DEV Community: Tristan Zhou

Why Do 3D Prints Warp? Causes and Solutions

Tristan Zhou — Mon, 26 Oct 2026 07:58:29 +0000

You've probably seen this happen: a 3D print starts off perfectly fine, but a few hours later, one corner begins lifting off the build plate. Sometimes the print finishes with a curved bottom. Other times, the nozzle catches the raised edge and ruins everything. It's tempting to blame poor bed adhesion, but warping usually starts with something more fundamental: how plastic behaves as it cools. Understanding this makes it easier to solve the problem instead of constantly adjusting printer settings.

What Is 3D Print Warping?

Warping happens when a printed part bends or curls away from its intended shape, usually near the bottom corners. In FDM 3D printing, molten filament is deposited layer by layer. As it cools, the plastic contracts. If different areas cool at different rates, internal stresses develop and may pull the part away from the build plate. Large, flat components and materials such as ABS and ASA are particularly vulnerable, although PLA can warp too.

Why Do 3D Prints Warp?

Several factors contribute to warping, and they often occur together.

1. Uneven Cooling and Shrinkage

Thermal contraction is one of the main causes. Imagine printing a large ABS enclosure. The bottom stays relatively warm against the heated bed, while the upper layers cool in the surrounding air. As the plastic contracts, it pulls against neighboring layers, creating stress that can lift the corners. Drafts from windows, air conditioners, or nearby fans can make this worse. Prusa also explains how temperature differences during printing contribute to shrinkage and warping, particularly with high-temperature materials.

2. Poor Bed Adhesion

Good adhesion helps the print resist shrinkage forces. However, oil, dust, an incorrect Z-offset, or an uneven build plate can prevent the first layer from sticking properly. If the nozzle is too far away, filament may barely contact the surface. If it's too close, extrusion becomes restricted. Both situations can cause trouble.

3. Material Properties

Not all plastics shrink or respond to temperature changes in the same way. PLA generally warps less under typical desktop printing conditions, while ABS, ASA, nylon, and polycarbonate often require more careful temperature control. For example, when comparing ABS and PETG, PETG is generally easier to print with less warping, while ABS offers better heat resistance but benefits from a controlled printing environment. Additives and fillers can also change material behavior, so two filaments with the same material name may print differently.

4. Part Geometry

A small bracket might print successfully while a long, flat enclosure made from the same filament warps badly. Large dimensions, sharp corners, and uneven wall thicknesses can influence how thermal stresses develop. Sometimes the solution isn't changing the printer settings but reconsidering the part's design.

How to Prevent 3D Print Warping: 7 Practical Steps

There's no universal setting that prevents every warped print. I recommend checking the simplest problems first before making bigger changes.

Step 1: Clean and Check the Build Plate

Start with a clean build surface. Fingerprints, dust, and leftover adhesive can affect first-layer bonding. Follow the manufacturer's cleaning instructions, then check bed leveling and Z-offset. A good first layer should have continuous, evenly flattened extrusion lines without being excessively squashed. If the first layer looks inconsistent, fix it before changing anything else.

Step 2: Adjust the Bed Temperature

A heated bed slows cooling near the bottom of the part and helps maintain adhesion. Typical starting temperatures include:

Material	Bed Temperature
PLA	50–65°C
PETG	70–85°C
ABS	90–110°C
ASA	90–110°C
Nylon (PA)	70–110°C

These are general ranges, not fixed requirements. Always follow your filament manufacturer's recommendations. Increasing the temperature slightly may improve adhesion, but excessive heat can soften the lower layers and cause dimensional problems.

Step 3: Improve First-Layer Settings

A reliable first layer makes everything else easier. Try reducing first-layer speed, checking extrusion width, and adjusting the Z-offset. Slower printing gives the material more time to contact the surface. The goal is consistent adhesion, not squeezing as much plastic as possible onto the bed.

Step 4: Control Cooling and Airflow

Cooling requirements depend on the filament. PLA usually benefits from part cooling, especially around small features and overhangs. ABS and ASA are more sensitive to rapid temperature changes. An enclosure can reduce drafts and maintain a more stable environment, making it especially useful for larger prints. However, completely disabling cooling isn't always ideal. Follow material-specific guidance and ensure appropriate ventilation and printer temperature management.

Step 5: Add a Brim

A brim adds extra first-layer material around the model, increasing contact area with the build plate. It's particularly helpful for large flat prints, sharp corners, and parts with small contact areas. A width of around 5–10 mm is a reasonable starting experiment. UltiMaker also recommends using brims or rafts to improve build plate adhesion, although rafts use more material and can affect the bottom surface finish. Remember, a brim resists lifting but doesn't eliminate internal stress.

Step 6: Reconsider Print Orientation and Design

If temperature and adhesion adjustments aren't enough, examine the model itself. Could you change its orientation, split a long part into smaller pieces, or replace sharp corners with rounded ones? More uniform wall thickness may also help certain designs. Just remember that changing print orientation affects layer strength, so the easiest orientation to print isn't necessarily the strongest for the final application.

Step 7: Check Filament Moisture

Moisture-sensitive materials such as nylon can absorb water from the air. Wet filament may cause bubbles, inconsistent extrusion, and poor surface quality. These problems aren't the same as thermal warping, but they can complicate troubleshooting. If you suspect moisture, dry the filament according to the manufacturer's instructions rather than assuming drying alone will solve warping.

Which Materials Warp the Most?

Here's a general comparison of common FDM materials:

Material	Warping Tendency
PLA	Low
PETG	Low to moderate
ABS	High
ASA	High
Nylon (PA)	Moderate to high
Polycarbonate (PC)	High

These are general tendencies. Actual results depend on formulation, print size, printer design, and environmental conditions. Some fiber-filled filaments experience less shrinkage than their unfilled equivalents, but that doesn't mean every reinforced filament is easier to print.

Can You Fix an Already Warped Print?

It depends on the severity. If a corner lifts slightly and dimensional accuracy isn't important, the print might still finish successfully. But once the bottom has noticeably curled away from the bed, it usually won't return to its original shape during printing. For functional parts, restarting is often the safer choice. Avoid pushing a lifted corner back into place while the machine is running. Instead, examine the failed print and identify where the problem started. When troubleshooting, change one major setting at a time so you can tell what actually helps.

Frequently Asked Questions

Why do my 3D print corners keep lifting?

Usually, thermal contraction creates more force than the bed adhesion can resist. Check your first layer, bed temperature, airflow, and whether a brim could help.

Can PLA warp on a heated bed?

Yes. PLA is generally less prone to warping than ABS, but poor adhesion, drafts, or large flat geometries can still cause problems.

Is an enclosure necessary for ABS?

Not always for small prints, but an enclosure is strongly recommended for larger ABS parts and more consistent results.

Final Thoughts

3D print warping isn't simply a problem with glue or bed temperature. It's the result of material shrinkage, uneven cooling, adhesion, and part geometry working together. Start by checking the first layer, then look at temperature stability, airflow, and material behavior. If those changes don't work, reconsider the model design. Understanding why plastic shrinks is far more useful than memorizing one set of slicer settings.

A Nebulizer Mouthpiece Looks Simple. It Isn't.

Tristan Zhou — Wed, 23 Sep 2026 06:47:11 +0000

Some products are so ordinary that we rarely stop to think about how they were designed.

A nebulizer mouthpiece is a good example.

At first glance, it is just a small piece of plastic attached to a breathing treatment device. There are no screens, electronics, buttons, or moving mechanisms.

It looks simple.

But simple-looking products are often where small design decisions matter the most.

Once you start looking closely at a nebulizer mouthpiece, there is more going on than you might expect.

The shape has to feel natural

The first thing you notice is the shape.

A mouthpiece has to sit comfortably between the lips without feeling too large, too sharp, or awkward to hold in place.

That sounds obvious, but it immediately creates design constraints.

The edges need to be smooth.

The transition between surfaces shouldn't feel abrupt.

The opening has to work with the airflow path while still being comfortable for the person using it.

Even a small change in width, thickness, or curvature can change how the product feels.

This is one of those cases where a few millimeters can make a surprisingly big difference.

Smooth doesn't mean featureless

A mouthpiece usually looks very clean from the outside.

That doesn't mean the geometry is simple.

Turn one over or look inside and you may find:

internal transitions
thin walls
ribs
connection features
sealing surfaces
small changes in wall thickness

Many of these details are almost invisible during normal use.

But they help the mouthpiece connect to the rest of the device, maintain its shape, and perform consistently.

This is something I find interesting about plastic product design in general.

The outside often looks simple because most of the engineering has been hidden.

Material choice changes more than appearance

When we think about plastic, it is easy to focus on whether something is clear, white, soft, or rigid.

But the material affects much more than appearance.

A mouthpiece may be handled repeatedly. Depending on the product, it may also need to tolerate cleaning, disinfecting, or other conditions defined by the device manufacturer.

The material therefore has to make sense for how the part will actually be used.

This is one reason medical doesn't describe a single type of plastic.

Two plastic components used in the same piece of medical equipment may have completely different requirements.

One might simply form part of the outer housing.

Another might come into direct contact with the user.

Their shapes may look equally simple, but their design considerations are not necessarily the same.

Manufacturing leaves clues

Plastic products often tell you something about how they were made.

If you look carefully at everyday molded parts, you can sometimes find a thin line where two mold surfaces met.

There may be small circular marks in less visible areas.

The wall thickness may be more uniform than you would expect.

Corners are often rounded instead of perfectly sharp.

These details aren't random.

They are connected to how molten plastic fills a mold, cools, shrinks, and eventually comes out as a finished part.

The same basic manufacturing process is used for everything from storage boxes and appliance housings to much smaller molded components.

But with medical injection molding, the manufacturing decisions also have to reflect how the finished component will actually be used.

That's what makes an apparently simple part more interesting.

Cleaning can influence design

Another detail that's easy to overlook is cleaning.

Whenever a product needs to be cleaned repeatedly, small geometric decisions become more noticeable.

Deep corners can be harder to reach.

Rough or complicated surfaces can hold residue more easily than simple ones.

Connections between components may create small spaces that need attention.

So a smooth shape isn't necessarily only about appearance or comfort.

It can also make the product easier to handle and maintain.

This is a good example of how one design feature can solve several problems at the same time.

A rounded surface can feel better, be easier to mold, and be easier to clean.

Good product design often works like that.

The connection matters too

The mouthpiece doesn't work by itself.

It has to connect to another component.

That connection might look like nothing more than one plastic tube sliding over another, but it still needs to feel right.

Too loose and the connection may not feel secure.

Too tight and assembly or removal becomes frustrating.

The geometry also has to remain consistent from part to part.

This is something that's easy to forget when looking at mass-produced products.

We usually judge one object.

Manufacturing has to produce thousands or millions of objects that behave the same way.

Making one part that fits is relatively easy.

Making the same fit repeatedly is the real challenge.

Simple products are rarely actually simple

The more everyday products I look at, the less I believe in truly "simple" parts.

A nebulizer mouthpiece doesn't have much visual complexity.

That's precisely what makes it interesting.

The shape has to feel comfortable.

The internal geometry has to support its function.

The material has to match its intended use.

The part has to connect correctly to the rest of the device.

And the same design has to be produced consistently over and over again.

None of these details are especially dramatic.

Most users will never think about them.

That's usually a sign that the design is doing its job.

A small plastic mouthpiece may look like almost nothing.

But somebody still had to think about every surface.

Humanoid Robots Look Ready for the Future. So Why Haven’t They Truly Become Widespread?

Tristan Zhou — Mon, 22 Jun 2026 02:41:21 +0000

When scrolling through videos, have you ever been amazed by humanoid robots? They can walk, carry objects, dance, talk to people, and even appear in factory or home scenarios. But here is the real question: if they are truly so advanced, why don’t we see them widely used in factories, warehouses, hospitals, shopping malls, or ordinary homes?
Humanoid robots are still blocked by several practical problems before they can become truly widespread: hardware reliability and cost, energy and battery life, intelligent decision-making and scenario adaptability, unclear replacement value, and safety and interaction risks.

1. Reliability and Cost

Many humanoid robot videos look impressive, but succeeding once in a video and working reliably every day in real-world environments are not the same thing.
Factories, warehouses, hospitals, and homes do not need a robot that can occasionally complete a task. They need one that can:
· run for long periods
· maintain a low failure rate
· be easy to maintain
· keep costs under control
· avoid serious consequences when something goes wrong
Humanoid robots contain numerous joints, transmission structures, sliding components, housings, brackets, and internal connectors. Wear, looseness, or fatigue in any critical part may affect the operation of the entire machine.
Improving reliability usually means using better motors, more precise components, and more durable materials. For example, in recent years, PEEK has been increasingly used in the manufacturing of humanoid robots. However, high-performance materials, precision machining, high-quality actuators, and complex assembly all drive up costs.

2. Energy and Battery Life Bottlenecks

Humanoid robots are expected to do too many things: walk, maintain balance, recognize their surroundings, process speech, drive multiple joints, and calculate movements in real time.
All of these functions require power. The problem is that battery choices create an obvious trade-off: if the battery is too small, battery life is insufficient; if the battery is too large, the entire robot becomes heavier; the heavier the robot, the greater the load on its joints; and the greater the load, the more energy it consumes.
If a robot has to recharge after working for only a short time, or frequently stops for maintenance, it will struggle to become a real productivity tool. It may attract media attention, but it will be hard to convince factories or households to use it over the long term.
For humanoid robots to become widespread in the future, they must find a better balance between batteries, actuators, lightweight structures, and low-power computing.

3. Insufficient Intelligent Decision-Making and Scenario Adaptability

Many people assume that as long as AI is powerful enough, humanoid robots will be able to handle all kinds of tasks like humans. That view is too optimistic.
The real world is not a demonstration stage. The floor may be uneven, object positions may change, lighting may be unstable, people may suddenly pass by, and tasks may shift unexpectedly.
A robot picking up a cup in a video does not mean it can handle every situation in a real kitchen. Cups may vary in size, their positions may change, the tabletop may be wet, and there may be people or pets nearby. More importantly, the cup may contain hot water, and the robot must determine whether its action is safe.
These details are ordinary for humans, but they are difficult for robots. Many demonstrations run smoothly because the environment has already been carefully prepared in advance.

4. Unclear Replacement Value

For humanoid robots to become widespread, they must answer a very commercial question: who exactly are they replacing, and is that replacement worth it?
In factories, many tasks already have mature automation solutions. Robotic arms, conveyor lines, AGVs, and dedicated fixtures may be cheaper, more stable, and more efficient than humanoid robots. If a fixed-position robotic arm can already complete the job, why would a company buy a more expensive, more complex, and harder-to-maintain humanoid robot?
The same applies to homes. Robot vacuums solve floor cleaning, dishwashers solve dishwashing, and smart speakers handle voice interaction. If a humanoid robot can only “do a bit of everything” while remaining expensive, failure-prone, and only moderately efficient, it will be hard for it to become a real necessity.
Humanoid robots cannot simply prove, “I can do it.” They must prove, “I am more cost-effective than the existing solution.” Until that point is proven, even the coolest demonstration is still just a demonstration.

5. Safety and Interaction Risks

If humanoid robots enter homes, hospitals, shopping malls, schools, or offices, they must coexist closely with people. The risks increase accordingly.
They need to avoid bumping into people, pinching them, falling over, misinterpreting commands, losing control while carrying objects, or making the wrong move in an emergency. Industrial robotic arms are usually separated by safety barriers, and their operating areas are clearly defined. Humanoid robots, by contrast, may operate in open environments, where the margin for error is much smaller.
Another easily overlooked issue is that humans may not be willing to trust them. When a robot looks increasingly human but moves unnaturally, it can actually make people uncomfortable. This is especially true in homes, healthcare, and elderly care settings, where users care not only about functionality but also about psychological acceptance.

Conclusion

So, can humanoid robots run reliably over the long term? Can they control costs? Can they handle complex environments? Can they be more cost-effective than existing solutions? Can they coexist safely with humans?
Until these questions are solved, humanoid robots will continue to attract attention, but it will still be difficult for them to truly spread across factories, warehouses, hospitals, shopping malls, and homes.
The future may belong to humanoid robots, but not to robots that can only perform. The ones that truly remain will be reliable, affordable, maintainable, and capable of creating real value.

Humanoid Robots Look Ready for the Future. So Why Haven’t They Truly Become Widespread?