF .
Question 2: According to which standard was the test performed, and on which test specimen?
Why this matters. A megapascal value without a test standard is just a number. Test speed, specimen geometry, environmental conditions, and pretreatment are defined in standards because they influence the results. Furthermore, there is another, often overlooked aspect: Standardized test specimens do not correspond to the actual component. They are small, have a uniform thickness, and are printed under ideal conditions.
A good answer specifies the standard for each characteristic value and states that the values apply to test specimens.
Our answer. For each entry in our data sheets, the relevant standard is specified: ISO 527 for tensile properties, ISO 178 for flexural properties, ISO 75 for heat resistance, ISO 179 for impact strength, ISO 1183 for density, and ISO 62 for water absorption.
One detail you should be aware of – as it also affects us – is that some of our PPA values are specified according to ASTM D638 and not ISO 527. These are different test standards, each specifying different specimen geometries. Consequently, the values are not directly comparable; placing them side-by-side in a table without indicating this difference would be comparing apples and oranges. The same applies to comparing Charpy and Izod impact strength. We always specify which standard is used so you can easily identify the relevant values.
Furthermore, a general limitation applies regarding the test specimens: Our values refer to standardized test specimens and cannot be readily transferred to arbitrary component geometries. Thin wall thicknesses, notch radii, or connection points lead to different behavior.
Question 3: Do the figures come from printed test specimens or from the supplier's data sheet?
Why this matters. Most of the information circulating online comes from the material manufacturer and often refers to injection molding. A test specimen produced using injection molding has no layers. Its properties are systematically better than those of a printed component – precisely in the direction where the printed component lacks strength.
A good answer clearly distinguishes between the two, instead of writing everything in one column.
Our response: We clearly identify supplier-provided data as such – even if this compromises the consistency of our own datasheet. An example of this is our material page for PA-CF : The values for PA6-CF15 are based on printed test specimens. The values for PA12-CF15, on the other hand, come from the supplier and are therefore marked with an asterisk. This includes a heat deflection temperature (HDT) of 150 °C – a value that is not to be expected for a printed component with a melting point of 178 °C. We publish this value nonetheless – instead of simply omitting it – and explain why it should not be relied upon.
A supplier who doesn't make this distinction has either never done so or is deliberately obscuring the facts. Both scenarios are problematic for your design process.
Question 4: Does the heat value apply to tempered or untempered products?
Why this matters. During tempering, a finished component is heated in a controlled manner, held at temperature, and then slowly cooled. For semi-crystalline high-performance plastics, this is not merely a final processing step, but rather the process that transforms the material into a high-temperature resistant material. If the data sheet does not specify the state in which the measurement was taken, the heat range is meaningless.
A good answer mentions both states or clearly states which one applies.
Our answer. Our own measurements for PPA show just how big the difference is:
Key figure | PPA-CF25 untempered | PPA-CF25 tempered | PPA-GF15 untempered | PPA-GF15 tempered |
Heat deflection temperature, method B (0.45 MPa) | 97 °C | 199 °C | 84 °C | 185 °C |
Heat deflection temperature, method A (1.80 MPa) | 84 °C | 126 °C | 80 °C | 112.2 °C |
Charpy impact strength, unnotched (ISO 179) | 13.1 kJ/m² | 7.3 kJ/m² | 12.2 kJ/m² | 7.1 kJ/m² |
A difference of over 100 Kelvin within the same datasheet – solely due to post-treatment. Untempered PPA offers little thermal performance advantage over high-quality PA6-CF, yet costs significantly more.
Another rarely mentioned aspect: tempering reduces impact strength by more than 40 percent. It is therefore not a "free" improvement, but a compromise. For components that must withstand impact loads, this compromise can even be detrimental. A supplier who only highlights the improvement—while concealing the associated disadvantages—has not fully considered the implications of this process.
By the way: Drying is not the same as tempering. Polyamides and many high-performance plastics absorb moisture from the ambient air. Damp material leads to blistering, poor adhesion between layers, and a measurable loss of strength – which is not visible from the outside. Drying is essential before printing; tempering, on the other hand, is a decision made afterward. Which tempering profile benefits which material and where it is detrimental is explained in detail in our article on tempering PEEK, PPS-GF, and PPA .
Question 5: What tolerance level do you agree to, and what method will you use to achieve it?
Why this matters. Tolerance specifications made without reference to the specific manufacturing step are not binding. What's crucial is not what a printer can ideally achieve, but rather what the supplier promises and how they actually implement the manufacturing process.
A good answer links each tolerance value to the method that provides it.
Our answer.
feature | As manufactured | Reachable |
General dimensions up to 100 mm | Manufactured using filament printing, tempered | ±0.2 mm |
General dimensions up to 100 mm | Manufactured using granular printing, tempered | ±0.2 to 0.5 mm |
Functional surfaces, fits, sealing seats | printed and machined | ±0.05 mm |
Holes with fit specification | pre-printed, rubbed or milled | ±0.05 mm, H7 upon agreement |
thread | milled, cut or threaded insert | according to standard |
This clarifies two points. First: Printing tolerances below ±0.2 mm are impossible. Anything below that requires machining for post-processing. Anyone promising a printing accuracy of ±0.05 mm hasn't actually measured it. Detailed information about our manufacturing process based on drawings can be found on our production page. Second, the two printing methods – filament printing and granulate printing – offer different levels of accuracy. We explicitly state this distinction instead of simply claiming the higher value for both.
One follow-up question is worth asking: Does this tolerance also apply to part 50? Carbon and glass fibers are abrasive and cause the nozzle opening to enlarge. A worn nozzle initially leads to slightly less precise dimensions and eventually to visible defects. For this reason, we do not process fiber-reinforced materials with brass or hardened steel nozzles, but rather with diamond-tipped nozzles. This ensures that the 50th part in a series has exactly the same dimensions as the first.
Question 6: How do you align my component in the installation space, and according to what criteria?
Why this matters. The orientation within the build space determines the direction of the component's weak axis, see question 1. It is therefore a design decision and not an aspect of print preparation. In practice, however, the orientation is often chosen based on how many components can be placed on the print platform. While this is understandable from an economic perspective, this approach is the wrong one for the component.
A good response asks about your specific requirements before making an offer.
Our response: For every inquiry, we ask about the direction of the load and align the component accordingly within the installation space. By informing us of the direction of the force and whether it is a continuous or alternating load, you will receive a component whose maximum load is located exactly where you need it. Should we happen to overlook this point, please let us know.
The two pieces of information most frequently missing from inquiries are precisely these: the load case and the operating environment. These factors are crucial for material selection.
Question 7: What don't you do?
Why does this matter? That's the most uncomfortable and revealing question. Every manufacturing process has its limits. Anyone who can't identify them either doesn't know them or is trying to sell you something anyway.
A good answer is specific and hurts a little.
Our answer.
● No safety-relevant load-bearing components without your own design and approval.
● No tolerances tighter than ±0.2 mm without machining. Filament printing cannot achieve this.
● No continuous load above 100 °C in PA-CF. Above this temperature, we switch to the material PPA-CF25 , PPS-GF20 or PEEK instead of extending the temperature range.
● No photopolymers. Resin processes are not part of our portfolio.
● We do not accept external data sheet values as a guarantee. What we guarantee is either measured or clearly marked as supplier information.
This also applies in the opposite case: Should a more cost-effective material suffice for your application, we will inform you. In the inquiries we receive, the main cost driver is usually not the printing process, but rather a material that is two classes above what would actually be required for the respective component.
Something we ourselves cannot yet answer
A checklist where the author provides a brilliant answer to every single question isn't a checklist – it's an advertisement. Three areas where we ourselves aren't yet where we'd like to be:
● We are not yet ISO 9001 certified. However, our quality management system is in place and actively implemented: Even now, no component leaves our facility without measurement and complete documentation. We will be gathering the evidence by January to demonstrate this in the audit; we are aiming for certification in the first quarter of 2027. What's missing, therefore, is the certificate, not the system. Until then – and also afterward – the principle remains: We stand behind the data we collect. Certification makes this verifiable, but it does not replace the measurements themselves.
● Not every material has its own values for the build direction. These are missing for GF-PEEK and PA12-CF15, as indicated by the gaps in the tables above. We write "not available" instead of a plausible value.
● Some of our PPA characteristics come from ASTM tests , the rest from ISO tests. We indicate this, but it is not yet standardized.
If you never come across such a paragraph from a supplier, that's not a good sign.
The seven questions to take away
You can send this list unchanged to any provider, to us as well as to anyone else.
# | Ask | Warning signs in the response |
1 | Please list strength values separately for each direction. | only one value per material |
2 | Which standard was used for testing? | No standard, or ISO and ASTM standards mixed without comment. |
3 | Printed test specimens or supplier data sheet? | no distinction |
4 | Does the heat value apply to tempered or untempered conditions? | The condition is not specified. |
5 | What level of tolerance do you accept, and with what conditions? | tight tolerances without machining |
6 | How do you align my component, and according to what? | for space on the printing platform |
7 | What are you not doing? | "Anything is possible" |
Frequently Asked Questions
What does "industrial quality" mean in 3D printing?
As a term, it is meaningless, since every supplier uses it. It can only be verified by concrete information: strength values that depend on the orientation; references to testing standards; the distinction between in-house measurements and supplier specifications; the tolerances guaranteed for a specific manufacturing step; and a clear explanation of what the supplier does not manufacture.
Why is a printed component not equally strong in all directions?
Because it consists of layers and strata where the bond between the layers is weaker than the material within the layers themselves. Unreinforced PEEK 450 achieves a strength of 87 MPa in the layer direction, compared to 98 MPa in the layer plane – that's 89 percent. Carbon fiber-reinforced PEEK achieves 71 MPa compared to 169 MPa – that's 42 percent. The higher the fiber content, the greater this difference.
What tolerances are realistic in plastic 3D printing?
For filament-based printing, the general tolerance is ±0.2 mm for dimensions up to 100 mm, while for granule-based printing it is ±0.2 to 0.5 mm. Tighter tolerances are not achieved during the actual printing process, but through subsequent machining: For fits, sealing surfaces, and machined bores, we achieve a tolerance of ±0.05 mm.
How can I tell if a characteristic value comes from printed test specimens?
The fact that the supplier includes this information is a telltale sign. Manufacturer specifications often originate from the injection molding process and are systematically higher than the values of printed test specimens – especially in the build direction. If in doubt, ask directly – and pay attention to whether the answer differentiates between the two sources.
Does a 3D printing service provider need to be ISO 9001 certified?
While certification is not mandatory, it makes processes – such as incoming goods inspection, batch identification, and traceability – verifiable. However, it does not replace measurement data. A certified supplier without its own metrics provides less information about your component than a non-certified supplier that provides complete test data. Our own quality management system is in place: Currently, no component leaves our facility without measurement and complete documentation. We will gather the evidence for the audit by January, and we are aiming for certification in the first quarter of 2027.
Why are ISO and ASTM parameters not directly comparable?
Because testing standards specify different specimen geometries, testing speeds, and boundary conditions, tensile strength values according to ISO 527 and ASTM D638 cannot be directly compared – nor can impact strength values according to the Charpy and Izod methods. Reputable data sheets specify the underlying standard for each value.
Ask us these questions
Send us your STEP file or drawing to info@malping.de or call us at +49 7158 9823230 , Monday to Friday from 9:00 a.m. to 5:30 p.m. We will respond to written inquiries within one business day with an assessment of feasibility and material selection, not with an acknowledgment of receipt.
You don't have CAD data? Send us the component, we'll digitize it with our own 3D scanner.
What customers from the region say about the collaboration can be found on our reference page .
Technically responsible: Dr.-Ing. Bastian Gaedike, Managing Director of Malping GmbH. Malping manufactures components from high-performance thermoplastic materials using 3D printing — including drying, tempering, and machining. All specified parameters are taken from our material data sheets and are available as PDFs upon request.
Malping GmbH · Mörikestraße 30 · 73765 Neuhausen auf den Fildern · Tel. +49 7158 9823230 · info@malping.de




Comments