Ground ore contains particles across a wide range of sizes. A sieve analysis passes a sample through a stack of screens with progressively smaller openings and weighs the material retained on each, which gives the proportion of the sample finer than any size.
The result is usually summarised by P80, the size through which 80% of the sample passes. P80 is used to set grind targets, to size mills with Bond's equation, and to assess cyclone performance. This article covers the sieve test, the calculations and the equations used to describe a full size distribution. The P80 calculator interpolates P80 from sieve results.
Sieve analysis procedure
- Split a representative sample to a mass that will not overload the sieves. Coarse samples require far more mass than fine ones.
- Wet screen the sample over the finest sieve, because dry fines adhere to coarser particles and blind the sieves. Dry the oversize and record the mass lost as fines.
- Stack the sieves with the coarsest on top. Standard series step by a ratio of about the square root of 2, so each opening is roughly 1.4 times the next.
- Shake for a fixed time on a sieve shaker, and confirm that a further minute of shaking changes the result little.
- Weigh each fraction and the pan, and add the wet-screened fines to the pan.
- Check that the fractions sum to the starting dry mass within a small tolerance; a larger loss indicates spillage or a damaged sieve.
Cumulative percent passing
The mass of each fraction divided by the total gives the percent retained. Summing the percent retained from the top sieve downward gives the cumulative percent retained, and 100 minus that value is the cumulative percent passing.
| Sieve (µm) | % retained | Cumulative % passing |
|---|---|---|
| 1,700 | 0.5 | 99.5 |
| 1,180 | 1.7 | 97.8 |
| 850 | 3.8 | 94.0 |
| 600 | 5.9 | 88.1 |
| 425 | 7.5 | 80.6 |
| 300 | 8.7 | 71.9 |
| 212 | 9.4 | 62.5 |
| 150 | 9.5 | 53.0 |
| Pan (−150) | 53.0 | — |
P80 by interpolation
The 80% passing point lies between the 300 µm sieve (71.9%) and the 425 µm sieve (80.6%). A straight line between these two points on log-log axes gives:
where
- size that x% of the sample passes (µm), for example P80
- ,
- the two sieve sizes either side of x% passing (µm)
- ,
- cumulative % passing at those sieves; the interpolation is a straight line on log-log axes
Worked example: k = ln(80 ÷ 71.9) ÷ ln(80.6 ÷ 71.9) = 0.1067 ÷ 0.1142 = 0.934, and P80 = 300 × (425 ÷ 300)0.934 = 415 µm. Linear interpolation on ordinary axes gives a slightly different result (416 µm here), so the method should be stated and applied consistently. P80 should never be extrapolated beyond the top or bottom sieve; an additional sieve should be used instead.
Size distribution models
where
- cumulative % passing size d
- ,
- Gates–Gaudin–Schuhmann size modulus (the size 100% passes) and distribution modulus (slope on log-log axes)
- ,
- Rosin–Rammler size at which 63.2% passes, and the uniformity coefficient
Fitting either equation to the sieve data by least squares gives two parameters that describe the whole curve. The Gates–Gaudin–Schuhmann form is a straight line on log-log axes and fits the fine end of many ground products well; the Rosin–Rammler form often fits crushed products and broad distributions better. The slope (m or n) indicates the width of the distribution, a steeper slope corresponding to a narrower size range. These fits are useful for comparing surveys and as inputs to simulation, but P80 by interpolation remains the reporting standard.
Laser sizing and other methods
Laser diffraction measures fine material quickly and is common below about 38 µm, where sieving is slow. It reports volume-based sizes of particles treated as spheres, so its results do not match sieve results exactly, and the two methods should be correlated on the material concerned before their results are combined in one report. Cyclosizers and sedimentation methods separate by settling velocity, which depends on density as well as size.
Glossary
- Sieve analysis
- Sorting a sample into size fractions with a stack of sieves and weighing each fraction.
- Cumulative % passing
- The share of a sample finer than a given sieve size.
- P80
- The size 80% of a sample passes; F80 is the same for a feed.
- Wet screening
- Washing fines through the finest sieve before dry sieving.
- Sieve series
- A set of sieve openings stepping by a fixed ratio, usually about the square root of 2.
- Rosin–Rammler distribution
- An equation describing a size distribution with a characteristic size and a uniformity coefficient.
- Gates–Gaudin–Schuhmann distribution
- An equation that plots as a straight line on log-log axes, with a size modulus and a slope.
- Laser diffraction
- Sizing particles by how they scatter laser light; reports volume-based equivalent sphere sizes.
Frequently asked questions
How is P80 calculated from a sieve analysis?
Identify the two sieves either side of 80% passing and interpolate between them, usually on log-log axes. For 71.9% at 300 µm and 80.6% at 425 µm, P80 is about 415 µm.
Should samples be wet sieved or dry sieved?
The fines should be wet screened first and the remainder dry sieved. Dry sieving alone leaves fines attached to coarse particles and overstates the coarse fractions.
Why does a laser sizer give a different P80?
Laser diffraction reports volume-based equivalent sphere sizes, whereas sieves measure the second-smallest particle dimension. The two methods should be correlated on the ore concerned.
How much sample is required?
Enough for a reliable result without overloading the sieves; coarse samples require much more mass. Gy's formula sets the minimum for a given precision.
Discuss your plant with a metallurgist
Tiegel is a Perth metallurgy company. We help with testwork, plant optimisation, metal accounting and investigations, and we build the Tiegel Suite software.
Talk to a metallurgistReferences and further reading
- Wills, B.A. and Finch, J.A. (2016). Wills' Mineral Processing Technology, 8th edition. Butterworth-Heinemann. Chapter 4, Particle size analysis.
- ISO 3310-1. Test sieves: technical requirements and testing, Part 1: test sieves of metal wire cloth.
Worked examples and charts marked illustrative use invented figures, not data from any operation.