Pull compensation in embroidery: how much to add and how to test it

What is the push and pull effect?

When the machine forms a stitch, the take-up lever pulls the top thread tight, drawing the two needle holes at each end of the stitch toward each other. Multiply that by hundreds of parallel stitches and the fabric is gathered sideways, across the stitch direction. That is pull. At the same time, all those stitches packed side by side add bulk along the direction they are stacked, and the fabric is shoved outward at the ends of the stitching. That is the push effect.

The practical result for a satin column: it finishes narrower than drawn (pull across the stitches) and slightly longer than drawn at its ends (push along the column). A fill behaves the same way relative to its row direction: it shrinks across the rows and grows along them. A circle filled with horizontal rows tends to sew as a slightly wide oval: narrower top to bottom, wider side to side.

What decides how much a design pulls?

• Fabric stretch. The biggest factor. Jersey, pique and spandex blends pull far more than twill or canvas.
• Fabric thickness and loft. Thicker fabric means more fabric for each stitch to wrap around, so the thread "uses up" width going over and under it. Fleece, towels and caps need more.
• Stitch density. More stitches per mm means more pull. See stitch density.
• Stitch length. Long satin stitches across a wide column pull harder than short ones.
• Thread tension. Tighter top tension pulls harder. A file tuned on a machine with balanced tension may look thin on one running tight. Check thread tension before re-tuning the file.
• Stabilizer and hooping. Firm cutaway and drum-tight hooping resist pull; loose hooping or floating multiplies it.
• Underlay. Good underlay locks the fabric first, so less compensation is needed.

Typical starting values

These are common starting points for 40 wt thread with suitable stabilizer and underlay. Software differs in how it applies the value (per side or total width), so check your program's documentation and treat these as values to test, not rules.

Percentage vs absolute compensation

Software offers pull compensation as a fixed distance (mm), a percentage of the shape's width, or both.

• Fixed mm adds the same amount to every column. A 0.2 mm boost is a lot on a 1.5 mm lettering stroke and almost nothing on an 8 mm border. It suits narrow satins and lettering.
• Percentage scales with width, which matches how wide columns pull more in absolute terms. But a small percentage on a 1.2 mm column adds only a few hundredths of a mm, not enough to matter.
• Combined (where available): a small fixed minimum plus a modest percentage covers both cases. Many digitizers set something like a small fixed amount and adjust the percentage by fabric.

Narrow satins under about 2 mm are the sensitive zone. Small lettering is where under-compensation makes strokes look like thin threads and over-compensation closes up the counters of letters like "e" and "a". See lettering for text-specific rules.

Push compensation: shortening the ends

Push is smaller than pull and fewer programs expose a setting for it. Where it matters most is the ends of satin columns and the leading edges of fills in the direction the rows are stacked: column ends can bulge past an outline or cross into a neighbor. Digitizers handle it by drawing satin ends slightly short, especially where a column meets another object end-on, and by keeping overlaps between objects modest. On a satin letter, for example, the tips of strokes may be pulled back a fraction of a millimeter so they do not overshoot.

Overlaps: the partner of compensation

Compensation keeps each shape at its intended size; overlaps cover what compensation does not. Where a fill meets another fill or an outline, digitizers typically extend the underneath object a few tenths of a mm under the neighbor (often around 0.3 to 0.5 mm, more on stretchy fabric). This is the main defense against the white gap between a fill and its border. If you see gaps, check overlaps and sewing order before raising compensation: the full picture is on registration, and the machine-side checks are on outlines not lining up.

How to test pull compensation

1. Build a test file. Draw satin columns at 1.5, 3 and 6 mm wide and a 40 mm square fill with an outline around it. Duplicate each at three compensation values, for example 0.1, 0.25 and 0.4 mm.
2. Use the real job setup. Same fabric, stabilizer, hooping, thread and needle. Compensation tuned on a different fabric is a guess.
3. Sew and measure. Measure column widths with calipers or a fine ruler at several points and compare with the design widths. On the fill square, look at whether the outline sits evenly around all four sides.
4. Look along the stitch direction too. Check whether column ends overshoot (push) and whether the square's sides parallel to the fill rows have grown.
5. Pick and record. Choose the value that hits the drawn width with a clean edge. Save it as a fabric preset in your software.

Symptoms of too little and too much compensation

Satins thinner than designed, fabric peeking at edges

Too little compensation, or tension too tight. Add 0.1 mm and retest.

Gaps between fills and outlines on one or two sides

Too little compensation across the fill rows, often combined with missing overlap. The gap appears on the sides parallel to the rows.

Lettering strokes merge, counters filled in

Over-compensation on narrow columns. Reduce the fixed amount for text.

Outlines look heavy or ragged, fills bulge past borders

Over-compensation, so shapes spill beyond their outlines.

Wavy satin edges

Usually underlay or stabilizer, not compensation; uneven pull along a column points to fabric shifting. See design distortion.

Round shapes sew as ovals

Pull across the fill rows and push along them. Add compensation and consider a fill angle change or slightly reshaping the object.

When compensation is not the fix

Compensation corrects predictable, even distortion. It cannot fix fabric that slides in the hoop, stabilizer that tears away halfway through, or a sewing order that lets distortion pile up on one side. If distortion is different each time you sew the same file, the cause is physical: check hooping (see how to hoop) and stabilizer weight first. If it is the same every time, the file is the place to fix it.