Just published an article on dimensional accuracy in 3D printing – with a dedicated section on the newly announced Bambu Lab H2D
I recently published an article that dives deep into dimensional accuracy in 3D printing, covering everything from theoretical tolerances to real-world challenges like filament quality, thermal effects, and machine limitations. The main focus is on helping designers and makers understand why parts sometimes don’t come out exactly as expected — and how to design smarter around those realities.
While the article wasn’t originally written about any specific printer, something interesting happened during the writing process: Bambu Lab officially announced the H2D. Given its focus on precision and its impressive spec sheet, I felt it made perfect sense to include a full dedicated section at the end of the article, discussing what the H2D brings to the table regarding dimensional accuracy.
Here’s what you’ll find in the article:
- An overview of what dimensional accuracy really means in FDM/FFF printing, beyond marketing numbers.
- The main factors that affect accuracy in practice – nozzle temp, humidity, cooling, tolerances in design, material behavior, and more.
- Why larger machines often struggle with tighter tolerances, and a reflection on the trade-offs between size, precision, and cost.
- A comparison to high-end industrial CNC machines, to give context to what true micron-level accuracy entails — and why it costs so much.
- My take on the Bambu Lab H2D, added as a new section, analyzing the Vision Encoder, auto-calibration, flow control, and what its design choices may imply in terms of precision.
If you’re interested in why parts fail to fit, how to design around printer limitations, or just want a clearer picture of what accuracy means in real-world 3D printing, I think you’ll enjoy it.
Read the full article on Medium
This blog is a personal project I’ve recently started, and it’s still in its early days. Although it’s hosted on Medium, all of my content is completely free to access.
I’m mentioning this because I’d really love to grow my subscriber base, and I know that when you go to subscribe, Medium often suggests a monthly payment. But I want to make it clear: you don’t need to pay anything to read my blog.
There are, of course, well-known authors on Medium whose content is behind a paywall — and many of them are worth it. I just want to share what I know, and if you find it useful, your support through a free subscription already means a lot to me!
So here it is, most of it:
One topic that often generates questions — especially among more technical users — is dimensional accuracy.
This topic was partially covered in the article: Filament Flow Calibration: One of Many Aspects to Dimensional Accuracy in 3D Printing, where my goal at the time was to provide some basic information on the subject. Today, we’ll take a deep dive into the topic of dimensional accuracy in 3D printing.
Introduction
For designers, makers, engineers, and hobbyists alike, achieving reliable dimensions isn’t just about machine specs or slicer defaults. It’s about understanding the complex interactions between hardware, materials, temperature, motion, and geometry, and using that knowledge to make better design and printing decisions.
In this article, we’ll explore what really influences print accuracy, what you can reasonably expect from a well-tuned desktop printer, and what practical steps you can take to design smarter, calibrate effectively, and adapt flexibly to different printing scenarios.
Whether you’re creating prototypes, functional parts, or commercial-quality models, this guide will help you rethink how you approach accuracy — and why that might just change how you design for 3D printing altogether.
1. What Affects Dimensional Accuracy?
Dimensional accuracy in 3D printing is not determined by a single factor, but rather by the interplay of multiple systems working together. Understanding these systems is crucial for identifying where inconsistencies come from and how to reduce them. Broadly speaking, the most important contributing factors fall into three categories: motion precision, material flow behavior, and post-processing deformation. Let’s take a closer look at each one.
1.1 Motion Precision
The first and most obvious factor is the mechanical precision of the printer itself. How accurately a machine can move its print head or build plate directly impacts the fidelity of the final print. Several sub-factors play into this:
- Condition of the Machine: Over time, mechanical parts such as belts, linear rails, bearings, and motors can wear out or loosen, reducing precision. Regular maintenance and calibration are essential to preserving motion accuracy.
- Installation Rigidity: The stability of the printer’s physical setup also matters. A machine placed on a wobbly table or exposed to vibration from nearby equipment may experience inconsistent movements, especially during high-speed travel.
- Motion Algorithms: Modern 3D printers often rely on firmware-level optimizations like input shaping or vibration compensation. These algorithms can significantly improve print quality by predicting and correcting for mechanical imperfections, but their effectiveness varies depending on the implementation and firmware version.
1.2 Material Flow Behavior
Even with perfect mechanical motion, a 3D printer can still produce inaccurate parts if the material flow isn’t precisely controlled. Thermoplastic filaments show non-Newtonian behavior within certain limits. This makes extrusion consistency a real challenge.
Molten thermoplastics (such as PLA, PETG, ABS, etc.) are non-Newtonian fluids, more specifically of the pseudoplastic type. This means their viscosity decreases as shear rate increases, a behavior known as shear thinning. In other words, the faster the material is pushed through the nozzle, the more easily it flows.
Key aspects that affect flow precision include:
- Extrusion Stability: The ability of the extruder to maintain a consistent flow rate depends on the quality of the gears, stepper motor, and feeding path. Inconsistent force can lead to over- or under-extrusion in specific layers or sections.
- Temperature Control: If the nozzle temperature is too high or too low for the material, it may result in poor flow characteristics or delayed solidification, which can distort dimensions.
- Moisture Content: Moisture absorbed by the filament can lead to bubbling, stringing, or inconsistent extrusion. Drying and proper storage of materials are essential steps for maintaining accuracy.
- Printing Speed: Faster printing means higher demands on both the extruder and the heater block. At high speeds, filament may not melt uniformly, leading to flow variation and dimensional drift.
- Nozzle Movement Speed: Sudden changes in speed or direction during printing can result in pressure build-up or lag in the nozzle, causing slight deviations in extrusion amount at corners or infill transitions.
1.3 Post-Processing Deformation
After material leaves the nozzle, it continues to behave dynamically. This phase — often overlooked — can introduce significant dimensional changes due to thermal and material-specific behaviors.
- Die Swell: As molten filament exits the nozzle, it slightly expands due to pressure relief. This “swell” can add a small but measurable offset to part dimensions, particularly in over-extruded sections or corners.
- Cooling Shrinkage: All thermoplastics shrink as they cool, and the rate of shrinkage varies based on the material, geometry, infill pattern, and cooling rate. For example, PLA does not shrink to much, while materials like ABS or nylon can shrink significantly more.
Source: Plastic Materials | Free Online Database for Plastic Industry
- Environmental Conditions: Room temperature and humidity also play a role, especially with hygroscopic filaments. Sudden drafts or fluctuations in ambient conditions can lead to warping, layer separation, or unpredictable contraction.
Together, these three categories — motion precision, material flow, and deformation during cooling — make up the core framework for understanding dimensional accuracy. The challenge is that these factors don’t operate in isolation; they influence each other in subtle but important ways. Only by considering the entire system can we begin to improve and optimize dimensional results in real-world 3D printing.
2. Dimensions and Geometry: It’s Not Just Size That Matters
When discussing dimensional accuracy in 3D printing, it’s easy to focus solely on overall size — how closely the printed part matches the target measurements in X, Y, and Z. But the issue goes deeper than just dimensions. Geometric fidelity — the ability to accurately reproduce the intended shapes and forms — is just as critical, and often more complex to control.
One of the most common areas where geometric accuracy breaks down is in small holes and internal features. While your outer dimensions might fall within acceptable tolerances, the fit and function of a printed part can fail due to slight distortions in fine geometry. And can a smaller diameter nozzle help, check chapter 6?
2.1 Small Holes and Internal Features
Holes — especially circular ones — are particularly vulnerable to deviations during printing. Several factors contribute to this:
- Layer-based approximation: Circular features are built up layer by layer using straight toolpaths. As a result, small holes can become slightly elliptical or polygonal due to the layer resolution and printer movement limits. If somo of you use AutoCAD or in fact, any CAD software know this effect, the polygonization of circles and arcs, Exporting the geometry with a low polygon number “resolution” STL only makes it worse.
- Over-extrusion at corners or seams: During extrusion, excess material can accumulate at segment junctions, leading to bulges near or retraction points — often found at the edges of holes.
Seam lines
- Shrinkage and cooling effects: When printing small internal features, material tends to cool faster than in bulk areas. This rapid cooling can cause uneven shrinkage, particularly in confined geometries, altering the final shape.
- Undersizing trend: A well-known behavior in FDM/FFF printing is that holes tend to come out smaller than designed, especially when using default slicer settings and generic flow parameters. The extent of this undersizing varies depending on the material, print temperature, and part orientation.
2.2 Shape Integrity Around Seam Lines and Thin Sections
In addition to holes, seam lines and thin features also present challenges for geometric accuracy. The start/stop points of each layer (seams) can create localized buildups or slight offsets, especially if the printer isn’t perfectly tuned or if the slicer settings are too aggressive.
Thin walls or delicate features may deform during printing due to:
- Thermal stresses from rapid cooling or uneven heat distribution
- Excessive retraction or pressure release at the nozzle
- Inconsistent flow at low speeds where precise material control is more difficult. But high speed has is challenges too…
These effects may not drastically alter the overall size of the part but can compromise its mechanical function, aesthetic quality, or fit with other components — particularly in snap-fit designs or articulated parts.
2.3 A Closer Look at Scale Dependency
Interestingly, the severity of geometric inaccuracies often scales non-linearly with feature size. A 0,2 mm distortion might be negligible in a 100 mm cube but devastating in a 5 mm diameter hole. This is why small design features demand closer scrutiny when evaluating print accuracy. The same printer and material combination that performs well for large functional parts may fail to meet tolerances for miniature components without specific adjustments.
Dimensional accuracy is more than just matching width, depth, and height. It’s about maintaining shape integrity, particularly in small, detailed features where tolerances are tight and any deviation can cause a part to malfunction or fail to fit. When testing your printer or validating a model, don’t just measure outer edges — inspect internal details, holes, and seams, as these are often the first places where accuracy breaks down.
3. The Problem with Comparing “Printer Accuracy”
When people talk about 3D printer accuracy, it’s common to hear statements like “This printer has 0.05 mm precision” or “Machine A is more accurate than Machine B.” While these claims may be based on manufacturer specs or limited test prints, they rarely tell the full story. In practice, comparing printer accuracy without context can be highly misleading.
Why? Because dimensional accuracy is not defined solely by the mechanical capability of the printer. It’s also shaped — sometimes significantly — by material properties, slicer settings, environmental conditions, and even part geometry, some of this I’ve already mentioned on this text. Two machines with theoretically identical accuracy specs can produce very different results under real-world conditions.
3.1 Accuracy vs. Tolerance Range
Before we go further, it’s important to clarify what kind of “accuracy” we’re talking about. There are two key concepts:
- Deviation from nominal: How close the actual dimension of a printed part is to the intended value in the 3D model (e.g., is a 30 mm cube really 30 mm?).
- Tolerance Range: The range of variation across multiple prints or across different parts of a single print (e.g., how much deviation is there across ten identical cubes printed under the same conditions?).
Sometimes, a printer might consistently produce parts that are slightly too small or too large — but within a very narrow range. This could be acceptable or even desirable, as it’s predictable and can be compensated for. On the other hand, a printer that occasionally nails the exact dimension but shows wide variation across prints could be far less reliable in practice.
3.2 Material Behavior Adds Complexity
Even if two printers use the same model file and same slicer profile, the results can diverge simply because of the material used. Thermoplastic filaments behave differently depending on their:
- Base polymer (PLA, ABS, PETG, etc.)
- Color additives and fillers (e.g., black and white filaments often contain more inorganic pigments, affecting flow and shrinkage they area worst)
- Moisture content (a hygroscopic filament can behave inconsistently if not properly stored)
- Batch variations (even two spools from the same brand may differ slightly)
So, if one user reports that their machine prints a hole perfectly at 5.00 mm and another says the same hole prints at 4.75 mm, the difference might not be due to the printer at all — it could be due to the filament, or even ambient temperature and humidity.
3.3 CAD Model Reproducibility and Design Tolerance
Another hidden factor in these comparisons is the tolerance built into the model itself, as already mentioned on chapter 2.1. Not all STL files or CAD designs are created with universal compatibility in mind. Many models are designed based on one machine’s behavior, filament shrinkage, or slicer defaults, which means they may print perfectly on one setup and poorly on another.
This is especially true for functional parts, such as hinges, gears, snap-fits, or sockets. If a designer assumes a 0.2 mm clearance is enough based on their personal printer and filament, others using the same file might encounter friction, warping, or non-functional assemblies.
However, a considerable number of people either don’t know or refuse to acknowledge these factors, often being very critical of certain projects I create — which are, for the most part, functional models.
https://makerworld.com/en/@JorgeRui
What I’m about to say might sound controversial or even come across as lacking modesty, but on some social platforms — especially Reddit, which I used to use to showcase my work — I was often targeted by a group of highly motivated “fans” seemingly intent on discrediting my work.
Some of them may have had bad experiences with one of my models, unaware that the failure might have actually been on their end. That said, like anyone else, I’m not immune to making mistakes. And whenever an issue is due to a flaw in the digital model itself, I always try to correct it as quickly as possible.
3.4 Why “Same Accuracy” Doesn’t Guarantee Same Results
Let’s say two users have the same printer model, with the same nozzle, firmware, slicer profile, and filament. Even then, subtle differences in setup or usage — bed leveling, flow calibration, machine wear, ambient temperature — can introduce enough variation to affect the final dimensions. That’s why even “identical” printers don’t always produce identical parts.
This variability is one of the reasons why, in recent projects, we’ve started recommending more forgiving design strategies — use of magnets with generous clearance, features that tolerate shrinkage, and other geometry tweaks that make printed parts more adaptable to different materials and machines.
4. Case Study: Real-World Testing with a Bambu Lab P-Series Printer
To move beyond theory and general assumptions, A real-world test was made to evaluate dimensional accuracy under controlled yet realistic conditions. The goal was to quantify what kind of variation we can expect from a well-maintained, mid-life consumer-grade printer using standard print settings.
Here’s a breakdown of the setup and the results.
4.1 Test Setup
- Printer model: Bambu Lab P-Series (approximately 3,000 hours of operation)
- Firmware version: Latest as of 2025–03–12
- Nozzle and layer height: 0.4 mm nozzle, 0.20 mm layer height
- Material: Bambu Lab PLA Basic — Color: Pink (properly stored and dried)
- Slicer: Bambu Studio with default “Standard” profile for PLA
- Environment: Room temperature between 20–25°C, relative humidity around 60%
- Calibration: Flow coefficient manually tuned for solid top layer quality
- Test model: Standard 30 × 30 × 30 mm calibration cube (solid)
This setup represents a common but well-prepared scenario, using factory presets with only minor flow tuning — similar to what many experienced users might apply in practice.
4.2 Measurement Method
After printing multiple cubes, the outer XY dimensions of each cube were measured using a calibrated digital caliper. Each dimension was recorded multiple times to account for tool variance and human error, then statistically analyzed to determine the standard deviation and overall tolerance range.
The Z dimension was also monitored, though less prone to variability in this specific test.
4.3 Results Summary
- XY tolerance (3σ bandwidth): 0.0846 mm
- 95% confidence interval: 0.0672 mm — 0.1218 mm
- Z-axis deviation: Similar range, though slightly less variance observed
- General result: Minor deviation from nominal (under 0.1 mm) and high consistency
These numbers suggest that under ideal but realistic conditions, the tested printer performs within a very respectable range for a desktop FDM machine. A deviation below ±0.1 mm is typically acceptable for most large functional part and more than optimal for visual parts.
4.4 Important Notes and Limitations
While these results are useful, they come with a few important caveats:
- Machine-specific: These results apply to this particular printer, in this specific condition. A brand-new machine or one with different wear or maintenance may behave differently.
- Material-specific: Almost PLA’s, and does from Bambu Lab have relatively stable properties, but different filaments — even from the same brand — can behave differently due to pigmentation, moisture, or formulation differences, an example is the black PLA that has more potencial for significant deviations (see point 3.2 for this)
- Setup-specific: A small change in nozzle temperature, cooling, bed adhesion method, or part orientation could shift the results outside this range.
This case study demonstrates that with proper tuning and environmental control, consumer-level printers like the Bambu Lab P-Series are capable of producing parts with impressive dimensional consistency. However, these values should be taken as a reference point, not as a universal benchmark.
The goal isn’t to chase perfect precision, but to understand what your own machine is capable of — and to design accordingly. That means using test prints to establish your own tolerances and adjusting designs and settings based on your real-world setup, not just manufacturer specs or community anecdotes.
5. Tolerance vs. Deviation: What Are We Really Measuring?
When we talk about dimensional accuracy, it’s easy to fall into the trap of treating it as a single, absolute value. But in practice, accuracy is a nuanced concept that depends on what exactly you’re measuring — and why.
Are we trying to understand how much a part deviates from its ideal size? Or are we concerned with how consistently it reproduces that size across multiple prints? This distinction is crucial, especially when designing functional parts or collaborating with others who may be using different printers or materials. This is an especially important aspect in some industries, such as the automotive industry — but also in 3D printing, especially when you’re creating parts for a large audience
5.1 Absolute Deviation from the Model
This is the most intuitive definition: the difference between the printed part and the original model. For example, if you print a 30 mm cube and it measures 29.85 mm, you have a deviation of -0.15 mm.
This type of deviation is important when:
- You need tight-fitting parts (e.g., press fits, mechanical assemblies)
- You’re working with toleranced engineering drawings
- You’re calibrating your printer or tuning your slicer
However, a small, consistent deviation can often be corrected through design compensation. For example, if your printer consistently undersizes outer dimensions by 0.15 mm, you can slightly scale your model or adjust your flow settings.
5.2 Tolerance Repeatability
This refers to how much variation there is between multiple prints of the same model, or even across different areas of the same print. In other words, how consistent is the machine, regardless of whether it’s slightly over or under the target dimension?
Let’s say you print five 30 mm cubes:
- Machine A: 29.84, 29.85, 29.83, 29.86, 29.85 → Very consistent, even if slightly off
- Machine B: 29.60, 29.90, 30.10, 29.75, 29.95 → Less predictable, harder to compensate
In this context, repeatability is often more valuable than precision alone. A predictable machine lets you design with confidence and build in the right tolerances.
5.3 Why the Distinction Matters
Many users chase after perfect absolute accuracy, adjusting settings endlessly in search of a 100% match with the model. But if the printer’s performance is not consistent, those adjustments may only work for one print or one geometry, and at a specific moment.
That’s why in design for FDM/FFF printing, I often recommend focusing on tolerant, adaptable designs that can absorb small deviations across machines and materials.
This includes:
- Larger clearances for sliding parts, or snap-fits (that I typically don’t use — due to mechanical issues)
- Rounded edges and chamfers to reduce the effect of over-extrusion
- Relief cuts os special geometry in holes to mitigate shrinkage
- Avoiding “perfect” dimensions for internal geometries (e.g., don’t design a 5.00 mm hole expecting a perfect fit)
5.4 Different Printers, Same Model — Different Results
Let’s consider a scenario where two users print the same model on two identical printer models. One uses PLA-CF, the other uses basic PETG. Even if both printers are well-tuned, differences in shrinkage rate, die swell, and cooling behavior could lead to functionally different results.
This is why standardized test models often produce varied outcomes in the community. The slicer settings, material batch, and environmental conditions can all cause dimensional shifts — even when using the exact same STL file.
To understand and manage dimensional accuracy, we must ask two separate but related questions:
- How close is the printed part to the model’s dimensions?
- How consistent are the results across prints and conditions?
Answering both helps define whether a deviation is just a minor offset — or a sign of unreliable performance. And in many cases, designing with realistic tolerances and variability in mind is more effective than chasing precision for its own sake.
6. Practical Ways to Improve Dimensional Accuracy
Achieving reliable dimensional accuracy in 3D printing doesn’t require an expensive industrial machine or endless trial and error. In fact, with a good understanding of your hardware, materials, and slicer, you can significantly improve print precision — even on a consumer-grade FDM printer.
Here are some practical strategies to help reduce deviations and improve the match between your printed parts and their original CAD dimensions.
6.1 Tune Speed and Acceleration
Slower is often better when it comes to accuracy. High speeds and rapid acceleration can introduce unwanted vibrations and inertia, especially during sharp direction changes or small feature transitions. These mechanical effects can cause slight overshoots, ringing, or positional drift.
Lowering print speed and acceleration can help:
- Improve motion precision
- Reduce ghosting and ringing
- Minimize overshooting in corners or seams
Tip: Try reducing speed by 20–30% and lowering acceleration to around 500–1000 mm/s² for dimensional calibration prints, normal print acceleration on a Bambu Lab P1 for PLA is around 6000 mm/s², that is at least six times more!
6.2 Adjust Printing Temperature
Extrusion temperature directly affects the viscosity of molten filament, which in turn influences flow behavior and shrinkage. Lower temperatures (within the safe range for your material) can help reduce over-extrusion, stringing, and uncontrolled flow.
However, dropping the temperature too much can cause under-extrusion or poor layer bonding. It’s a balance.
Tip: After slowing down your print, you can usually lower the nozzle temperature by 5–10°C while maintaining good extrusion quality.
6.3 Use a Smaller Nozzle
A smaller nozzle (e.g., 0,2 mm) provides finer control over extrusion and can improve the accuracy of small features like holes, embossed text, or thin walls. While it increases print time, it reduces over-deposition in tight areas and allows better resolution.
This can be particularly helpful when:
- Printing small functional parts
- Working with parts that require post-assembly fit
- Creating detailed prototypes with complex geometry
Keep in mind: Small nozzles require more precise tuning of flow and retraction settings and may struggle with large or heavily infilled parts, and it will take much more time to print your part.
6.4 Select Materials with Low Shrinkage
Some materials are inherently more dimensionally stable than others. PLA, for example, has relatively low shrinkage and good layer adhesion. But for even greater stability, fiber-reinforced materials like PLA-CF or PPA-CF offer significantly reduced warping and shrinkage thanks to the stiffness of the embedded short fibers.
These materials are ideal when maximum accuracy is critical — especially in mechanical or engineering parts.
Important: Fiber-filled materials are abrasive and require hardened steel or ruby nozzles. They may also demand higher printing temperatures and more powerful part cooling.
Please do not use 0.2mm nozzles for fibre-added materials, — it will clog, and since it’s not made of hardened metal, its diameter won’t stay at 0.2 mm for long. In these cases, the ideal choice is really a 0.6mm nozzle that in all Bambu Lab printer is made of hardened metal.
6.5 Dry Your Filament
I’ve reached the limite for the post … This go on referring to the new H2D and a bit more…
Quite refreshing to see detailed facts versus opinionated observations!