How to Read Pull-Out Test Results — Load–Movement Curves from 23 Real Test Points

The article What Is a Pull-Out Test? already covers the principle, equipment and step-loading procedure. This article picks up from there: once the report is in your hands, how do you read the tables and curves? It uses real data from 23 points tested on a smooth concrete surface, from DB16 deformed-bar tests SPN carried out in late 2025 at a pharmaceutical plant in Phan Thong District, Chon Buri.

What Decides Whether a Pull-Out Test Passes or Fails?

The acceptance criterion for a proof-load Pull-Out Test is that the anchor carries the full test load, holds it for the specified time, and shows no failure mode of any kind. ASTM E488 lists five failure modes: concrete cone breakout, concrete cracking followed by anchor pull-out, anchor pull-out, bond failure, and fracture of an anchorage component.

A common misunderstanding is that bond failure only counts if you see the bar come out. The standard defines continuing movement under constant or decreasing load as failure. On site, the signs are a dial-gauge needle that keeps creeping during the hold, or a pressure gauge that drops by itself even though the valve has not been released.

This job used the project specification's criterion: hold the test load for at least 3 minutes with no observable movement, e.g. the washer under the nut does not loosen. As for the "magnitude" of total movement, ASTM E488 sets no fixed number. If the designer wants to limit movement, the limit must be written into the specification before testing.

During the hold at test load, the site supervisor can therefore check three things directly: the dial-gauge needle, the pump's pressure gauge, and the concrete surface around the jack base, as shown in the model below.

← Scroll to see the full image →

Where Does the Test Load Come From? Example: DB16 Bar, 39 kN Design Load

The test load is the design load multiplied by a factor set by the designer or the project specification. This job tested DB16 deformed bars embedded 26 cm in concrete, with a design load of 39 kN and a factor of 1.25.

When testing post-installed rebar, also check that the test load is comfortably below the bar's yield load, because the aim is to test bond in the concrete, not to pull the bar to yield. Example: DB16 has a cross-sectional area of 2.01 cm². If it is SD40 steel (fy = 4,000 kg/cm²), the yield load is about 8,040 kg, or roughly 78.9 kN, so the 48.75 kN test load is about 62% of yield. Confirm the actual steel grade from each project's construction drawings.

Another point visible on the data sheets is that pressure-gauge readings depend on the specific jack. This job used two hydraulic jacks: one read 145 bar and the other 107 bar at the same load of 4.97 t, because their piston areas differ. The report must therefore state the jack and gauge numbers used at each point and attach the calibration certificates that match those numbers.

Real Results from 23 Points: All Passed, Movement 1.04–3.76 mm

Every point carried the full test load, held it, and showed no concrete cracking. Movement at maximum load ranged from 1.04 to 3.76 mm. Even though all points passed, the highest value is about 3.6 times the lowest. That difference is what the PASS column does not tell you, and it is why the individual curves must be read as well.

← Scroll to see the full table →

Summary of 23 points at the 48.75 kN test load (smooth concrete surface)

Item	Value
Number of points	23
Minimum movement	1.04 mm
Median	1.79 mm
Maximum movement	3.76 mm
Result	PASS at all points

Reading the Load–Movement Curve: 4 Patterns Seen in Real Jobs

The load–movement curve plots pull-out load (vertical axis) against the movement of the bar end read from the dial gauge at the end of each 2-minute step (horizontal axis). The shape of the curve tells you more than the final number. The chart below plots point POT-24 from the real data sheet against the range of all 23 points, and the bottom strip summarizes four curve patterns worth knowing.

← Scroll to see the full image →

1. Steep, Uniform Straight Line — The Expected Shape

A steep straight-line curve means movement increases by roughly equal, small amounts at every step. For example, POT-24 moved 0.18–0.27 mm per step, totaling 1.67 mm at maximum load. This pattern indicates the anchorage is still working in its elastic range and the test setup is stable. All 23 points on this job moved no more than 3.76 mm at the test load.

2. Steep at First, Then Softening at the End — Check the Hold Period

A curve that softens at the end means the movement per step keeps growing as the load rises. At low loads the line is still steep, but over the last 2–3 steps the movement per step increases several-fold. This pattern shows that the bond is starting to work beyond its elastic range.

The deciding factor is the constant-load hold. If movement stops within the hold time, the anchor is still carrying the load. But if the needle keeps moving or the pressure drops, that is bond failure as defined by ASTM E488. Even when a point with this kind of curve passes, the site supervisor should review its installation, such as hole cleanliness and embedment depth.

3. Early Jump, Then Steep — Usually Seating of the Test Setup

An early-jump curve shows a large chunk of movement in the first few steps, then becomes steep again in the following steps. If the bar itself were slipping, the curve would not return to a steep slope after the jump. This pattern therefore usually comes from components of the test setup that were not yet fully seated, such as a bearing plate or packing under the jack base compressing, or free play in threads and nuts.

ASTM E488 handles this in two ways: apply a seating load of no more than 5% of the maximum load so that all components are fully seated before zeroing, and correct the displacement by extrapolating the curve through the data points back to zero load, then measuring movement from that new zero point (corrected displacement).

4. Uniform Flat Slope from the First Step — Suspect the Measurement Before the Bar

A uniformly flat curve shows large, equal movements at every step starting from low load. A system that is "soft" from as little as 15% load, with no steep straight-line portion first, is not typical bond-slip behavior, which usually becomes evident only at high load. Check the measurement first, for example settlement of the jack base or a dial-gauge stand moving with the loaded area. If the point still holds the load and shows no cracks, it can still pass, but the report should note that the movement values may be affected by the condition of the test setup.

What Makes Up the Measured Movement?

The measured value is the sum of four movement components, not bar slip alone. The dial gauge does not measure "bar slip" directly; it measures the movement of the steel plate on top of the jack, which rises together with the bar grip, relative to the stand's reference point. If the jack base is not flush with the concrete, the gap closes under load, and if the dial-gauge stand is placed close to the area under compression, the reference point moves too. The measured value therefore includes movement of the test setup itself.

What Does the Measured Movement Consist Of?

• A. Elastic elongation of the bar length above the concrete — can be calculated from PL/AE (see box below)
• B. Slip of the bar relative to the concrete — the part that reflects the actual bond quality
• C. Seating of components — bearing plates, nuts, threads and couplers; usually occurs in the first few steps
• D. Settlement of the bearing base and movement of the reference point — depends on how flat the setup area is and where the stand is placed

The section below shows where each of the four components arises in the type of setup used on this job. The dial-gauge stem touches the steel plate on top of the jack, which rises with the bar grip, so the measured value is always A + B + C + D. Only component B reflects the actual bond quality.

← Scroll to see the full image →

ASTM E488 reduces the error from tilt of the test setup by measuring at two positions and averaging: ΔT = ½(A_N − A_I + B_N − B_I). Where site conditions are uneven, using two dial gauges and mounting the stands on a reference beam supported outside the loaded area helps the movement values reflect the bar's behavior more accurately.

7-Point Checklist Before Accepting a Pull-Out Test Report

1. Test load and its basis — clearly states the design load, the factor and the final test load, e.g. 39 × 1.25 = 48.75 kN
2. Equipment numbers match the calibration certificates — the jack and gauge numbers on each point's data sheet must match the attached calibration certificates
3. Step-by-step tables and curves for every point — load, gauge reading and movement at every step, not just a summary table saying PASS
4. Load steps and hold times — no more than 15% per step, each step held 2 minutes, and the final load held for the time stated in the project specification
5. Damage pattern at each point — states that there were no cracks, or describes the ASTM E488 failure mode if one occurred
6. Photos and locations — photos of every point with its ID sign, and point IDs consistent across the summary table, data sheets and photos
7. Site-condition notes — if any condition affected the measurement, such as an uneven concrete surface or the need to pack under the jack base, the report should record it so the designer can read the numbers in the right context

Summary: How Do You Read Pull-Out Test Results Completely?

Reading Pull-Out Test results takes two steps: first, decide pass/fail against the criteria; second, read the load–movement curve to assess quality.

• Acceptance criterion — full load carried, held, and no failure
• Steep straight line — the expected shape
• Softening at the end — check the hold period
• Early jump — usually seating of the test setup
• Flat from low load — check the measurement first

All 23 real points on this smooth concrete surface passed, with movements of 1.04–3.76 mm. Every measured value is the sum of bar elongation, bar slip and movement of the test setup, so it must always be read together with the shape of the curve.

If you are planning to test chemical anchors, threaded studs or post-installed rebar, see the service details at Non-Destructive Structural Testing (NDE), or the factors that affect cost in Structural Testing (NDT) Prices 2026. Pull-Out Test fees depend on the number of points, the test load and site conditions. Contact us for a free quotation within 24 hours. Prices exclude travel and equipment mobilization, which are charged additionally according to the distance to site.