Temperature Towers and Flow Calibration: Dialing In a New Filament

Every spool of filament ships with a temperature range printed on the side, and that range is almost useless for dialing in a print. It’s wide on purpose — it has to cover every printer, every altitude, every ambient condition. Your actual best temperature is one specific value inside that range, and the only way to find it is to test. The same goes for flow rate, which controls how much plastic actually comes out. This guide covers both, in the order to do them: temperature first, then flow.

Two calibrations, run once per spool, and the rest of your prints just work. It’s about forty-five minutes of setup that saves a week of “why does this filament print badly.”

Why Temperature Comes First

Temperature affects almost everything visible in a print: stringing, bridging, overhang quality, surface finish, and layer adhesion. Higher temperatures improve layer bonding but increase stringing and soften overhangs; lower temperatures sharpen overhangs and reduce stringing but eventually cause weak layers and under-extrusion. There’s a sweet spot, and it moves with the filament brand, color, and even batch.

You calibrate temperature before flow because flow rate is measured against extruded plastic that’s behaving correctly — and plastic at the wrong temperature doesn’t behave correctly.

Running a Temperature Tower

A temperature tower is a single print divided into bands, each printed at a different temperature. When it finishes, you have a physical record of how the filament performs across a range, and you pick the best-looking band.

Getting the Tower

Most modern slicers can generate a temperature tower for you. OrcaSlicer and Bambu Studio have it built into their calibration menus — the slicer inserts the temperature changes at the right layers automatically, so there’s no separate model to download. If your slicer doesn’t, a community post-processing script for PrusaSlicer or Cura does the same job by inserting temperature-change commands at set layer heights.

Set the range to span your filament’s published window:

• PLA: roughly 190-230C
• PETG: roughly 220-250C
• ABS/ASA: roughly 230-260C

Use a 5C step between bands.

Print Settings for the Tower

Keep everything except temperature constant. Use your normal speed, your normal cooling, and your current retraction. The point is to isolate temperature as the only variable — if you slow down or change cooling between bands, you can’t attribute the differences to temperature.

Reading the Tower

Examine each band under good light:

• Stringing: look for fine threads between the tower’s features. Less is better, and stringing usually gets worse toward the hotter bands.
• Bridging: flat horizontal spans should stay tight, not sag. Bridges tend to improve at lower temperatures because the plastic solidifies faster in air.
• Overhangs: angled sections should hold their shape without curling. Better at lower temperatures with good cooling.
• Surface finish: the walls should be smooth, without blobs or rough patches.
• Layer adhesion: if a band visibly delaminates while printing, it’s too cold.

You’re choosing the lowest temperature that still gives clean layer adhesion and a smooth surface — low enough to keep stringing and overhangs sharp, but not so low that layers separate. For PLA that’s commonly around 200-215C; for PETG, often around 230-240C. These are starting points, not answers — your spool decides.

Once you’ve found the best band, you can run a second short tower in 2C steps around that value if you want to be precise. The difference between 210C and 212C is sometimes visible in the surface.

Calibrating Flow Rate

With temperature set, move to flow. Flow rate (also called flow ratio or extrusion multiplier) is the software percentage that scales how much filament the printer pushes. Too high and you over-extrude — bulging surfaces, rough tops, dimensions running large. Too low and you under-extrude — gaps between lines, weak walls, dimensions running small.

A note on hierarchy: flow rate is a per-material software trim, not a fix for hardware. If your extruder’s steps-per-millimeter (e-steps) are wrong, every material extrudes wrong and you’ll be papering over it with flow. Calibrate e-steps once at the hardware level, then use flow per filament for the fine adjustment.

The Single-Wall Method

The cleanest flow test prints a hollow open box or cylinder with a single perimeter and no top, bottom, or infill, so the wall is exactly one nozzle-width of plastic.

1. Slice and print the single-wall test object.
2. Measure the wall thickness with calipers at several points and average.
3. Compare to the expected wall width (your line-width setting — for a 0.4mm nozzle, often 0.42-0.45mm).
4. New flow = current flow × (expected width ÷ measured width).

If your line width is set to 0.44mm and you measure 0.46mm, you’re slightly over-extruding, and you’d scale flow down a touch. Most slicers, including OrcaSlicer, have a built-in flow calibration that walks through this and writes the result straight back into the profile.

Verifying with a Calibration Cube

Print a 20mm calibration cube and measure all three dimensions with calipers. They should land within about 0.1-0.15mm of 20mm. If a dimension is consistently large, flow is high; consistently small, flow is low. Persistent dimensional error after flow calibration usually points back to e-steps or a mechanical issue, not flow.

The Per-Spool Routine

When you open a new spool, this is the whole protocol:

1. Name the slicer profile with the brand and color so you don’t lose the values.
2. Run a temperature tower (about 45 minutes) and pick the best band.
3. Run a single-wall flow test, or your slicer’s built-in flow calibration.
4. Print a 20mm cube and check dimensions.
5. Print your first real part and glance at it for stringing. Only retune retraction if it’s clearly excessive — see our stringing and retraction guide ↗.

Different batches of the same brand can drift several degrees in optimal temperature, so it’s worth a quick tower even on a familiar filament. For where each material’s range sits and which ones are fussiest, see our filament comparison ↗.

Calibrate once, label the profile, and that filament prints predictably from then on. The time spent here is the difference between a printer that’s a reliable tool and one that’s a constant negotiation.

For more context, Bambu Lab printer reviews ↗ covers related topics in depth.