TPU 68D and TPU 90A are both flexible, but they are not remotely interchangeable. One is easier to print and stiffer in service, while the other is softer, stretchier, and much better at grip, comfort, and sealing. TPU 68D, sold by Bambu Lab as TPU for AMS, is faster, cleaner, and more manageable in an automated system; TPU 90A behaves more like a rubber-like material in real use and is better for grip, seals, and comfort. The right answer depends on the job, not just the material label.
Why these two TPUs behave differently
Flexible filaments open up a whole new world of 3D printing possibilities, but not all TPUs behave the same way. The main difference comes down to hardness and elasticity.
TPU 68D is firmer and more rigid while still being flexible. It is a great choice when you need a part to stay stable and print cleanly without fighting the machine.
TPU 90A is softer and more elastic. That makes it excellent for grip, wearables, and sealing parts, but it is also more demanding to print reliably and less friendly to an AMS workflow.
How I tested them
Every part shown here was printed using default Bambu Studio settings with a 0.4 mm nozzle, 15% infill, and two wall loops. The goal was to evaluate how these filaments perform in real, everyday applications without exotic tuning or heavily customized profiles.
This is the kind of test that matters because most people are not printing TPU in a lab. They are printing a useful part and deciding whether it feels right, behaves well, and survives actual use.
Tank treads: TPU 90A wins the grip battle
To kick things off, I printed tank treads for a Cyberbrick Forklift kit. This is a perfect application for flexible material because the tread has to flex, grip, and survive repeated mechanical loads without binding the motors.
TPU 68D results
- Print time: 27 minutes
- Clean appearance with minor stringing
- Too stiff for the application
- Bound up the motors
- Poor grip
- Tread snapped when removed from the forklift
TPU 90A results
- Print time: 49 minutes
- More flexible and grippy
- Allowed smooth control of the forklift
- No motor binding
- Much better traction
Strength test
I hung weight from the treads to compare durability.
- TPU 68D: broke with less than 9 pounds
- TPU 90A: withstood 27–36 pounds
That is a large gap, and it is exactly the kind of result you want from a material that is supposed to absorb shock and tolerate repeated flexing.
Verdict for the forklift
TPU 90A clearly came out on top here. It was softer, had more traction, and handled the mechanical load much more gracefully. If the part needs to grip, flex, and survive force, the softer filament is the better choice.
Custom furniture pads: function matters more than softness alone
For a more practical test, I designed custom chair and recliner feet to replace the common stick-on pads that never fit quite right. These pads need to hold the furniture in place without being too spongy or leaving a mark on the floor.
TPU 68D furniture pads
- Prints were nearly flawless
- Firm feel
- Slid easily on hardwood floors
- Held up well under weight
- Great fit and clean print quality
TPU 90A furniture pads
- Softer and grippier
- Better at keeping furniture in place
- Slight deformation under heavy weight
Verdict for furniture pads
For furniture that needs to stay put, TPU 90A is the better option. For furniture that should glide slightly, especially on hardwood floors, TPU 68D is a cleaner, more stable choice.
For TPU 90A pads, increasing the infill to about 30% and adding extra wall loops can help improve rigidity without giving up the grippy feel.
Wearables: 3D printed belts
Could TPU printing replace a traditional leather belt? I decided to test that idea directly.
A belt needs to be comfortable, quiet, flexible enough to move with the body, and durable enough to survive repeated load. That is a different problem from a rigid mechanical part.
TPU 68D belt results
- Print time: 8 hours 22 minutes
- Felt stiff, like a brand-new belt
- Noisy when moving in a seat
- Very durable
TPU 90A belt results
- Print time: 20 hours 30 minutes
- Much more comfortable and flexible
- Quiet to wear
- Minor stringing, but the result still looked great
Strength test
I hung chairs from the belts to compare their real-world load capacity.
- TPU 68D: broke at about 40 pounds
- TPU 90A: held up to 50 pounds
Because both were strong, the deciding factor ended up being how they felt in everyday use. TPU 90A stretched more and was noticeably more comfortable.
Verdict for belts
I have personally worn a TPU 90A belt for more than a year with no issues. It is the winner here, even though TPU 68D would probably still work for many people.
Coaster feet: TPU 68D is easier to print, TPU 90A is grippier
The coaster test was a smaller but still useful application. The main question here was how each material handled the first layer and the surface contact behavior.
The most obvious difference with TPU 90A is that it tends to ooze more from the nozzle, which can make first layers trickier. I printed these coasters with the logo side down and paid close attention to the first layer to avoid contaminating the visible top surface.
Observations
- TPU 68D prints cleaner and more predictably.
- TPU 68D has less tendency to ooze from the nozzle.
- TPU 90A offers more grip on the table.
- TPU 90A can contaminate the top surface if the first layer is not watched closely.
Verdict for coaster feet
Both materials work well here. For single-nozzle printers, especially with an AMS, TPU 68D is the easier and more practical option because it stays easier to manage and can remain loaded in the AMS while other materials are printed around it. TPU 90A is grippier, but it is a little more demanding to use in a multi-material setup.
Water-resistant project box: stiffness and seal performance are different jobs
A more advanced test was a small electronics enclosure that used TPU in three different roles:
- TPU gasket seal
- TPU sidewalls for cable pass-through
- TPU screw mounts
This test was especially useful because it separated properties that often get lumped together under the word “flexible.”
Seal performance
- TPU 68D gasket allowed water intrusion.
- TPU 90A gasket kept water out.
That difference is meaningful. A TPU gasket only works if it can compress into a reliable seal. TPU 90A did that much better.
Drillability
- TPU 68D drilled cleaner holes and produced a better seal around wires.
- TPU 90A deformed when drilling and sometimes tore, making the seal less reliable.
Verdict for the project box
Use both if you can:
- TPU 90A for the top gasket where sealing matters most
- TPU 68D for sidewalls where clean drilling and easier manufacturing matter more
If you only have a single-nozzle printer, you may have to use TPU 68D, but you should not expect it to be fully waterproof in the same way as TPU 90A.
Which TPU should you choose?
This is the short version.
Pick TPU 68D when you want:
- Cleaner, faster prints
- Better compatibility with an AMS workflow
- A firmer, more stable flexible part
- A material that is easier to manage on a single-nozzle printer
Pick TPU 90A when you want:
- More grip
- Better flexibility and stretch
- Stronger resistance to tearing and flex failure
- Better sealing performance and comfort in wearable parts
Final verdict
TPU 68D is the better choice when you want a flexible material that is easier to print and easier to live with on a regular printer. It is more print-friendly, more predictable, and more compatible with the automated filament system.
TPU 90A is the better choice when the part needs to grip, stretch, seal, or feel more comfortable in contact with the user or the environment. It was stronger in the load tests, more flexible in the wearable applications, and more reliable as a gasket material.
Both filaments have a place. The right TPU is the one that matches the job, and in practice, the softer one usually wins when the part has to do real mechanical work.