Reading a sieve analysis and spotting a failing gradation

The curve tells you what the pavement will do long before the pavement does it.

How the test is run

Following ASTM C136 with the wash step from C117:

Sample and split. Take a representative bulk sample (ASTM D75), then reduce it with a riffle splitter or by proper quartering (ASTM C702). Hand-scooping introduces bias you cannot recover from later.
Dry and weigh. Oven-dry at 110 ± 5 °C to constant mass, cool, and record the dry mass to 0.1 g. Call it M₁.
Wash over the 0.075 mm sieve. Cover with water, agitate hard to release the dust clinging to coarse particles, and pour the wash water over a protective 1.18 mm sieve nested above a 0.075 mm sieve. Repeat until the water runs clear. Re-dry everything retained and weigh again: M₂. The difference M₁ − M₂ is the fines removed by washing.
Dry sieve. Stack sieves largest to smallest over a pan, load the washed dried sample, and shake mechanically for the calibrated time, typically 10 to 15 minutes.
Weigh each fraction to 0.1 g, including the pan.
Check mass balance. The sum of retained masses must agree with M₂ within 0.3%. If it does not, the test is invalid — repeat it rather than reporting it.
Compute. Cumulative retained, cumulative percentage retained against M₁, and percentage passing as 100 minus that. Remember to include the washed-out fines in the material passing 0.075 mm.

Reading the curve

Plot percentage passing against sieve size, on a semi-log scale or a 0.45 power chart, with the job mix formula envelope drawn on.

A good dense-graded curve is a smooth continuous S, running roughly parallel to the control limits and sitting between them rather than hugging one. On a 0.45 power chart it tracks close to the straight line from the origin to the maximum aggregate size. The slope is even, which means every fraction is represented and the voids in mineral aggregate come out where the mix design intended.

A curve heading for trouble shows one of three shapes:

A hump around 0.300 to 0.600 mm. Too much middle-sized sand. This is the classic tender mix that shoves under the roller instead of compacting.
A flat stretch across the intermediate sieves. A missing size fraction — gap grading.
A steady drift toward one limit. Not a failure yet, but it tells you the plant cold feeds or the crusher screens are moving. Act on the trend, not on the first result outside the line.

What a bad gradation does on the road

Too fine. Fine particles add enormous surface area. With binder content unchanged, the aggregate ends up under-coated, which shows up as ravelling and cracking. The fines also fill the voids in mineral aggregate, so the mix loses internal friction and you get rutting and bleeding in summer. Under the roller it stays soft and pushes into a wave ahead of the drum.

Too coarse or gap-graded. Not enough fine aggregate to fill the stone skeleton leaves large interconnected voids. The mix becomes permeable, water gets in, and stripping and subgrade softening follow. It is also harsh under the screed, which gives surface honeycombing and high in-place air voids, and it segregates easily during loading and laydown because there is no cohesive mortar holding it together.

Why the wash step matters

Ultra-fine dust and clay cling to the faces of larger stones, held by moisture and electrostatic forces. Dry sieving does not break those bonds, so the fines stay stuck on the coarse sieves and get reported as coarse material.

Skip the wash and the material passing 0.075 mm can be understated by 1% to 4%. That single error moves your dust-to-binder ratio, which then drives a mix adjustment in the wrong direction. Washing is mandatory for all compliance testing, and it matters most on recycled asphalt pavement and crushed fine aggregates, where the dust coating is heaviest.

Lab errors that produce a wrong answer

Bad sampling. Scooping from the outside of a stockpile, where the coarse stone has rolled down the sides, gives a coarse-biased result before the sample reaches the lab.
Overloaded sieves. Material several layers deep blinds the mesh so small particles never reach an opening. The mass retained on that sieve is then wrong, and so is everything downstream.
Wrong shaking time. Too short and fines are stranded on coarse sieves. Too long on soft limestone and the shaker itself creates fines by grinding the edges off the stone.
Worn or damaged sieves. Stretched mesh, torn cloth or openings clogged from being scraped with a spatula.
Incomplete drying. Damp material weighed as M₁ skews every percentage in the report.
Balance not re-zeroed, or the 0.3% mass conservation check skipped.

When a result falls just outside

Do not stop production on one marginal test. Work through it in order.

Rule out the test. Check the mass balance. Inspect the specific sieves around the failing size for blinding, tears or trapped stones. Run a retest on a fresh split from the same bulk sample.
Check the tolerance you are judging against. Single-operator precision is typically around ±1% to ±2% on intermediate sieves and ±0.5% on the 0.075 mm sieve. Compare against JMF operational tolerances, not the broad master specification band.
If the retest confirms it, go to the plant. Check cold feed belt speeds and gate openings — one bin running fast changes the whole blend. Check stockpiles for segregation and for moisture causing fines to bridge in the hopper. Check the screen decks for blinding, pegging and broken wire.
Adjust and confirm. Change the cold feed proportions to pull the failing fraction back toward the middle of the JMF band, run 15 to 30 minutes, then sample again before releasing more material to site.

This guide describes general practice and typical figures. Your project specification, job mix formula and laboratory results govern the actual values on your job. Nothing here is a design or a substitute for the engineer of record.