The I-test and H-test: how to check embroidery tension from the back

The I-test turns tension from guesswork into something you can see and measure. It takes a few minutes and works on every lockstitch embroidery machine, from a 4x4 home machine to a multi-needle head.

Updated 5 min read Reviewed against our editorial standards

Quick answer

Stitch a satin column (a block letter I, or an H to test two directions) on your usual fabric and stabilizer, then look at the back. Balanced tension shows about one third bobbin thread in the center and one third top thread on each side. More bobbin means the top is tight; less means it is loose.

On this page
  1. What the I-test shows
  2. How to run the I-test
  3. The H-test: two directions and many needles
  4. Reading the result and what to change
  5. Why the test can mislead you
  6. Make your own reusable test sheet

What the I-test shows

A satin stitch carries the top thread across the column, down through the fabric, and back. At each needle point the top thread locks with the bobbin thread. With balanced embroidery tension, each lock sits just past the back of the fabric, a little in from the column edge. Seen from the back, the top thread wraps around the edges and the bobbin thread lies in a band down the middle. The proportions of that band are the measurement.

Back of three satin columns: top too tight, balanced, top too loose Top too tight mostly bobbin Balanced about 1/3 bobbin Top too loose little or no bobbin
Rust is top (embroidery) thread, blue is bobbin thread, viewed from the back of the hoop.

The one-third figure is a convention, not a law. Many experienced embroiderers accept anything from about one quarter to about one half bobbin as working tension, as long as no bobbin shows on the face and the satin lies smoothly. Aim for one third and you will have margin both ways.

How to run the I-test

  1. Get a test file. Many machines have built-in block fonts: a capital I or a simple bar in a sans-serif font, sized about 25 to 40 mm tall, gives a good satin column. You can also draw a plain satin column in any digitizing program, including free Ink/Stitch.
  2. Use the real materials. Hoop the fabric and stabilizer you will actually use. A test on stiff cutaway can hide a problem that appears on a thin knit.
  3. Choose contrasting colors. A dark top thread over a white bobbin makes the stripes easy to see.
  4. Stitch at your normal working speed. Tension behaves differently at slow speed; test where you plan to run.
  5. Unhoop and turn it over. Look at the middle of the column, not the ends, where tie-ins and tie-offs crowd the stitches.
  6. Estimate the proportion. Divide the column width by eye into thirds. Note whether the bobbin band is centered and even along the length.

The H-test: two directions and many needles

Tension can behave differently depending on which way the hoop is moving. A block capital H has two vertical satin legs (stitches running side to side) and a horizontal crossbar (stitches running up and down), so one letter checks both directions. Read each part separately: if the legs are balanced and the crossbar is not, the issue is often thread feed or hooping, not the tension setting alone.

On multi-needle machines the H-test becomes a whole-head check. Stitch a row of H letters, one per needle, assigning each needle a different color in the design's color sequence. One stitch-out shows the balance of every needle side by side, and the needle that stands out is the one to inspect. Typical causes of a single odd needle are a missed thread guide, thread wrapped twice around the break sensor wheel, a worn needle, or lint in that needle's tension discs.

Reading the result and what to change

I-test patterns and the first adjustment to try
Back of the columnFace of the columnMeaningFirst change
About 1/3 bobbin, centeredClean, no bobbinBalancedNone. Record your settings.
More than 1/2 bobbinBobbin dots at edgesTop tight or bobbin looseLower top tension one step
Little or no bobbinSatin looks loose or loopyTop loose or bobbin tightRethread top, then raise top one step
Bobbin band off centerEdges unevenThread feed or pull compensation, not tensionCheck thread path and the design
Band varies along the columnPatchyInconsistent feed: spool catching, uneven bobbin, lintFix feed before touching tension
Good on test, bad on the real designVariesDensity, stabilizer or flaggingSee density and hooping

Change one thing at a time and retest. Lowering the top is the first move for "too much bobbin" because it is reversible and the bobbin is usually right; see embroidery thread tension and, only if needed, bobbin tension.

Why the test can mislead you

Testing on the wrong base

Stabilizer and fabric change the result. A heavy cutaway on denim resists draw-in and can make tension look tighter than on a lightweight tearaway. Always test on a stack that matches the job.

Very narrow or very wide columns

Under about 2 mm, the bands are too small to judge and small text has its own issues (see small lettering problems). Over about 7 mm, home-machine satins become long floats that can sag, and many digitizers split them; see satin stitch problems.

Specialty threads

Metallic, 12 wt and 60 wt threads change the balance. The one-third target still applies, but you will reach it at different settings. Run a separate test and record the settings for each.

Underlay

Heavy center-run or zigzag underlay under the column lifts the satin and can hide part of the bobbin band. For the cleanest reading use a test column with light or edge-run underlay.

Make your own reusable test sheet

A single test file that you run every time you change a variable saves hours. A useful layout includes:

  • A 5 mm satin capital H, stitched in each needle or color you use.
  • A 25 x 25 mm fill block with a satin border, to show registration and push and pull.
  • A line of small text around 6 to 8 mm tall, the size where weaknesses appear first.
  • A run-stitch outline offset slightly from the fill, to show alignment.

Label each sample on the stabilizer with the date, thread, needle, bobbin and settings. Over time this becomes your machine's reference book. The stitch-out tuning method uses exactly this kind of sheet in a fixed order of checks.

Frequently asked questions

How wide should the satin column be for an I-test?

Somewhere around 4 to 6 mm works well. Narrow columns under about 2 mm make the stripes too small to judge, and very wide satins can float loosely and give a misleading result.

Does the one-third rule apply to fill stitches?

It applies to satin columns. On fills the stitches are short and overlap, so the pattern is harder to read. Tune on a satin column, then confirm that the fill looks right on the face with no bobbin peeking through.

What color bobbin thread is best for reading the test?

Use a top thread color that contrasts strongly with your bobbin, such as a dark blue or red over a white bobbin. The stripes are then easy to measure by eye.

How often should I run an I-test?

Whenever you change thread brand or type, needle size, bobbin type or a major fabric category, after any service or cleaning that disturbed the tension, and at the start of a production run on a multi-needle machine.

Sources and further reading

  • Madeira USA tension testing guidance from a thread maker
  • Isacord embroidery thread and tension resources
  • Ink/Stitch free digitizing software for making a custom test file

Found something wrong or out of date? Tell us and we will fix it.