How the two differ in kind
A standard penetration test gives a number, and the number is routinely used as though it described the soil. It does not, on its own.
The RAW N-VALUE is the blow count to drive a standard sampler a standard distance, and it depends on how confined the soil is at the point of the test — which depends on the depth. The same granular soil at the same actual density gives a higher blow count at twelve metres than at three, because there is more overburden pressing on it and the sampler has to displace soil against that pressure. So an uncorrected log shows N rising with depth in a uniform deposit, and reading that as increasing density is a misinterpretation of the test rather than an observation about the ground.
The OVERBURDEN CORRECTION normalises the count to a reference stress, so that values from different depths can be compared and so that published correlations — which were derived against corrected values — can be used. There is a second correction that matters as much and is less discussed: the ENERGY actually delivered by the particular hammer. The test's blow count assumes a standard energy transfer, and real rigs vary substantially depending on the hammer type and its condition, so a count from one rig is not directly comparable with one from another until both are normalised to a common energy ratio.
RELATIVE DENSITY is the property all of this is aiming at: where the soil sits between its loosest and its densest achievable states. That is what governs a granular soil's strength, its stiffness, its settlement under load, and its susceptibility to liquefaction — none of which follow from a raw count.
The practical consequence is that a foundation designed from uncorrected N-values in a deep deposit is designed on the assumption that the soil is denser than it is.
The factors that actually differ
Show
| Corrected SPT N-value | Relative density | |
|---|---|---|
| What it is | A field blow count, corrected for overburden stress and for the hammer's delivered energy. | Where the soil sits between its loosest and densest achievable states. |
| Depends on depth | The raw count does, strongly — which is exactly what the correction removes. | No. It is a property of the soil's packing. |
| Comparable between boreholes | Only after correction, and only if both rigs' energy is accounted for. | Yes, being a normalised property. |
| What it governs | Nothing directly — it is an input to correlations. | Strength, stiffness, settlement and liquefaction susceptibility. |
| Energy correction | Essential and frequently omitted. Hammer type and condition change the delivered energy substantially. | Not applicable. |
| Where it comes from | A field test, cheap and universally available, which is why it persists. | Derived from the corrected count via published correlations, or measured in a laboratory on recovered samples. |
| Soils it suits | Granular soils. In clays the count is indicative and other tests are preferred. | Granular soils — the concept of loosest and densest states applies to them. |
| What distorts it | Gravel and cobbles, which inflate the count; a disturbed base; groundwater inflow loosening the soil before the test. | Inherits whatever distorted the count it was derived from. |
| The misinterpretation | Reading a rising raw count with depth as increasing density. | Assuming a density from a raw count without correcting it. |
| Which to design from | The corrected value, which is what correlations and codes are written against. | The property, where the design method calls for it directly. |
Which one, and when
Choose corrected spt n-value when…
- Interpreting a borehole log, where the raw counts need correcting before they mean anything.
- Comparing values from different depths in the same deposit, or between boreholes.
- Using published correlations, which were derived against corrected values.
- Liquefaction assessment, where the corrected, energy-normalised value is the standard input.
Choose relative density when…
- Assessing a granular soil's strength, stiffness or settlement behaviour.
- Compaction acceptance on granular fill, where the density achieved against the achievable range is the criterion.
- Where the design method calls for relative density directly.
- Comparing the condition of a deposit against what compaction could achieve in it.
Now run your own numbers
This page holds no prices on purpose — a national average is wrong for almost every real project. Quantify both options with your dimensions and your local quotes.
Frequently asked questions
- Why does the blow count rise with depth in uniform soil?
- Because the sampler has to displace soil against the confining stress, and that stress rises with depth. Driving the sampler requires pushing the grains apart locally, and the deeper the test, the more overburden is pressing down and the more resistance that displacement meets — independently of how tightly the grains are packed. So a uniform deposit of constant density produces a log in which N rises steadily with depth, and reading that as the soil getting denser confuses the test's behaviour with the ground's. The overburden correction removes it by normalising every value to a reference stress, which makes the corrected profile flat in a uniform deposit and makes genuine changes in density visible as changes in the corrected value rather than being buried in the depth trend.
- What is the energy correction and why does it matter?
- The test defines a hammer mass and drop height, but the energy actually transferred into the rods varies substantially with the hammer mechanism, its condition and how it is operated — so two rigs testing the same soil can produce meaningfully different counts. Corrections normalise the count to a standard energy ratio so that values are comparable between rigs, between contractors and against the published correlations, which were themselves derived with a particular energy characteristic. Modern practice is to measure the delivered energy on the rig rather than to assume a value for the hammer type, and reputable ground investigation reports state the energy ratio used. Where a report gives raw counts with no statement of energy, the values are indicative rather than comparable, which is worth knowing before designing from them.
- What does relative density actually tell you?
- Where a granular soil sits between the loosest and densest packings it can achieve, and that position governs almost everything about its behaviour. A dense sand is strong, stiff, settles little, and dilates as it shears — it tries to expand, which increases its strength. A loose sand of exactly the same grains is weak, compressible, and contracts as it shears, which under rapid loading can generate pore pressure and lead to liquefaction. So two deposits identical in grading and mineralogy behave completely differently depending on their relative density, which is why it is the parameter the correlations aim at and why compaction specifications for granular fill are written in terms of achieving a density within the achievable range.
- What makes an SPT unreliable?
- Several things, and the reports do not always say. Gravel and cobbles inflate the count, because the sampler hits a particle rather than penetrating soil — so a high count in a gravelly stratum can mean an obstruction rather than dense ground. A poorly cleaned borehole leaves debris the sampler drives through first. Groundwater inflow into the base of a borehole can loosen the soil before the test, giving an artificially low count in exactly the granular soils the test is meant for, which is why maintaining the water level in the borehole matters. Rod length, borehole diameter and sampler type all have their own corrections. None of this makes the test useless — it is cheap, universal and well correlated — but it does mean interpreting a single value rather than a profile is unwise.
- Does the SPT work in clay?
- Indicatively, and it is not the test of choice. The standard penetration test was developed for and correlates best in granular soils, where it relates to relative density and friction angle. In clays the count relates loosely to undrained shear strength and the correlations are much weaker, partly because the test is a dynamic disturbance in a material whose strength depends on its structure and its stress history. Better options exist and are usually specified where clays govern: undisturbed sampling with laboratory testing, the cone penetrometer test which gives a continuous profile, or a vane test in soft clays. An SPT profile in clay is useful as a stratification tool and as a rough indication rather than as the basis for a design parameter.
- How does this relate to liquefaction assessment?
- Directly, because the corrected, energy-normalised value is the standard input to it. Liquefaction is the loss of strength in a saturated granular soil under cyclic loading, and the susceptibility depends on how loose the soil is — which is what the corrected count estimates. Assessment methods compare a cyclic resistance derived from the corrected value against the cyclic demand from the design earthquake, and they are explicitly calibrated against counts corrected for overburden and normalised to a stated energy ratio, with further corrections for fines content. Using a raw count in that framework overstates the resistance, and it overstates it most at depth where the overburden effect is largest — which is exactly where a liquefiable layer is most likely to be missed.
- How many tests are enough?
- Enough to see a PROFILE rather than a point, which is the practical answer and the reason single values are so misleading. A borehole with tests at regular intervals shows how the corrected value varies with depth, which identifies the strata, their boundaries and any weak layer — and a weak layer a metre thick beneath a competent one is exactly what governs a foundation and exactly what a sparse test schedule misses. Across a site, enough boreholes to establish how the profile varies laterally matters as much, since ground rarely repeats. What decides both is the structure and the consequence of getting it wrong, which is why investigation scope is proportionate to the project rather than standard — and why the cost of an extra borehole is trivial against the cost of a foundation redesigned during construction.
- Is the CPT better?
- For many purposes yes, and it is increasingly the default where the ground suits it. A cone penetration test pushes an instrumented cone at a constant rate and records tip resistance, sleeve friction and often pore pressure continuously — so it produces a profile at millimetre resolution rather than values every metre or two, it identifies thin layers an SPT would step over entirely, and it is far more repeatable because there is no hammer energy to vary. Its limitations are that it cannot penetrate dense gravel, cobbles or rock, and that it recovers no sample, so where material has to be seen or tested in a laboratory a borehole is still needed. In practice the two are frequently used together, with CPT profiling between boreholes that provide the samples.