You moved the seam to the back, then to the nearest corner, then to random, and the blob is still there. Seam position decides where you see the defect. It does nothing about why there is a defect. Every wall loop starts with the nozzle at the wrong pressure and ends with it at the wrong pressure, and the seam is just the visible scar of that. Fix the pressure and the seam nearly disappears wherever it sits.
| What you see at the seam | What it means |
|---|---|
| Small gap or divot, then a bump a few mm later | Pressure deficit at loop start, overshoot at loop end: pressure advance is off |
| Blob at the same spot on every layer, stacking into a vertical zipper | Seam is aligned with no mitigation; the overlap adds up layer after layer |
| Little threads or zits scattered at random around the part | Ooze during travel, not the seam; this is an ooze-budget problem |
| Blob is worse on small-diameter circles and tight corners | Loop is too short to settle: the start and end events sit almost on top of each other |
| Blob shrinks when you slow the outer wall to ~40–60 mm/s | Pressure lag is the cause; speed is multiplying it |
| Seam is clean on PLA, ugly on PETG or PCTG | Stickier, more compressible melt: same settings need a retuned pressure advance value |
A Loop Has a Start and an End, and Both Cost You
A closed wall loop is a path that must begin and finish at the same point. At the start the nozzle has just traveled or retracted, so the melt chamber is under-pressured and the first millimeters of line are thin. That is the divot. At the end the extruder has been pushing at cruising pressure and the nozzle is about to stop, so residual pressure keeps squeezing plastic out after the motion slows. That is the overshoot. Then the end of the loop lands on top of the start, so you have a thin line and a fat line trying to occupy the same place.
Pressure advance (called linear advance on some firmware) exists to correct exactly this lag. If it is off, the slicer's seam settings are treating a symptom. A direct-drive extruder usually lands near 0.02–0.06 s in Klipper; a Bowden setup is more like 0.4–1.0 s. Your own value comes from a calibration print, not from a forum thread.
In This Order
- 1Calibrate pressure advance for this filamentRun a PA pattern or tower on the exact spool you are printing. A value tuned for PLA will under-correct PETG or PCTG. If the corners look right and the seam still blobs, move on; if the corners are also bulging or rounded, PA is wrong and nothing else will hold.
- 2Put the seam where geometry hides itAn inside corner or a concave edge hides a seam far better than a flat or curved face. On round parts, a single aligned seam placed at the back beats random placement, because random scatters tiny blobs across the whole surface instead of one controllable line.
- 3Slow the outer wallDrop the outer wall to roughly 40–60 mm/s. Pressure lag is proportional to how fast you were going when the loop ended, so halving the speed halves the overshoot. It is the cheapest fix you have; it costs time on the visible surface only.
- 4Turn on wipe-on-loop and a short retractionWiping moves the nozzle a short distance back along the path (about 1–2 mm) while retracting, so the pressure releases over the model instead of dumping onto the seam. Direct drive needs only 0.4–0.8 mm of retraction here; more just creates a new start-up divot.
- 5Try scarf seam on smooth, tall surfacesA scarf joint ramps the flow down over a long stretch (commonly 10–20 mm of lead-in and lead-out) so the start and end overlap as two tapered wedges rather than two full-height ends. It works very well on cylinders and gentle curves and is lost on small or sharp-cornered features, where there is no run-up length to use.
Why “Random” and “Nearest” Disappoint
Random seam positions turn one visible stripe into hundreds of specks, which is why they look fine from a distance and rough in the hand. Nearest keeps the seam close to the previous end point, so it wanders unpredictably and tends to run along a visible diagonal. Neither one changes the amount of plastic misplaced at each loop; they only redistribute it. For parts that get painted, sanded, or mated against another surface, an aligned seam in a known place is easier to hide and easier to post-process than a scatter.
The same logic applies to tolerances. A seam blob adds roughly 0.05–0.15 mm to the wall at that spot, which is enough to bind a press fit or rub on a mating part. On functional parts, put the seam on a face that does not touch anything.
PCTG is a confident choice when seams matter. Its modified CHDM structure gives a clean glossy outer wall, and its moisture uptake is around 0.15%, lower than PLA and a clear step up from PETG, so you are tuning pressure and not fighting steam-bubble blobs on top of it. It does ooze more than PLA, so give it its own pressure advance value and a hidden seam rather than reusing PLA settings.
If you have changed seam position three times, stop and tune pressure advance instead. DuffAM calibrates pressure advance per material and places seams on hidden geometry before a part ships, which is part of the process we run on every order.
Seams You Have to Look For
DuffAM tunes pressure advance per material and places seams on hidden geometry, so mating faces and outer walls come off the printer clean.
Order Direct at DuffAM →Bulging corners and a messy seam on the same print: how to calibrate the pressure value behind both.
Stringing Isn’t a Retraction Problem — It’s an Ooze BudgetScattered threads and zits instead of one seam line: the ooze, not the loop end, is the culprit.