A purchase order says the shirting is 133 x 72. The rolls arrive, the fabric looks right, the colour is approved — and nobody checks whether the construction matches the paper. Six weeks later the shirts are cut, and the marker yield is off by three percent because the actual fabric is 128 x 68. The dispute that follows is expensive, slow, and entirely avoidable.
EPI (ends per inch) and PPI (picks per inch) are the two numbers that describe how tightly a woven fabric is constructed. They are simple to verify, require no laboratory, and take about thirty seconds per check once you know the method. This guide explains how to find fabric EPI and PPI by yourself, when the numbers matter most, and how to turn a quick count into an acceptance decision at goods receiving.
What EPI and PPI Actually Tell You
EPI counts the warp yarns — the ends — running lengthwise in one inch of fabric. PPI counts the weft yarns — the picks — running across the width in one inch. Together they describe fabric construction: a 133 x 72 poplin has 133 warp ends and 72 weft picks per inch.
Construction density drives properties buyers care about. A higher count generally means a tighter, stronger, more stable weave with a smoother surface; a lower count means a more open, lighter, sometimes weaker fabric. Construction also feeds directly into cost: more yarns per inch means more yarn per metre, and yarn is the largest single cost component in most woven fabrics. When a supplier substitutes a 128 x 68 construction for an agreed 133 x 72, they are not making a cosmetic change — they are shipping less yarn than you paid for.
For buyers, EPI and PPI are therefore acceptance data, not trivia. They belong on the fabric specification checklist next to width, weight, and colour — and they are the fastest construction parameter to verify in the warehouse.
The Three-Step Method
The standard manual method, consistent with ASTM D3775 (end and pick count of woven fabric) and ISO 7211-2 (threads per unit length), works as follows:
Step 1: Take a representative swatch
Cut or select a swatch large enough to give you a full inch in both directions, away from the selvedge. The selvedge is woven more densely than the body of the fabric and will inflate your count. For a receiving check, the best practice is to cut the swatch at least 10 cm from the fabric edge and from a random roll in the shipment, not the top roll only.
Step 2: Mark a one-square-inch window
Using a ruler, mark a square exactly one inch by one inch on the fabric. In the warehouse, a pick glass — also called a thread counter or linen tester — does this job in one motion: it is a small folding magnifier with a one-inch (or half-inch) aperture etched into its base. Place the aperture on the fabric and count directly through the lens.
Step 3: Count ends and picks at five points, then average
Count the warp yarns crossing one edge of the square (EPI) and the weft yarns crossing the adjacent edge (PPI). Repeat at five locations across the swatch — left, centre, right, and two diagonal points — and take the average. The five-point average matters because construction is never perfectly uniform: tension varies across the loom width, and a single count near the edge can mislead you by several threads.
A Worked Example
Suppose the purchase order specifies a 133 x 72 cotton poplin. Your five EPI readings are 132, 134, 133, 131, 133 — average 132.6, effectively 133. Your five PPI readings are 71, 72, 73, 71, 72 — average 71.8, effectively 72. The fabric matches the specification within normal counting tolerance, and you can sign off the construction in under two minutes.
Now suppose the EPI readings average 126. That is not counting error; it is a different fabric. A seven-end shortfall on a 133 construction is roughly a five percent yarn reduction, which will show up in fabric weight, tear strength, and — because the fabric is more open — potentially in seam slippage and opacity. That roll should be held, re-tested on a larger sample, and raised with the supplier before cutting begins.
Pick Glass vs Ruler: Which to Use
A pick glass is the proper tool. The magnifier resolves individual yarns in fine constructions where a naked eye blurs them together, and the fixed aperture removes the marking step. A ruler works for coarse, open fabrics — denims, canvas, heavy twills — where individual yarns are visible. The common failure mode with rulers is parallax error: reading the scale at an angle shifts the count by one or two yarns, which is exactly the tolerance band that matters. If your fabrics are finer than about 100 ends per inch, buy the pick glass; it costs less than one disputed roll.
Where the Method Does Not Apply: Knits
EPI and PPI are woven-fabric parameters. Applying them to knits is a category error that appears surprisingly often in buyer specification sheets. Weft-knitted fabrics are described by courses per inch (rows) and wales per inch (columns of loops), and the counting method, the tools, and the interpretation are all different — knitted loops stretch and distort under the magnifier in ways woven yarns do not. If your product is jersey, interlock, or fleece, write courses and wales into the specification, not EPI and PPI. Verifying the wrong parameter carefully is worse than not verifying at all, because it creates false confidence.
Counting on Stripes, Checks, and Yarn-Dyed Fabrics
Patterned wovens need one extra discipline. On stripes and checks, always take your five readings inside the ground area — never across a stripe or a check intersection. Pattern yarns are often heavier, doubled, or woven at a different density than the ground, so a count that includes them describes the pattern, not the fabric. The same applies to seersucker and dobby textures: count the flat ground between the textured zones. If the purchase order specifies construction for a yarn-dyed fabric, confirm with the supplier whether the number refers to the ground or includes the pattern repeat; the two answers can differ by several threads, and the ambiguity has caused more than one avoidable claim.
Warehouse Check vs Laboratory Test
The thirty-second count is a screening tool, not a laboratory result. A formal construction test conditions the fabric to standard atmosphere before counting, because cotton and other hygroscopic fibres swell and relax with humidity, shifting the count by a thread or two. In practice this means the warehouse check and the lab report can legitimately disagree by a small margin — and the tolerance on your purchase order should be wide enough to absorb that difference. Use the manual count to catch the five-percent shortfalls that indicate a real problem; use the lab for borderline cases and for building the evidence file if a claim goes formal. Both have their place, and neither replaces the other.
Turning the Count into an Acceptance Decision
A count is only useful inside a decision rule. Three practices make the thirty-second check commercially effective:
- Specify a tolerance, not an exact number. Construction varies slightly even in compliant production. A common buyer tolerance is ±3 ends or picks on fine shirting and ±2 on coarse fabrics — but the number is a commercial agreement, so write it on the purchase order alongside the nominal construction.
- Sample the shipment, not the top roll. Mills sometimes ship the inspected roll on top and average fabric beneath. Pull swatches from randomly selected rolls across the delivery, and record which roll each swatch came from.
- Connect construction to the rest of the inspection. A construction shortfall should trigger a weight check and a garment AQL inspection conversation, because under-constructed fabric fails downstream in predictable ways — seam strength, dimensional stability, and pilling. Log the EPI/PPI readings on the same receiving report as the fabric defect scoring so the full picture sits in one document.
The Bottom Line
Finding fabric EPI and PPI is one of the highest-return checks in apparel quality control: it needs no lab, no power, and no training beyond a single demonstration, yet it detects the most common form of quiet specification drift in woven fabric — the missing yarn. Build the three-step count into your receiving procedure, give it a tolerance and a sampling rule, and treat a failed count the way you would treat a failed shade check: hold the roll, document the readings, and resolve it before the cutting room opens the first lay. Suppliers who know their buyers count threads ship the construction they promised.
