How to Tune Flow Rate for Better Surface Finish (October 2026)

Tuning flow rate is how you tell your slicer how much plastic to actually push out compared with what the toolpath asks for. If that number is wrong, you get ridged scratchy top surfaces, gaps between wall lines, parts that will not fit, or a nozzle that drags on the layer below. The whole process takes about fifteen minutes of printing once your E-steps and temperature are already correct.

One note before we start. If you searched for this and landed on pages about lathe feed rates and tool marks, that is machining, not FDM. Feed rate in turning means how fast the cutter moves, and nothing in this guide applies there. Here, flow rate means an extrusion multiplier, a percentage applied to every extrusion move in the sliced file.

There is also a second meaning floating around, which is max volumetric speed, the limit of how much material per second your hotend can melt. Same phrase, completely different setting. We will come back to that in step 2.

Table of Contents

• 1What You Need
• 2Step-by-Step: How to Tune Flow Rate for Better Surface Finish
• 31. Establish a Clean Baseline
• 42. Set a Starting Flow Rate
• 53. Run a Flow-Rate Test
• 64. Calculate the Corrected Flow Multiplier
• 75. Validate the New Flow on a Printed Part
• 86. Save the Flow Rate by Profile
• 9Common Mistakes
• 10Tips for a Consistent Surface Finish
• 11Frequently Asked Questions
• 12What flow rate should I use for good surface finish?
• 13Is 100 percent flow rate always correct?
• 14How can I tell if my 3D printer is under- or over-extruding?
• 15Should I tune flow rate separately for every filament?
• 16Does changing flow rate fix rough top and bottom surfaces?
• 17Why does a flow test look good but my full print still looks bad?
• 18Conclusion

What You Need

A stable extrusion setup needs four things: the printer, the filament, a slicer, and a way to measure.

Start with a printer whose E-steps are already calibrated and whose nozzle temperature is dialed in for the material you are running. Flow rate cannot fix a nozzle that is grinding filament because the drive gear is worn, a partially blocked nozzle, or filament that is creased or knotted. Fix those first or your measurement will be garbage.

Next, a digital calipers. The cheap ones with 0.01 mm resolution are enough. A resolution of 0.1 mm is genuinely not enough here, because 0.01 mm on a 0.42 mm wall is more than two percent error in your final multiplier.

Then a slicer. OrcaSlicer, PrusaSlicer, Cura, Bambu Studio and Creality Print all support this, but they do not agree on the name. OrcaSlicer and PrusaSlicer call it Flow Ratio, Cura calls it Flow, Bambu Studio hides it inside a calibration menu instead of the print settings, and Creality Print labels it Extrusion Multiplier. Same control, four names. Know which one you are clicking before you start.

Finally, a clean, dry spool of the filament you actually intend to print with. Nominal 1.75 mm filament is a label, not a measurement. Real filament usually runs somewhere around 1.74 to 1.78 mm, and a 2 percent difference in diameter becomes a 2 percent error in your multiplier, which is the entire size of the adjustment you are trying to make.

If you have a micrometer, measure the filament itself and enter that value in your filament profile rather than trusting the box label. If you do not, keep the diameter setting at its default and just accept that your multiplier absorbs the difference.

Step-by-Step: How to Tune Flow Rate for Better Surface Finish

1. Establish a Clean Baseline

Reload the filament properly. Pull the old filament out, wipe the drive gear and the idler arm, feed the new spool in fresh, and purge roughly 200 mm at printing temperature so you clear whatever colour or material was in there before.

Then confirm three things before touching any flow setting.

• The filament diameter in your filament profile matches the material you loaded, including whether it is 1.75 mm or 2.85 mm.
• The printer profile matches the machine you are printing on, so nozzle diameter, max temperatures and speeds are right.
• The nozzle temperature is one you have already validated for this filament, not a value you guessed this morning.

This step is the one people skip, and skipping it is why two people can print the same model on the same machine and get different answers. If the baseline is not stable, every measurement after it is noise.

How to tell it worked: the purge line comes out smooth and even in colour, with no bubbles, no clicking from the extruder and no grinding. If you hear clicking during the print, stop and fix the filament path before continuing.

2. Set a Starting Flow Rate

The default starting point is 100 percent. Every mainstream slicer ships with that value, and for most printers with correct E-steps it lands within a few percent of correct.

Here is what the number actually does. Your slicer computes the volume of plastic needed for each line based on line width, layer height and filament area, then converts that into an E value in millimetres of filament. The flow multiplier scales that E value on every extrusion. So a flow of 103 percent means every extrusion move is pushed out 3 percent longer than calculated.

That is why a new flow value does not mean your E-steps are wrong or right. E-steps govern how many millimetres the firmware moves for a given E command. Flow governs what the slicer asks for. They are two separate layers, and changing one does not repair the other.

Nor does flow have anything to do with volumetric capability. If your hotend cannot melt plastic fast enough at high speed, raising flow to compensate produces thin, under-melted lines and a rough surface. The fix there is more temperature or a slower print, not more flow.

How to tell it worked: at 100 percent, a well-tuned PLA machine should give you a wall that measures within about 0.02 mm of your configured line width. Anything further out is what the next step measures properly.

3. Run a Flow-Rate Test

Generate a single-wall test model. In OrcaSlicer and PrusaSlicer it is a calibration object under flow ratio; in Cura, enable Spiralize Outer Contour with 1 perimeter, 0 infill, 0 top layers and 1 bottom layer on a 20 mm cube. Bambu Studio and Creality Print ship their own flow calibration gcode that does the same job.

Print it standing on its corner so you get one continuous vertical wall, not a stack of horizontal lines. Keep the nozzle size and layer height you normally use. If you calibrate at 0.1 mm layer height and then print at 0.28 mm, the result is a multiplier for a condition you do not have.

Use enough wall height that the measurement is not dominated by the rounded start and end. Roughly 40 to 50 mm tall is comfortable.

Now read the wall. Measure it at three points around the height, away from the seam, and average them.

What you see	Direction	Do this now
Clean straight wall, lines merge with no ridge	Correct	Run the same test twice and confirm it repeats before you leave it alone
Visible V-shaped gaps between wall lines, sponge texture	Under-extrusion	Raise the multiplier in 2 percent steps and reprint
Raised ridge along the top of every line, wobbly wall	Over-extrusion	Lower the multiplier in 2 percent steps and reprint
Wall much thicker than line width, corner blobs	Over-extrusion	Check filament diameter setting is not set to 2.85 mm, then recalibrate
Wall thinner than line width even at 100 percent	Under-extrusion	Re-run E-steps first, then repeat the test
Top layer bumpy but walls clean	Usually not flow	Check cooling fan percentage and ironing settings before changing flow
Surface rough only near one side, seam area	Usually not flow	Align the seam or check retraction and wipe distance

How to tell it worked: the wall reads within 0.02 mm of your line width, has no gaps and no ridge, and looks the same at the bottom, middle and top.

4. Calculate the Corrected Flow Multiplier

Here is the formula. Write it down, because most wrong multipliers come from arithmetic slips rather than bad tests.

New flow multiplier = (configured line width / measured wall thickness) x current flow multiplier

A worked example. Your configured line width is 0.42 mm and your current flow is 100 percent. You measure the wall at 0.45 mm. Corrected flow = (0.42 / 0.45) x 100 = 93.3 percent. Enter 93.3 and reprint the same test.

Another one in the other direction. Line width 0.42 mm, measured wall 0.40 mm, current flow 96 percent. Corrected flow = (0.42 / 0.40) x 96 = 100.8 percent. Set it, retest, and expect the new wall to land within 0.01 mm of 0.42.

Measure at least three points and average them, because that averaging is where people lose accuracy. A single reading taken over a seam or a rounded corner can shift your final multiplier by several percent. And let the part cool to room temperature before you measure it, because warm plastic reads noticeably thinner than cold plastic, and measuring hot is the most common reason a correct calibration gets undone.

Do not make repeated large jumps. If you are more than about 5 percent away from correct, change the number once, reprint, and measure again rather than estimating the final value in your head. Two clean measurements beat five guesses.

5. Validate the New Flow on a Printed Part

The single-wall test tells you about one line at one speed. A real part has corners, curves, solid infill and top surfaces, and those behave differently.

Print something small that has all of them. A calibration cube or a small bracket with sharp corners and a curved wall works. Set your normal layer height, your normal line width, and your normal speeds so the test reflects your actual slicing settings.

Look at four things. Dimensional accuracy, using the calipers on a dimension you care about. Layer consistency, meaning every line has the same width from bottom to top. Top surface, which should be flat and consistent with no ridges or blobs. And seam appearance, since a seam always shows a little but should not show gaps or a raised lip.

Then adjust in small increments only. If the walls are good but the top is still rough, do not drop the whole multiplier, because that will open up gaps in the walls. Change 1 to 2 percent and change one other setting at a time.

Keep other causes in mind while you judge the result. Nozzle temperature that is too low shows up as rough surfaces and weak layer bonding. Insufficient cooling leaves stringy or lumpy tops. Retraction problems create blobs at the seam. Z-axis or first-layer problems show up on the bottom of the part. None of those are fixed by flow.

How to tell it worked: walls match your line width, no gaps, no ridge, top surface flat, and printed dimensions within about 0.1 mm of the model on a 20 mm feature.

6. Save the Flow Rate by Profile

Put the tuned value in a named filament or material profile, not in the global printer profile. The moment you set it globally, it will follow you into a PETG profile and be wrong there.

Each slicer puts it somewhere different:

Slicer	Setting name	Where it lives
Cura	Flow	Settings, Printer, Extruder, Flow
PrusaSlicer	Flow Ratio	Filament settings, custom G-code section, or right-click a filament preset to duplicate it
OrcaSlicer	Flow Ratio	Filament settings, or inside a filament profile customisation
Bambu Studio	Flow ratio	Calibration menu on the print screen, then written into the filament profile
Creality Print	Extrusion Multiplier	Filament settings for the active material

Record the details alongside the number, in the preset name or a notes file: filament type and brand, spool batch, nozzle size, nozzle temperature, your measured filament diameter, the final multiplier, and the date you tested. Without those, a number sitting in a preset is meaningless six months later when you swap nozzles.

If your firmware supports it, M500 saves flow settings to EEPROM and M221 sets the flow percentage at runtime. Worth knowing when you want to nudge a running print, though the value you set with M221 does not rewrite the sliced file.

How to tell it worked: start a new print with the profile loaded, check the sliced preview, and confirm the flow value is present before you hit print.

Common Mistakes

• Tuning flow before E-steps. E-steps set how many millimetres the firmware moves per E command. Tune those first, every time.
• Measuring the filament at one point. Take three readings along a metre of filament and average. One spot is not a diameter.
• Leaving filament diameter at 1.75 mm when the material is different. A 2.85 mm filament left at the default will hand you a multiplier far outside any sane range.
• Misreading under-extrusion as a retraction problem. Gaps between wall lines are flow. Blobs and stringing at a seam are retraction. They look similar in a photo and are completely different fixes.
• Changing flow while chasing temperature or Z-axis faults. If the first layer is bad or layers are separated, no multiplier will help.
• Calibrating on a test pattern that does not match your printing. Calibrating at a layer height and speed you never use gives you a number for the wrong condition.
• Setting the value globally instead of per material. One PLA number does not transfer to PETG, ABS or a filled filament.
• Jumping in big steps. Change 2 percent at a time and reprint. Six percent corrections from one guess usually mean the baseline was wrong.
• Expecting one number per material family. Batches and brands vary. Your PLA at 97 percent says nothing about a different spool of PLA.

Tips for a Consistent Surface Finish

Once flow is right, a few habits keep it there.

• Change flow in 2 percent steps and reprint rather than adjusting by feel mid-print.
• Retest after any temperature change, new filament brand, new nozzle or nozzle swap.
• Keep slicer settings identical between the calibration test and real prints so the multiplier stays valid.
• Check nozzle consistency occasionally by watching the tip shape, because a partially blocked nozzle produces the same symptoms as bad flow.
• Keep the filament path clean and the spool dry, since moisture in PETG and nylon bubbles and ruins surfaces regardless of flow.
• Tune seam position and cooling after flow, not before. Both change how the surface looks without changing how much material goes down.

Ironing deserves one honest note. Community advice on ironing flow rates varies widely, often citing 20 to 30 percent flow at 30 to 60 mm/s as a starting point. Ironing works far better when your extrusion is already correct, because it re-heats and re-flows a consistent bead. On a wall with gaps, ironing smears the gaps into a flat smear instead of fixing them.

And be realistic about the ceiling. Correct flow gives you consistent dimensions and clean walls. It does not fix elephant’s foot, seam visibility, stringing or a dull finish caused by low cooling. Those are separate settings with separate fixes.

Frequently Asked Questions

Conclusion

Start here: get a clean baseline with correct E-steps, a validated nozzle temperature and a known filament diameter. Print one single-wall test model, measure it at three points once it has cooled, and apply the formula once. Then verify on a small part with corners and a top surface before you save the value into that filament’s own profile.

That is how to tune flow rate for better surface finish without chasing your own tail for weeks. Accurate, repeatable extrusion is the foundation everything else sits on. Once walls match your line width and tops stop rippling, the remaining finish problems are cooling, temperature and seam work, and you will know exactly which knob to turn.