Foundry sand: calculate AFS GFN and move from 49 to 72

Direct answer: what proportion moves AFS 49 to AFS 72?

The Minerallum team reviewed the American Foundry Society calculation methods, the Silica and Moulding Sands Association guidance and several experimental studies on moulding sands.

The mathematical answer is direct:

To move an AFS 49 sand towards an AFS 72 target, the added sand must have an index strictly higher than 72.

When the available fine sand has an AFS GFN of 100, the theoretical starting blend is:

• 54.9% AFS 49 sand;
• 45.1% AFS 100 sand;
• proportions calculated on a dry-mass basis.

For one metric tonne of base sand:

• 549 kg of AFS 49 sand;
• 451 kg of AFS 100 sand;
• theoretical blend AFS: 72.

This calculation establishes the initial mass balance. The industrial formulation is then confirmed through:

1. sieve analysis of the actual blend;
2. review of the distribution on every sieve;
3. control of the fine fraction;
4. permeability and strength testing;
5. a trial in the relevant moulding system.

AFS GFN is useful for setting a target, while the complete particle-size distribution shows how that target has actually been produced.

Why this question has strong industrial value

The objective is not merely to define AFS GFN. It is to address a specific production requirement: turning an available raw material into a formulation compatible with a moulding target.

A foundry may seek a higher fineness to:

• reproduce finer details;
• modify mould-face density;
• reduce certain forms of metal penetration;
• adapt sand to a lighter or more complex casting;
• stabilise a feed assembled from several sources;
• compensate for changes in reclaimed sand;
• prepare a facing sand different from the backing sand.

Our review of the references shows that changing fineness modifies several balances at the same time.

The Kansas Geological Survey links grain size and size distribution to mould density and permeability. The distribution must allow gases to escape while producing a structure dense enough to resist penetration by molten metal [3].

More recent experimental research also treats AFS GFN together with variables such as:

• moisture;
• binder type and dosage;
• mould hardness;
• compactability;
• compressive or bending strength;
• permeability;
• grain morphology [4–6].

We therefore retain a clear working rule:

The AFS target defines the grading direction; tests on the moulding system determine the final formulation.

What does AFS GFN actually mean?

AFS GFN, or American Foundry Society Grain Fineness Number, is calculated from the mass or percentage of sand retained on a standard series of sieves.

The American Foundry Society describes it as an estimate of the average sieve size of a loose, dry sand sample [1].

The calculation principle is:

1. dry and prepare the sample;
2. pass it through a stack of sieves;
3. weigh the material retained on each sieve;
4. convert the masses into percentages;
5. multiply each percentage by its corresponding AFS factor;
6. sum the products;
7. divide by the total percentage recovered.

The general equation is:

AFS GFN = Σ (% retained × AFS factor) / Σ (% retained)

When 100% of the sample is recovered, the denominator is 100.

How to read the number

As a general interpretation:

• a lower AFS value corresponds to a coarser sand;
• a higher AFS value corresponds to a finer sand.

This should not be converted into a unique relationship between AFS and particle diameter.

SAMSA distinguishes between:

• AFS, based on older mesh numbers and their factors;
• AGS, an average grain size calculated in micrometres.

The association states that there is no universal direct conversion between the two indicators [2].

Why AFS does not describe the full curve

Two sands can obtain the same AFS value even when:

• the first is concentrated on three adjacent sieves;
• the second is distributed across five or six sieves;
• the third combines a coarse fraction with a very fine fraction;
• their amounts below the last sieve differ;
• their grain shapes are not the same.

The average number may be identical, while the granular structure is not.

This is why our review of a report begins with AFS GFN and then systematically returns to the percentage retained on every sieve.

The equation for blending two sands

When two sands have been analysed with the same method and sieve series, the theoretical AFS of their blend follows a mass balance.

Let:

• G₁ be the AFS of the initial sand;
• G₂ be the AFS of the added sand;
• Gᵗ be the target AFS;
• x be the dry mass fraction of the added sand.

The relationship is:

Gᵗ = (1 − x) × G₁ + x × G₂

The fraction of added sand is therefore:

x = (Gᵗ − G₁) / (G₂ − G₁)

For a target of 72 starting from 49:

x = (72 − 49) / (G₂ − 49)

This relationship follows directly from the definition of GFN as a weighted sum of the retained fractions.

Conditions for a coherent calculation

We apply the equation when:

• both results are established on dry material;
• both sands use the same sieve series;
• sample preparation methods are comparable;
• recovered masses are properly documented;
• the blend does not yet contain binder, water or additives;
• the added fine fraction has an AFS higher than the target.

When these conditions are not met, the blend should be recalculated from the mass retained on each sieve, rather than from the two AFS values alone.

Complete example: AFS 49 + AFS 100

The objective is:

• initial sand: AFS 49;
• fine sand: AFS 100;
• target: AFS 72.

Calculation:

x = (72 − 49) / (100 − 49)

x = 23 / 51 = 0.451

The fine sand therefore represents 45.1% of the dry blend.

The fraction of initial sand is:

1 − 0.451 = 0.549

or 54.9%.

Preparing a 100 kg pilot batch

Component	AFS	Dry proportion	Mass
Initial sand	49	54.9%	54.9 kg
Fine sand	100	45.1%	45.1 kg
Theoretical blend	72	100%	100 kg

Check:

(54.9 × 49 + 45.1 × 100) / 100 = 72

Theoretical preparation of one metric tonne

Component	Dry mass
AFS 49 sand	549 kg
AFS 100 sand	451 kg
Total	1,000 kg

In production, charging masses are adapted to the accuracy of the dosing equipment, after which the homogenised blend is sampled for analysis.

Influence of the selected fine-sand AFS

The closer the added sand is to AFS 72, the larger the required proportion.

Fine-sand proportion required to move AFS 49 to AFS 72

Theoretical dry-mass calculation according to the AFS GFN of the available fine sand.

Fine sand AFS 80 74.2 %

74.2% AFS 80 sand and 25.8% AFS 49 sand.

Fine sand AFS 90 56.1 %

56.1% AFS 90 sand and 43.9% AFS 49 sand.

Fine sand AFS 100 45.1 %

45.1% AFS 100 sand and 54.9% AFS 49 sand.

Fine sand AFS 120 32.4 %

32.4% AFS 120 sand and 67.6% AFS 49 sand.

Decision table

AFS of added sand	Added-sand proportion	AFS 49 proportion	Interpretation
72	100%	0%	No blend containing AFS 49 can reach 72
80	74.2%	25.8%	High consumption of fine sand
90	56.1%	43.9%	Nearly balanced blend
100	45.1%	54.9%	More flexible industrial ratio
120	32.4%	67.6%	Lower proportion; ultrafines require close control

Selection should therefore not be based solely on price per tonne.

The comparison should include:

• required proportion;
• delivered cost;
• lot-to-lot stability;
• exact particle-size distribution;
• dust content;
• grain shape;
• binder interaction;
• effect on mould properties;
• losses and reclamation.

Why blending AFS 49 with AFS 72 is not enough

This question requires an explicit answer.

With:

• G₁ = 49;
• G₂ = 72;
• target = 72;

the equation gives:

x = (72 − 49) / (72 − 49) = 1

The blend therefore requires 100% AFS 72 sand.

Any presence of AFS 49 sand lowers the average.

For example:

AFS 49	AFS 72	Theoretical AFS
50%	50%	60.5
25%	75%	66.25
10%	90%	69.7
0%	100%	72

To retain part of the AFS 49 sand while reaching 72, the second sand must be finer than the target.

Sieve-by-sieve calculation: the more informative method

The two-index calculation is fast. A sieve-by-sieve calculation shows the actual structure of the blend.

For every sieve i:

Rᵐᵢ = (1 − x) × R¹ᵢ + x × R²ᵢ

where:

• R¹ᵢ is the retained percentage of the AFS 49 sand;
• R²ᵢ is the retained percentage of the fine sand;
• Rᵐᵢ is the predicted retained percentage in the blend.

This produces a complete predicted curve before the first trial.

What the curve can reveal

It can show:

• dominant sieves;
• continuity of the distribution;
• excessive concentration on one sieve;
• a bimodal distribution;
• increases below 106 or 75 µm depending on the sieve series;
• the pan fraction;
• movement towards very fine sizes;
• the difference between the target average and the actual structure.

The Kansas Geological Survey report illustrates this point with several sands having similar AFS values but different retained percentages [3].

Example of a hidden problem

Two formulations can both calculate to AFS 72:

• formulation A: most grains are retained on adjacent sieves;
• formulation B: a coarse fraction is blended with a very fine fraction.

Formulation B may have the correct average while also showing:

• more dust;
• a higher specific surface;
• different permeability;
• different binder demand;
• segregation during transport.

Sieve-by-sieve control avoids confusing a compliant average with a controlled distribution.

Expected effect on permeability

Permeability describes the ability of a mould or core to allow gases to pass.

As fineness increases, the channels between grains can become smaller and more tortuous. The general tendency is therefore towards lower permeability, especially when the distribution contains more fines or compacts strongly.

The Kansas Geological Survey report already relates grain size and distribution to density and permeability [3].

A 2026 experimental study observed lower permeability in its reclaimed-sand system in association with several combined factors:

• higher GFN;
• more irregular grains;
• contaminants;
• residual clay;
• moisture;
• a broader distribution [4].

We therefore do not isolate AFS GFN as the only cause.

Useful controls

After changing the grading, we examine:

• permeability of the prepared mixture;
• moisture;
• compactability;
• mould hardness;
• green strength or core strength;
• gas release;
• blowholes or porosity;
• metal penetration;
• surface finish;
• lot-to-lot stability.

Effect on strength and binder demand

A finer sand generally presents a larger total surface to be coated for the same mass.

Research from the University of Miskolc compared new and reclaimed sands using:

• particle-size analysis;
• specific surface;
• loss on ignition;
• scanning electron microscopy;
• X-ray diffraction;
• bending strength;
• thermal distortion [6].

This approach confirms that core quality is not controlled by a single index.

When AFS GFN is increased, we monitor:

• binder quantity required to coat the grains;
• binder distribution;
• mulling time;
• flow through the tooling;
• initial strength;
• final strength;
• shakeout;
• emissions and residues associated with the selected system.

In green sand, Karunakar and Datta studied GFN, clay, moisture, mulling time and hardness together, then measured strength and permeability [5].

This reinforces our method:

When grading changes, binder formulation and preparation parameters must be examined with it.

Grain shape and moulding behaviour

Two sands with the same distribution can still behave differently when their grains are:

• rounded;
• subrounded;
• subangular;
• angular;
• cracked;
• coated with residues.

Shape influences:

• flowability;
• packing;
• number of contacts;
• actual surface area;
• binder consumption;
• strength;
• permeability;
• reclamation behaviour.

Grinding used only to increase AFS GFN may create new edges and dust.

The study by Dai et al. shows that mechanical and thermo-mechanical loading can strongly increase the AFS fineness of certain sands through grain fracture [7].

We therefore distinguish between two strategies.

Blending characterised fractions

Advantages:

• predictable mass balance;
• known particle-size curves;
• limited change in morphology;
• rapid ratio adjustment;
• better traceability.

Particle-size reduction by grinding

Points to control:

• generation of fines;
• change in shape;
• heat;
• equipment wear;
• dust;
• width of the distribution;
• need for classification after grinding;
• production stability.

For a defined target such as 49 to 72, controlled blending of screened fractions generally provides the clearest method.

Minerallum’s proposed laboratory procedure

Step 1 - Obtain representative samples

Sample separately:

• the AFS 49 sand;
• the candidate fine sand;
• several points when the material is in bulk;
• several bags when it is packaged.

Record:

• product;
• batch;
• date;
• origin;
• apparent moisture;
• sampling location.

Step 2 - Work on a dry basis

Measure moisture when the materials are not fully dry.

Blend proportions are expressed on a dry-mass basis so that differences in moisture do not change the actual mass of grains being introduced.

Step 3 - Sieve both components

Use the same sieve series, duration and procedure.

Record:

• initial mass;
• mass on every sieve;
• pan mass;
• total recovered mass;
• retained percentage;
• cumulative passing or retained percentage;
• calculated AFS GFN.

Step 4 - Calculate several scenarios

Prepare at least:

• the theoretical target scenario;
• a slightly lower scenario;
• a slightly higher scenario.

For a target of 72 using AFS 100 fine sand, the following can be compared:

AFS 49 sand	AFS 100 sand	Theoretical AFS
60%	40%	69.4
54.9%	45.1%	72.0
50%	50%	74.5

This small matrix shows the sensitivity of the formulation to dosing errors.

Step 5 - Prepare a pilot batch

The batch should be large enough to:

• limit weighing error;
• permit proper homogenisation;
• supply the sieve analysis;
• prepare test specimens;
• retain a reference sample.

Step 6 - Repeat the sieve analysis

Compare:

• theoretical AFS;
• measured AFS;
• theoretical distribution;
• measured distribution;
• recovered mass;
• fine fraction;
• repeatability.

A difference may result from:

• poor homogenisation;
• weighing error;
• moisture;
• segregation;
• sampling differences;
• attrition during preparation.

Step 7 - Prepare the moulding system

Then add, depending on the process:

• bentonite;
• water;
• sea coal or another additive;
• resin;
• catalyst;
• silicate;
• hardener;
• coating.

Dosages previously used with AFS 49 should not automatically be transferred to AFS 72.

Step 8 - Measure functional properties

The test plan depends on the process.

System	Priority controls
Green sand	Moisture, compactability, permeability, green strength, hardness, active clay
No-bake resin	ADV or pH as required by the system, work time, strength, loss on ignition, shakeout
Cold-box / hot-box	Flowability, binder dosage, initial and final strength, gas, distortion
Silicate	Strength, curing, moisture, shakeout, reclamation
Facing sand	Fineness, surface strength, local permeability, penetration defects

Step 9 - Produce a trial casting

Compare with the reference system:

• mould filling;
• mould integrity;
• gas release;
• surface finish;
• penetration;
• sand inclusion;
• blowholes;
• dimensions;
• shakeout;
• binder consumption;
• rejection rate.

Tolerances and production control

An industrial formulation should not merely reach 72 once. It should remain stable.

We recommend defining:

• target AFS;
• acceptance range;
• maximum percentage on each critical sieve;
• fine-fraction limit;
• minimum recovered mass;
• sampling method;
• test frequency;
• dosage-adjustment rules;
• batch identification;
• retention of results.

Sensitivity to dosage

For the AFS 49 + AFS 100 blend:

• each percentage point of added fine sand changes the theoretical AFS by approximately 0.51 point;
• a two-point dosing error moves the target by approximately 1.02 AFS points.

This relationship supports decisions about:

• balance accuracy;
• dosing equipment;
• minimum batch size;
• verification frequency.

Segregation during storage

A large difference in particle size between the two components can promote separation during:

• discharge into a silo;
• conveying;
• vibration;
• big-bag filling;
• transport;
• screw or belt reclaim.

Sampling should therefore be performed on the stream or according to a representative plan, rather than only from the surface of the stock.

Purchasing specification for the fine sand

To adjust an AFS 49 sand consistently, the order for the fine sand should go beyond a description such as “fine sand”.

We structure the request around the following elements.

Identification

• foundry silica sand;
• use: moulding, coremaking, facing or grading correction;
• origin and preparation process;
• batch and traceability.

Particle-size distribution

• sieve series;
• percentage retained on each sieve;
• target AFS GFN and tolerance;
• AGS where requested;
• fraction below the minimum sieve;
• pan fraction;
• lot-to-lot repeatability.

Physical properties

• grain shape;
• bulk density;
• moisture;
• specific surface where useful;
• flowability;
• attrition resistance when the sand is recycled.

Chemical characteristics

Depending on the metal and binder system:

• SiO₂;
• Al₂O₃;
• Fe₂O₃;
• CaO;
• MgO;
• alkalis;
• loss on ignition;
• pH;
• ADV;
• foreign matter.

Logistics

• quantity;
• bags, big bags or bulk;
• unit mass;
• moisture protection;
• loading conditions;
• destination;
• schedule;
• preliminary sample.

Silica dust during preparation and shakeout

Adding a finer fraction increases the attention given to dust emissions during dry operations.

NIOSH studied the cleaning of castings produced in sand moulds and measured the effectiveness of a local exhaust ventilation system. The described system reduced exposure to respirable dust by 59% to 77% across the tools evaluated [8].

Within an industrial operation, relevant points include:

• bag emptying;
• dosing;
• mixing;
• transfers;
• sieving;
• conveyor drops;
• shakeout;
• shot blasting;
• grinding;
• cleaning;
• sweeping.

Technical measures can combine:

• capture at source;
• enclosure;
• closed transfer;
• extraction;
• wet methods where compatible;
• industrial vacuum cleaning;
• filter maintenance;
• control of leakage points.

This control also supports product stability by limiting uncontrolled losses of fine fractions.

Frequent errors when moving from 49 to 72

1. Adding AFS 72 sand to AFS 49 sand

The blend remains below 72 while it still contains AFS 49 material.

2. Calculating on wet mass

Two products with different moisture contents do not provide the same dry mass of grains.

3. Using only the average AFS value

The average can hide a bimodal distribution or excessive fines.

4. Combining analyses produced with different sieve series

The factors and fractions must be comparable.

5. Changing sand, binder and moisture at the same time without a test plan

It becomes difficult to identify the cause of the observed changes.

6. Reusing the binder dosage from AFS 49

The surface to be coated and the packing structure have changed.

7. Neglecting homogenisation

A local sample may meet the target while the complete batch remains heterogeneous.

8. Ignoring segregation

Coarse and fine fractions can separate during handling.

9. Accepting a large pan fraction

AFS GFN may increase while dust becomes excessive.

10. Approving only one laboratory specimen

Filling, compaction, venting and contact with molten metal should be observed in a production trial.

Information to send for a Minerallum study

To review an AFS target, we use the following information.

Current sand

• measured AFS;
• sieve table;
• moisture;
• grain shape;
• available composition;
• origin and batch.

Candidate fine sand

• measured AFS;
• sieve table;
• fine fraction;
• moisture;
• shape;
• packaging;
• monthly availability.

Process

• green sand, resin, silicate or another system;
• mould or core;
• metal and pouring temperature;
• casting mass and geometry;
• compaction system;
• reclamation where applicable.

Objectives

• target AFS;
• permeability;
• strength;
• surface finish;
• defect to reduce;
• binder consumption;
• production rate;
• rejection rate.

Logistics

• tonnage per batch;
• monthly tonnage;
• dosing equipment;
• storage;
• destination;
• packaging.

Conclusion: target AFS 72 with a controlled distribution

Our analysis leads to a straightforward and rigorous method.

To move from AFS 49 to AFS 72:

1. select a fine sand with an AFS higher than 72;
2. calculate the proportion on a dry-mass basis;
3. predict the sieve-by-sieve distribution;
4. prepare a pilot batch;
5. measure the actual AFS;
6. control fines, permeability and strength;
7. adjust binder and preparation parameters;
8. confirm the system through a trial casting;
9. establish production tolerances.

With AFS 100 fine sand, the initial calculation is:

54.9% AFS 49 sand + 45.1% AFS 100 sand

This value answers the mass balance. Industrial quality then comes from the combined control of:

• particle-size distribution;
• grain shape;
• fines;
• binder;
• moisture;
• compaction;
• permeability;
• strength;
• casting process.

AFS 72 is therefore not merely a number to reach. It becomes a documented grading specification integrated into the foundry’s actual operating system.