How Is Soil Compaction Tested?
A compaction specification is easy to write and hard to prove. A set of plans will say something like "fill shall be compacted to 95 percent of maximum dry density," and from that moment on the real question is not whether the roller ran, but whether the soil actually got dense enough. Soil compaction testing is how that gets answered. A laboratory test establishes the target density for the material, a field test measures what is actually in the ground, and the two numbers are compared.
Here is how the process works, what each test method measures, how often testing is required, and what happens when a lift fails.
What Percent Compaction Really Means
Nearly every earthwork specification is written as a percentage, and that percentage is always a comparison. Percent compaction is the in-place dry density of the soil divided by the maximum dry density of the same material, multiplied by 100. If the lab says a soil's maximum dry density is 122 pounds per cubic foot (pcf) and the field test reads 118 pcf, that lift is at roughly 96.7 percent.
Notice that "dry" density is doing the work here. Soil is weighed wet in the field, so moisture content has to be measured and removed from the calculation before the comparison means anything. A compaction test is really two measurements — density and moisture — reported as a single number.
Typical requirements run from about 90 to 95 percent of standard Proctor density for non-structural fill and embankments, up to 95 to 98 percent or more of modified Proctor for anything that will carry pavement or heavy loads. The specification always states which test method the percentage is based on, and that detail matters more than the number itself. For background on how density and moisture interact in the field, see our explanation of the science of soil compaction.
Step One: The Lab Proctor Test Sets the Target
Before anyone can test the field, someone has to establish what "maximum dry density" means for this particular soil. That is the Proctor test, formally the Standard Proctor (ASTM D698 / AASHTO T 99) or the Modified Proctor (ASTM D1557 / AASHTO T 180).
Both tests work the same way. A sample of the soil is dried, split into several portions, and each portion is mixed with a different amount of water, then compacted into a mold with a specified hammer and drop height. The standard test uses three lifts, 25 blows per lift, and a 5.5-pound hammer falling 12 inches, for a compactive effort of 12,375 ft-lbf per cubic foot. The modified test uses five lifts with a 10-pound hammer falling 18 inches, for 56,250 ft-lbf per cubic foot — roughly four and a half times the energy.
Each compacted portion is weighed and its moisture measured, and the resulting dry densities plot into a curve. The peak of that curve is the maximum dry density (MDD). The moisture content at that peak is the optimum moisture content (OMC). Because modified Proctor applies more energy, it produces a higher MDD and a lower OMC on the same soil.
That single laboratory result becomes the benchmark every field test is judged against, which is why it matters that the lab sample came from the same borrow source the crew is actually placing. Most specifications require a new Proctor whenever the fill material changes, and a common convention is one Proctor per 1,000 to 5,000 cubic yards of fill. When the source or the material changes, see our comparison of select fill and native soil for why the numbers move.
Step Two: Field Density Testing
Out on the job, three methods do almost all of the work. All three measure the dry density of the soil in place, and all three are compared against the Proctor result to produce percent compaction.
Nuclear density gauge
The nuclear gauge is the standard tool on construction sites. It measures density by passing gamma radiation between a source in a probe and detectors in the base of the instrument, and it measures moisture with a separate source. In "direct transmission" mode, a steel rod is driven to create a pilot hole, and the probe is lowered up to about 12 inches into it; readings take about a minute. In "backscatter" mode no pilot hole is needed, but the results are considered less reliable. The governing standard is ASTM D6938 / AASHTO T 310.
The trade-offs are administrative as much as technical. The gauge is fast and handles many tests per day, but it holds a licensed radioactive source, so storage, transport, use, and personnel dosimetry all fall under regulatory requirements, and operators need training and certification. The gauge also has to be standardized (a standard count) before use each day.
Sand cone
The sand cone method (ASTM D1556 / AASHTO T 191) is the classic, non-nuclear approach and is still widely accepted by state DOTs. A base plate with a known opening is placed on the prepared surface, a hole of known geometry is excavated through the plate, and all of the excavated soil is collected and weighed. The volume of the hole is then determined by filling it with free-flowing sand of known density. Wet density is the wet weight divided by that volume, and dry density follows once moisture is known.
The method is accurate and requires no radioactive materials, but it takes 30 minutes or more per test, it needs the excavated material to be recovered completely, and it should not be used on saturated or highly plastic soils or where significant amounts of coarse rock are present.
Rubber balloon
The rubber balloon method (ASTM D2167) works like the sand cone but replaces sand with a calibrated water vessel that presses a rubber membrane into the excavated hole. Volume is read directly from graduations on the vessel, and the water is retained for the next test. It is faster than the sand cone, but the membrane can puncture and the method is intended for fine-grained or granular soils without much coarse material.
Measuring Moisture Content
Whichever density method is used, moisture has to be determined to convert wet density into dry density. A nuclear gauge reports moisture directly. For sand cone and balloon tests, moisture is measured separately, using the oven method (ASTM D2216) as the reference, or faster field alternatives — the calcium carbide gas pressure tester (ASTM D4944), the microwave method (ASTM D4643), or direct heating (ASTM D4959).
This step is where seemingly contradictory results often come from. Two tests on the same lift can read very different percent compaction if the moisture samples were taken from material that had dried unevenly, which is one more reason to read the moisture number alongside the density on every report. If the moisture has drifted high, our guide to how wet is too wet to compact soil walks through the field signs and the fix.
How Often Tests Are Required
Test frequency is set by the project specification or the geotechnical engineer, but the usual conventions are fairly consistent. At a minimum, one test per lift in each distinct fill area, plus additional tests whenever the soil type or moisture changes. In areas that matter most — under foundations, slabs, retaining walls, and pavement — the frequency is typically tightened to one test per 1,000 to 2,500 square feet per lift.
For structural fill placed in thicker sections, a common requirement is one density test per two feet of fill height per 2,500 to 5,000 square feet of area. Footing and utility trenches are often tested on a linear basis, frequently one test per 50 linear feet of trench with a minimum number of tests in each area. Some specifications also reduce frequency after a run of passing tests and revert to the original frequency after any failure.
Separately, a "one-point" field check (AASHTO T 272) is used to confirm that the material being placed still matches the laboratory sample the Proctor was run on. It is a material-identity check, not a substitute for a density test.
What Makes a Test Fail, and What Happens Next
A failing test almost always traces back to one of a handful of causes: the lift was too thick for the compactor to reach the bottom of it, the soil was above optimum moisture and pumping under the roller, the compactor was the wrong type or too small for the material, or the crew simply did not make enough passes.
Two causes deserve special mention because they are easy to miss. Organic material — roots, topsoil, and other decomposable matter — will not hold density no matter how many passes you make, which is why stripping and stockpiling topsoil properly at the start pays off later. And frozen soil cannot be compacted at all; ice blocks particle rearrangement, and the resulting density reading is not meaningful. Our guide to cold weather earthwork covers what to do when the temperature drops.
When a lift fails, the response is usually one of four things: make additional passes with the same machine, aerate and dry the material before rolling again, thin the lift and re-compact, or remove and replace the material. When a soil is persistently wet and weak, engineered stabilization with lime or cement may be the practical answer, which is what our article on soil stabilization for unstable subgrade addresses.
Reading a Compaction Test Report
A useful report is more than a pass or fail stamp. Look for the field dry density and moisture content, the percent compaction relative to the specified Proctor value, the test location described by station and elevation or by grid coordinates, the lift number and depth, and the test method and standard used. The report should also identify the technician and the certifying laboratory, and note that the gauge was standardized before use.
That location data is what lets an engineer decide whether a failure is isolated or systemic. A single low test in one corner of a large fill area means something very different from three failures in a row along a single trench.
Why a Field Test Is Only Half the Story
A density test verifies a spot roughly the size of a dinner plate. It is evidence that one small area meets the specification, not proof that the whole lift does. That is why compaction testing is normally paired with other checks: visual observation during placement, careful control of lift thickness and moisture, and in many cases proof rolling to expose soft spots that a gauge reading can miss.
For contractors, the practical lesson is that choosing the right compaction equipment and controlling lift thickness from the first load are more effective than trying to test your way out of a marginal lift afterward. Density testing documents quality; it does not create it. The jobsite handoff checklist explains how the test records fit into the closeout package a general contractor or owner will expect.
Get those fundamentals right, and the test results tend to follow: uniform lifts, moisture near optimum, enough passes, and a lab benchmark that matches the material in the ground. When any one of those slips, the density test is simply the instrument that tells you so — before the slab is poured and the mistake gets expensive.
