# Drying Filament: Why We Do It and When You Should Care

> Drying filament explained: what moisture does inside a 220 °C nozzle, which materials care most, the symptoms, and why drying filament in an oven is risky.

Source: https://x3dstudios.com/blog/why-we-dry-filament  ·  2026-11-07  ·  X3D Studios  ·  9 min read

Drying filament is a policy here because printing plastic is mildly hygroscopic and water is destructive at 220 °C. Every common filament pulls moisture out of the air over time, and when that strand reaches the nozzle the water flashes to steam inside the melt, blowing the extruded bead apart from within. What comes out is stringing, a pitted surface, layers that never fully weld, and in PLA a strand brittle enough to snap inside the AMS. You should care when the material is thirsty, when the spool has been sitting open, and when the print is long enough that failing at hour ten actually costs you something.

## Why Drying Filament Matters: What Water Does in a Hotend

Water boils at 100 °C. A PLA nozzle runs around 220 °C, and PETG and ABS run hotter still. By the time a strand of filament reaches the melt zone, any water it carried is not water any more — it is steam, expanding roughly a thousand-fold inside a viscous fluid that has nowhere convenient to vent. The bubbles travel with the melt and burst as the bead leaves the nozzle. Each burst leaves a small crater, sprays a whisker of plastic across the part, or simply means that some of what the machine thought it was extruding was gas. /docs/materials compares all five materials, their rates and what each is for.

There is a slower, meaner effect underneath that one. PLA and PETG are polyesters, and at melt temperature water attacks the polymer chains directly, cutting long chains into shorter ones. Shorter chains mean a weaker part, and unlike the steam, that damage does not reverse when the spool dries out. A badly waterlogged spool that has already been through a hot nozzle prints acceptably after drying but does not necessarily recover its full strength, which is why the answer is to keep filament dry rather than to rescue it later.

![Flow diagram showing humid air absorbed by a spool, water boiling at 100 degrees Celsius inside a 220 degree nozzle, steam bursting the bead, and the four resulting defects: stringing, pitted skin, weak layers and a snapped strand in the AMS](https://x3dstudios.com/blog-images/why-we-dry-filament/moisture-in-the-hotend.svg)

*Water is steam long before the plastic around it melts. The bubbles are what you see on the part.*

## The Symptoms, in the Order You Notice Them

1. Sound first. A wet spool makes the hotend tick or pop audibly, a few times a minute. A dry spool is silent.
2. Whiskers and stringing between features that printed cleanly last week with the same file and the same settings.
3. A surface that reads rough and slightly matte where it should be smooth, covered in craters too small to see one at a time.
4. Inconsistent extrusion: thin patches in solid infill, gaps in the top layer, a first layer that looks starved in places.
5. Parts that split along a layer line under light hand pressure, because the beads never welded properly through the steam.
6. In PLA specifically, a strand that snaps where it bends. In an AMS that is the hub or the buffer, and the job stops there.

That last one is the failure mode a print farm sees most, because an AMS puts a permanent bend into every strand at two points, and brittle filament breaks where it is bent. The tell is a multi-colour job that runs perfectly on three trays and dies the moment it reaches the fourth — the one holding the oldest spool. /blog/ams-errors-decoded walks through what each AMS error code actually means and how to tell a moisture fault from a mechanical one.

> **INFO** — Stringing has three common causes: moisture, retraction settings, and too high a nozzle temperature. Moisture is the one that also makes noise. If the hotend is ticking, stop tuning retraction and go dry the spool.

## Which Materials Actually Care

Every filament on our menu absorbs some water, but they do it at very different speeds and they need very different dryer settings. Two independent things are going on here, and mixing them up is the usual mistake: how fast a material drinks is a property of the polymer, while how hot you have to run the dryer is a property of its softening point. ABS is slow to absorb and needs a hot dryer; PLA is quicker to absorb and must be dried cool.

| Material | How thirsty | Typical dryer temp | Typical time | Our rate |
| --- | --- | --- | --- | --- |
| PLA | Moderate | 45–55 °C | 6–8 h | $0.12/g |
| PETG | Moderate to high | 60–65 °C | 6–8 h | $0.12/g |
| ABS | Low to moderate | 70–80 °C | 6–8 h | $0.15/g |
| TPU | High | 60–70 °C | 8–12 h | $0.20/g |
| ASA | Low to moderate | 70–80 °C | 6–8 h | $0.15/g |

![Bar chart of typical drying temperatures by material: PLA 55, PETG 65, TPU 70, ABS 80 and ASA 80 degrees Celsius, colour coded by how quickly each material absorbs moisture](https://x3dstudios.com/blog-images/why-we-dry-filament/drying-temps-by-material.svg)

*Bar length is dryer temperature; colour is thirst. TPU is the shortest hot bar and the thirstiest material on the list.*

Carbon-fibre-filled versions of any of these belong a full step to the right. The chopped fibre gives water far more internal surface to cling to, so PETG-CF or PLA-CF picks up moisture noticeably faster than the plain version of the same polymer and shows it sooner as a fuzzy surface. /blog/carbon-fiber-filaments-explained covers what the fibre buys you and what it costs in handling. If you are still choosing a material at all, /blog/pla-vs-petg-vs-abs-vs-asa-vs-tpu compares them on everything else.

> **WARNING** — Never dry a spool above the temperature its own manufacturer prints on the label. PLA's glass transition sits around 50–60 °C, so a dryer set for ABS will soften a PLA spool, fuse the windings together, and hand you back a spool that will not feed at all.

## When You Should Care, and When You Should Not

Drying is a habit worth having selectively rather than religiously. The cases that reliably justify it:

- Any TPU, ever. It is the thirstiest thing most people own and it is the least forgiving to print wet.
- Any spool that has been open on a shelf for more than a month, in any material.
- Any carbon-filled filament, regardless of how recently the bag was opened.
- Any print long enough that a failure hurts. A twelve-hour job is twelve hours of machine time you do not get back.
- Any structural part, where the layer bonding you are quietly giving up is the whole point of the part.
- Any multi-colour job, because a snapped strand at tray three ends the plate for all four colours.

And the cases where it is honestly not worth the eight hours: a vacuum-sealed spool you opened this week, a thirty-minute PLA test print, or any decorative part where a whisker you will trim with a knife is not a defect worth preventing. The decision is really about the value of what is at risk, not about a rule.

Geography matters more than people expect. Our farm is in Austin, Texas, where the outdoor dew point sits high for a long stretch of the year, so a spool left on an open shelf here reaches equilibrium with the room far faster than the same spool would in a dry climate. If you print somewhere humid, treat an unsealed spool as a spool that is already absorbing water, because it is.

## How the Farm Handles It

Spools live in sealed containers with desiccant, not on a shelf, and they go into a drying cycle before long or thirsty jobs rather than after something has already gone wrong. Each machine feeds from a 4-tray AMS with its own desiccant inside, and the lid stays shut between jobs — that desiccant is a consumable that gets changed, not a decoration that ships with the unit.

The rest is catching the failure early when one slips through. Every printer carries a live camera stream, and a moisture failure has a long visual runway: the surface goes fuzzy and pitted for many layers before anything actually detaches from the plate. /blog/how-a-print-farm-detects-failures covers what telemetry sees, what a camera sees, and why you need both. Sliced files that customers upload also pass G-code safety validation before they reach a machine, which is a different problem but the same principle: cheaper to check first than to clean up after.

None of that changes your invoice, which is the point. A 62 g PLA part is $7.44 at the $0.12 per gram all-in rate whether it printed the first time or the third, with a $1.00 minimum per part and $7 standard US shipping, or $15 express. A moisture failure costs us grams and a printer-hour, and that is exactly why drying is a policy here instead of a suggestion. The live rates for every material are on /pricing.

## Drying at Home Without Buying Much

A purpose-built filament dryer is the easy answer and the one worth the money if you print weekly. The common cheap substitute is a food dehydrator with the stacking trays removed, which holds a spool and runs at exactly the low, stable temperatures this job needs. A kitchen oven is the option to be careful with: most household ovens swing by tens of degrees at settings this low, and PLA's softening point sits inside that error band.

Storage matters more than drying does, though, because drying a spool and then leaving it on the bench undoes the work within days. A sealed bin with a handful of colour-indicating silica gel is cheap, keeps a dry spool dry indefinitely, and the beads regenerate in the same dryer you used for the filament. The whole discipline is: dry once, then never let it get wet again.

---

## FAQ

### How can I tell if my filament is wet?

Listen to the nozzle. A wet spool ticks and pops as steam escapes, a few times a minute, and a dry one does not. The visual confirmation is a rough, slightly pitted surface plus whiskers between features on a file that used to print clean. The definitive test is to print the same small part before and after an eight-hour drying cycle; if the second one is smooth and quiet, the spool was the problem.

### Is drying filament necessary for PLA, or only for nylon?

PLA needs it less often than nylon or TPU, but it absolutely absorbs water, and the symptom it produces is different enough that people miss it. Wet PLA is brittle rather than just stringy: the strand snaps where it bends, which in an AMS means it breaks at the hub or the buffer and stops the job. A spool that has spent a humid summer open on a shelf deserves a cycle at 45–55 °C before it goes near a long print.

### Can you dry filament in an oven?

You can, but it is the riskiest of the options. Household ovens control poorly at the bottom of their range and can overshoot a set point by a wide margin, and PLA starts softening in the 50–60 °C range, so an overshoot fuses the windings into a solid disc. If the oven is what you have, use an independent thermometer inside it, stay well under the spool's labelled limit, and check on it. A dehydrator is a far safer twenty-five dollars.

### Do I need to dry filament before ordering a print from you?

No. You send a file, not a spool. Upload an STL or 3MF file (a sliced .gcode.3mf counts) at /print and the filament, its storage, its drying and any reprint caused by a bad one are all our problem. That is genuinely most of what a print farm is selling: not the printer, but everything around it that has to be right before the printer matters.

[Print from spools someone else keeps dry](https://x3dstudios.com/print)