Nuclear Gauge Failures Backed by a One-Point Proctor
You are the CMT technician on a four-story steel-framed building with a slab-on-grade. Structural fill beneath the slab is placed in 12-in. loose lifts and must reach 95% of maximum dry density per ASTM D698 at moisture from 1% below to 2% above optimum. At Lift 4, between elevations 3 and 4 ft, the contractor's QC technician runs an ASTM D6938 nuclear gauge at eight locations and reports three results of 93.1%, 92.4%, and 94.0%. The gauge moisture reads about 2.8% above optimum on minus-3/4-in. material. The technician defends the results with a one-point Proctor: a single laboratory point of 118.9 pcf dry density at 11.2% water content, compared against a maximum of 124.8 pcf from a curve developed three weeks ago for a different borrow area. The borrow now runs roughly 15% gravel. The contractor wants to place the next lift tonight to keep the crane schedule.
1. What verification do you require before the next lift goes in, and in what sequence?
Read the worked answer
The gauge result is not yet an acceptance result. Sequence: (1) hold the lift; (2) run a standard count and confirm the daily reference count is within the manufacturer's tolerance, because a drifted gauge biases every reading; (3) run ASTM D1556 sand cone at the three failing locations plus one passing location, with a hole large enough to capture the gravel; (4) take split moisture samples and run ASTM D2216 oven drying, not the gauge; (5) obtain the actual D698 report you are being held to and confirm which borrow and gradation it represents.
Reason: D6938 is an indirect, comparative method and must be correlated to direct density on the same material. With 15% gravel and no rock correction applied, the gauge can be biased in either direction, so the failing numbers prove nothing until a sand cone confirms them. The sand cone is the referee test.
2. Is the one-point Proctor defensible as the basis for acceptance here?
Read the worked answer
No, as presented it is not defensible. A one-point Proctor is a shortcut valid only when a full curve or an established family of curves already exists for that soil, the material is unchanged, and the single point plots in the accepted relationship to the curve (typically near the peak, within a defined offset).
None of those conditions is met. The 124.8 pcf reference curve came from a different borrow, and the borrow has since moved to about 15% gravel, which raises maximum dry density and usually lowers optimum moisture. A point at 118.9 pcf and 11.2% sits 5.9 pcf below the assumed maximum, which is not a near-peak point at all. Because 124.8 pcf is the denominator of every relative compaction value on this lift, the whole lift's numbers are unverified, not just the three failures.
3. Write the disposition and list what goes into the report.
Read the worked answer
Disposition: reject Lift 4 as a nonconformance and do not allow Lift 5 over it. Require a full compaction test on a representative sample of the current borrow, using the method the specification designates (confirm whether D698 or D1557 is specified; switching methods is not a field decision), plus a gradation.
Retest by sand cone at the three failing locations and at the five reported passing locations, since they share the same invalid denominator. If confirmed relative compaction is below 95%, require scarifying the lift, moisture conditioning into the specified window, recompaction, and retesting.
Report contents: gauge serial number and standard count, sand cone data with hole dimensions, oven moisture results, the Proctor curve with sample identification, technician and date, and a written nonconformance to the engineer of record. Waiving the acceptance criterion is not the technician's call.