Stop the Gaps: Dialing In Push–Pull Compensation in PE-DESIGN So Your Fills and Satins Actually Meet the Outline

If you’ve ever stitched a design that looked perfect on-screen—then watched a thin “halo gap” appear between the fill and the outline—you’re not alone. In my 20 years of embroidery education, I call this the "Gap of Doom." Most people blame hoop slip first. Sometimes that’s true. But very often the real culprit is push–pull distortion inside the stitch physics, and the fix lives in your digitizing choices.

This post rebuilds the exact lesson shown in the video (five visual examples), then adds the missing “shop-floor” context: how to estimate compensation, how to test without wasting blanks, and how to tell when hooping/stabilizer is the real problem.

Push–Pull Compensation in PE-DESIGN: The calm explanation behind scary registration gaps

In the video, Kathleen’s key point is simple: running stitches don’t distort much when properly stabilized, but fills and satins do—and that’s why your registration issues usually show up at the edges of fills/satins, not in the running outline.

Here’s the mental model I teach new digitizers to visualize the physics: Think of your fabric as cookie dough.

• “Pull” (Contraction): When the needle penetrates repeatedly in a line, it shortens the fabric in that direction. Think: Rolling the dough out, then watching it shrink back when you let go.
• “Push/Pooch” (Expansion): As thread builds up, it pushes the fabric outward perpendicular to the stitch. Think: Squeezing a balloon—if you squeeze the sides, the top expands.

So when you digitize a perfect rectangle, the fabric doesn’t sew a perfect rectangle. It sews a rectangle that has been “argued with” by tension (usually 100g-130g on top thread), stitch direction, and fabric recovery.

A lot of beginners try to solve this by cranking stabilizer or tightening the hoop until the fabric is a drum. That can create puckers, hoop burn, and still won’t fix a design that simply needs compensation.

Vertical vs. horizontal fill stitches: why the same square behaves like two different animals

Kathleen demonstrates this with two fill blocks:

• A green block with vertical stitches
• A red block with horizontal stitches

Her rule (worth memorizing):

• Vertical stitches tend to pooch out left/right and pull in top/bottom.
• Horizontal stitches tend to pooch out top/bottom and pull in left/right.

That’s why a running-stitch rectangle (your “truth line”) can end up showing gaps on certain sides depending on stitch direction.

Sensory Check: Run your finger along the finished embroidery. If the edge feels hard and raised significantly higher than the outline, you have "Pooch" (Push expansion). If there is a visible gap where the fabric shows through, you have "Pull" (Contraction).

The “outline is truth” habit (and why it saves you hours)

If you’re digitizing for clean borders, treat the running stitch outline as the reference that should remain visually correct. Then you adjust the fill/satin objects to meet it after distortion.

This is also where hooping comes back into the conversation: if your hooping is inconsistent, you’ll chase compensation settings forever.

If you’re doing a lot of repeat work (logos, team shirts, workwear), a consistent hooping workflow matters as much as the digitizing. One practical upgrade path many shops take is moving from hand-marking and “eyeballing” to a station-based workflow using a hooping station for embroidery. This isn't just about speed; it's about mechanical consistency so that your "Push-Pull" variables remain constant from shirt #1 to shirt #50.

Fill-to-fill registration: where the gap *should* land (and why it lands in the middle)

In the fill-to-fill example, Kathleen explains that even if the blocks don’t look like they’ll meet the outline perfectly on-screen, the distortion can cause them to “stack up” and meet where you expect.

Her key warning is about scale: these are “rather large objects,” and large objects accumulate more pull force. Translation: the bigger the fill area, the more you must plan for distortion.

If you’re testing a new compensation approach, don’t test on a tiny 1-inch sample and assume it scales. A 4-inch fill behaves differently.

Prep Checklist (Pre-Flight Safety Check)

Before you touch pull compensation settings, secure your physical baseline.

• Check the Bobbin: Look at the underside of a test stitch. You should see 1/3 white bobbin thread in the center and 2/3 top thread color on the sides. If it's all white, your tension is too tight, exaggerating pull distortion.
• Audit Object Types: Identify which objects are running, fill, and satin.
• Map Directions: Note stitch directions for each fill block (vertical vs. horizontal matters).
• Define "Truth": Decide what is your “truth line” (often the running outline).
• Match Materials: Plan a test stitch-out on the exact same fabric type (with the same stabilizer) you’ll actually use. You cannot test a T-shirt design on stiff felt.

Fill-to-run in PE-DESIGN: the classic “gap on the left and right” problem

Kathleen opens the “Fill to Run” example: a green horizontal fill inside a black running stitch box.

Because the fill stitches are horizontal, she explains you’ll typically see:

• Pulling in on the left and right (gaps appear there)
• Pooching on the top and bottom (less likely to gap there)

The fix she teaches is not mystical: you need pull compensation so the fill extends slightly beyond the run line in the software, so that after the fabric pulls in, it lands right on the outline.

A practical rule of thumb (so you’re not guessing forever)

The comments ask the question every digitizer asks: “How much pull compensation do I need?”

Kathleen explains that larger fills need more compensation and looser fabrics (like knits) distort more. She mentions that for a design of that size she usually uses about 0.3mm pull comp.

Here is my "Sweet Spot" table for setting Pull Compensation based on 20 years of trial and error:

Workflow to avoid "Guessing":

1. Start with the value from the table above.
2. Stitch a test.
3. Don't panic-adjust. Adjust in small increments (0.1mm at a time).

Don’t jump from “gap” to “max compensation.” Overcompensation creates a different defect: outlines that look swallowed or edges that feel heavy/bulletproof.

Fill-to-satin borders: why a 2.0 mm satin can hide sins a running stitch never will

Kathleen opens “Fill to Satin,” then goes to the Sewing Attributes panel and points out the border is a 2.0 mm satin stitch (zigzag width).

Her insight is gold: satins have width, so they’re more forgiving than a running stitch. You can overlap the fill under the satin and still get a clean edge.

She also notes another lever: you can give the satin stitch “a little more length (width)” to ensure coverage.

This is one of those moments where beginners over-focus on “perfectly meeting lines.” In commercial production, we want controlled overlap because it acts as a safety net across different fabric batches.

Setup Checklist (to make satin borders stitch cleanly)

• Confirm Object Type: Ensure the border is truly satin (not a run or a bean stitch).
• Verify Width: Check the satin width. Sweet Spot: Beginners should aim for 2.5mm to 3.5mm for borders. Anything under 1.5mm is prone to sinking into the fabric pile.
• Check Overlap: Ensure the fill extends under the satin enough (at least 30-40% of the satin width) to survive pull-in.
• Watch Conclusions: Keep an eye on corners—coverage failures show there first.
• Density Check: If you widen satin for coverage, re-check density so it doesn’t get stiff. A standard density is 0.40mm-0.45mm.

Warning: Physical Safety
When testing borders or watching closely for gaps, keep fingers clear of the needle bar. Never “help-feed” fabric near the needle area. If a needle hits a hard spot or the hoop frame, it can shatter, sending metal shards flying. Always wear glasses when observing closely.

Satin-to-run with a 9 mm column: the moment pull compensation stops being optional

Kathleen opens “Satin to Run” and calls out what many of us learn the hard way: satins distort the most, especially long satins.

In the video, the satin column is about 9 mm wide. She goes into Sewing Attributes and changes the Pull compensation setting. On-screen, you can literally see the Green Satin bar widen/extend past the black outline.

That visual jump is the whole lesson: you’re intentionally digitizing “too wide” so the sewn result becomes “just right.”

When you should redesign instead of compensating

If you’re running very wide satins (like ~9 mm) on unstable fabric, compensation alone may not save you. 9mm is approaching the jump-stitch limit for many machines (often 10mm-12mm), which can confuse the trimmer.

Often, you may need to:

• Break the satin: Split it into smaller segments or use a "Split Satin" fill pattern.
• Structural Underlay: Add a heavy underlay (Edge run + Double Zigzag). Think of this as the rebar in concrete.
• Slow Down: Reduce speed on dense satin areas. If your machine runs at 1000 SPM, drop to 600 SPM to reduce the "whip" effect on the thread.

If you’re stitching on bulky garments (like Carhartt jackets) or slippery performance wear and you’re seeing “random” gaps that change from hoop to hoop, your issue likely isn't software—it's heavy fabric shifting in the hoop. Many operators switch to magnetic embroidery hoops for these scenarios. They provide consistent, powerful tension all around the hoop without the "tug-of-war" required by traditional screw hoops, reducing hoop burn and shifting.

Warning: Magnetic Field Hazard
Magnetic embroidery hoops use powerful neodymium magnets. They can pinch fingers severely if they snap shut unexpectedly. Keep them at least 6 inches away from pacemakers or other implanted medical devices.

Satin-to-satin overlap: why “too much overlap” is often exactly right

In the final example, Kathleen opens “Satin to Satin” with two adjacent satin columns (red and green). She zooms in and shows a heavy overlap in the middle.

Her explanation is the truth: both satin columns will pull inward away from their digitized edges, so they need to be digitized overlapping so that after pull-in they meet somewhere in the middle.

This is also why purchased files sometimes disappoint: the digitizer may have tested on one fabric type, but your fabric behaves differently. Overlap is one of the first things that reveals whether a file was tested thoughtfully.

“I can’t adjust pull compensation”—what that usually means (and what to do next)

A commenter asks what to do if they can’t adjust pull compensation. Kathleen’s reply is a key diagnostic regarding file formats:

• Object-Based (e.g., .EMB, .PES in native editor): You can change density, compensation, and underlay. Example: editing a Word document.
• Stitch-Based (e.g., .DST, .JEF, .EXP): You are mostly stuck with the raw coordinates. Example: trying to edit a flattened PDF.

The Pro Fix: If you run a shop, you must build a library of editable master files. If you only buy .DST files, you are at the mercy of the original digitizer's settings.

The jagged-edge newbie problem: running vs satin isn’t the real question

One newbie comment asks how to specify satin stitches over running stitches, then clarifies they used a built-in BMP clipart in PE11 and the edges are jagged.

That’s a classic Auto-Digitizing Trap. Jagged edges usually come from the source artwork (low resolution, rough thresholding) and how the software auto-traces it.

In plain shop language: Garbage In, Garbage Out. Before you argue about satin vs running, make sure the shape you’re stitching is actually clean. If you’re digitizing logos with small details (like hands/fingers), you’ll often need to simplify shapes manually. Don't trust the "Magic Wand" tool for professional results.

Decision tree: is your “gap” a digitizing problem or a hooping/stabilizer problem?

Use this structured troubleshooting logic before you change any software settings:

Symptom Check: "My embroidery has white gaps."

1. Is the gap consistent in the EXACT same location on 3 different tests?
   • YES: It's a Digitizing/Push-Pull issue. (Go to Step 2)
   • NO (It moves around): It's a Hooping/Stabilizer issue. (Go to Step 5)
2. Is the gap between a fill/satin and an outline?
   • YES: Add overlap or pull compensation (0.2mm start) on the fill/satin object.
   • NO: Check sequencing and underlay choices.
3. Is the outline a running stitch?
   • YES: Expect it to “tell the truth.” Compensate the fill/satin to meet it.
   • NO (Satin Border): Widen the satin slightly (increase by 10-20%) to cover the gap.
4. Is the satin very wide (>7mm)?
   • YES: Expect heavy distortion. Maximize underlay. Reduce Speed.
   • NO: Small compensation changes should show results quickly.
5. Does the fabric shift, pucker, or show "Hoop Burn" (shiny residue marks)?
   • YES: This is a physical failure.
     • Solution A: Switch from Tearaway to Cutaway stabilizer.
     • Solution B: Use spray adhesive (like 505 spray) to bond fabric to stabilizer.
     • Solution C: Many professionals search for how to use magnetic embroidery hoop tutorials to eliminate hoop burn and uneven tension.

The “hidden” prep pros do before blaming pull compensation

Push–pull is real, but it’s not a free pass to ignore fundamentals. Before you chase settings, make sure your stitch-out conditions aren’t sabotaging you.

• Hidden Consumable 1: Temporary Spray Adhesive. A light mist prevents the fabric from "flagging" (bouncing) which ruins registration.
• Hidden Consumable 2: New Needles. A dull needle pushes fabric down before piercing it, causing massive distortion. Change needles every 8-10 hours of runtime.
• Stabilizer Selection: If it stretches, it needs Cutaway. If it's stable, use Tearaway.
• Hoop Tension: It should sound like a drum when tapped (thump-thump), but not so tight that the grain of the fabric looks warped.

If you’re hooping on a brother embroidery machine and you’re seeing results that vary wildly between hoops, the fastest improvement is usually consistency: same stabilizer, same hooping tension, same placement method.

Operation Checklist (a repeatable test stitch-out that actually teaches you something)

To master compensation, you must isolate variables:

• Baseline: Stitch the design with default settings.
• Variable: Stitch a second time with ONE specific change (e.g., increased Pull Comp on the fill by 0.3mm).
• Documentation: Photograph the result flat (don’t judge while it’s still stretched in it hoop).
• Measurement: Measure the gap location (Is it left/right? That's Pull. Is it top/bottom? That's Push).
• Targeted Fix: Adjust only the object that needs it, not the global settings.

The upgrade path: when better hooping tools save more money than “perfect” digitizing

Digitizing skill fixes a lot—but production reality is that you also need speed and repeatability.

• Level 1 (Hobby): If you’re doing occasional personal projects, mastering software compensation and using good stabilizers (Level 1 fix) is sufficient.
• Level 2 (Side Hustle): If you’re doing batches (work shirts, club gear), time lost to re-hooping and inconsistent placement becomes your biggest hidden cost. That’s where tools like a magnetic embroidery hoop pay off—they clamp fabric instantly without the "screw-tightening mechanics" that cause hand fatigue and uneven tension.
• Level 3 (Business Scaling): If you are consistently running orders of 20+ pieces, the bottleneck becomes the single-needle machine itself. Moving to a multi-needle platform, like the brother pr680w or a dedicated commercial machine, allows for higher speeds, fewer thread changes, and the stability needed for perfect registration on every run.

Good digitizing makes the design possible; good tools make the business profitable.

FAQ

• Q: In PE-DESIGN, why does a “halo gap” appear between a fill stitch area and a running-stitch outline even when the on-screen preview looks perfect?
  A: This is commonly push–pull distortion, so the fix is usually adding pull compensation/overlap to the fill (not tightening the hoop harder).
  • Confirm the outline is a running stitch and treat it as the “truth line.”
  • Increase pull compensation on the fill in small steps (about 0.1 mm per test); a safe starting point is often 0.2 mm, and some designs land around 0.3 mm depending on size/fabric.
  • Re-stitch on the same fabric + same stabilizer you will use in production.
  • Success check: the fill stitches land cleanly under/against the running outline with no fabric showing through at the edge.
  • If it still fails: follow the “gap moves vs. gap repeats” test—if the gap shifts between hoopings, troubleshoot hooping/stabilizer consistency instead of digitizing.
• Q: In PE-DESIGN, how do horizontal vs. vertical fill stitch directions predict where gaps will show around a running-stitch box?
  A: Stitch direction determines which sides “pull in” (gap) and which sides “pooch out,” so adjust compensation based on direction.
  • Expect horizontal fills to pull in on left/right (gaps there) and pooch on top/bottom.
  • Expect vertical fills to pull in on top/bottom (gaps there) and pooch on left/right.
  • Change only one variable at a time (direction or compensation) and test again.
  • Success check: gaps consistently reduce on the predicted sides after a small compensation change.
  • If it still fails: verify top/bobbin tension first, because overly tight tension can exaggerate pull.
• Q: What is the correct bobbin tension “success check” before adjusting pull compensation for registration gaps?
  A: Use the underside thread balance as the baseline—bad tension can mimic or amplify pull distortion.
  • Stitch a small test and flip it to the underside.
  • Look for roughly 1/3 white bobbin thread in the center and 2/3 top thread color on the sides.
  • Avoid “panic-adjusting” compensation until this baseline looks reasonable.
  • Success check: the underside shows a balanced blend (not all bobbin showing).
  • If it still fails: correct tension first, then re-test pull compensation in 0.1 mm steps.
• Q: What pull compensation starting values are recommended for PE-DESIGN based on fabric type (woven, denim, polo knit, T-shirt jersey)?
  A: Start with a fabric-matched baseline, then fine-tune in small increments after a real stitch-out.
  • Use these starting points: Woven cotton/twill 0.2–0.3 mm; canvas/denim 0.1–0.2 mm; pique knit (polos) 0.3–0.4 mm; T-shirt/jersey 0.4–0.5 mm.
  • Match stabilizer to fabric (stable fabrics can use tearaway; stretch fabrics need cutaway/no-show mesh cutaway).
  • Adjust only 0.1 mm per test to avoid overcompensation.
  • Success check: the gap closes without the outline looking “swallowed” or the edge feeling overly heavy.
  • If it still fails: re-check that the test fabric and stabilizer exactly match the final garment—testing on a different material will mislead compensation settings.
• Q: In PE-DESIGN, how should a fill-to-satin border be set up to hide small registration errors without creating stiff, bulletproof edges?
  A: Use controlled overlap and a sensible satin width; satin borders are forgiving, but density and corners must be checked.
  • Confirm the border object is truly satin (not running/bean stitch).
  • Set satin width in a beginner-friendly range (often 2.5–3.5 mm); very thin satins (under ~1.5 mm) tend to sink and reveal gaps.
  • Ensure the fill extends under the satin by about 30–40% of the satin width for reliable coverage.
  • Success check: corners stay covered first—no fabric peeking at corner tips after stitching.
  • If it still fails: widen the satin slightly (about 10–20%) and re-check density so the border doesn’t get overly stiff (commonly around 0.40–0.45 mm density).
• Q: In PE-DESIGN, why do wide satin columns (around 9 mm) show major distortion, and when should the design be redesigned instead of only increasing pull compensation?
  A: Very wide satins distort heavily, so compensation may not be enough—redesign plus underlay and speed control often becomes necessary.
  • Increase pull compensation on the satin so the digitized column intentionally extends past the outline.
  • Add stronger structure with underlay (edge run + double zigzag) to stabilize the satin.
  • Reduce machine speed on dense/wide satin areas (for example, dropping from 1000 SPM to about 600 SPM) to reduce whip and shifting.
  • Success check: the satin lands on the outline without random width changes or edge gaps along the column.
  • If it still fails: split the satin into smaller segments (e.g., split-satin approach) because a ~9 mm column can be near the practical limit where trimming/jumps become unreliable.
• Q: What safety precautions should be followed when checking embroidery gaps close to the needle area during test stitch-outs?
  A: Keep hands completely clear of the needle bar—never “help-feed” fabric while watching borders or gaps.
  • Stop the machine before repositioning fabric or touching the hoop area.
  • Observe from a safe distance and wear glasses when watching closely for registration issues.
  • Treat any needle strike (hard spot/hoop contact) as a hazard—needles can shatter.
  • Success check: adjustments and inspections are done with zero finger contact near the needle path.
  • If it still fails: pause, re-hoop, and re-check clearance to prevent hoop/frame contact before resuming.
• Q: How can an embroidery shop decide whether a recurring “white gap” problem needs digitizing changes, hooping/stabilizer changes, or a hooping tool upgrade?
  A: Use a simple three-level approach: diagnose repeatability first, then apply the least-cost fix that stabilizes results.
  • Diagnose: stitch 3 tests—if the gap is in the exact same location each time, treat it as digitizing/push–pull; if it moves, treat it as hooping/stabilizer shift.
  • Level 1 (technique): match fabric + stabilizer correctly, use spray adhesive to reduce flagging, replace dull needles, and adjust compensation in 0.1 mm steps.
  • Level 2 (tooling): if results vary hoop-to-hoop on slippery/bulky garments or hoop burn is frequent, a magnetic hoop can improve clamping consistency and reduce shifting.
  • Level 3 (capacity): if orders are consistently 20+ pieces and thread-change/time becomes the bottleneck, consider moving from single-needle workflow to a multi-needle platform for repeatability.
  • Success check: the same file produces the same edge registration on shirt #1 and shirt #50 with minimal re-hooping and rework.
  • If it still fails: standardize one “baseline recipe” (same stabilizer, same hooping tension, same placement method) before changing more software settings.