Skip to contentUnited States Department of Transportation - Federal Highway AdministrationSearch FHWAFeedback
Highways for LIFE

Arrow Asphalt Binder Cracking Device to Reduce Low-Temperature Asphalt Pavement Cracking: Final Report

Asphalt Binder Cracking Device Interlaboratory Study

For the ABCD ILS, each laboratory received instructions and the ABCD test procedure to review before receiving the ABCD unit and binder samples. After a few practice tests, each laboratory performed 3 ABCD runs of 4 specimens each (a total of 12 specimens) on each of the 3 binder grades.

Average ABCD cracking temperature and strain jump for each binder and each laboratory are presented in tables 3 and 4 and figures 3 and 4. ABCD strain jump (e) at fracture may be converted into the fracture stress of the asphalt binder (σƒ) as follows:

σƒ = (K) ε•EABCD•AABCD/Abinder

where

  • K = stress concentration factor, 2.02
  • EABCD = Young’s modulus of ABCD ring, 140 GPa (20,305 ksi)
  • AABCD = Cross-sectional area of ABCD ring, 22.6 x 10-6 m2 (27.0 x 10-6 yd2)
  • Abinder = Cross-sectional area of asphalt binder, 40.3 x 10-6 m2 (48.2 x 10-6 yd2)

To use the ABCD for accurate strength measurement, each ring should be calibrated for load-strain relationship. The ABCD rings used in this ILS were not calibrated. The precision of ABCD strain jump results may be improved when calibrated ABCD rings are used. For easy comparison, the Y-axis scales of figures 3 and 4 are kept the same for all three binders.

Before determining the precision of the test, ASTM C802 requires the operator to identify and remove erratic data by plotting variance versus lab, as shown in figures 5 and 6. Even though Lab 1 (EZ Asphalt laboratory) was the only laboratory with significant prior experience with ABCD testing, many labs produced lower variance for the ABCD test (meaning the results of three runs of one sample are similar). There are quite few noticeably large variance values. However, it is not clear if they are erratic data. Since the number of participating laboratories (23) was significantly larger than the recommended minimum (10), the effects of few data with the large variance values could be insignificant. Data analysis was performed without removing these data, and the precision estimates of ABCD test results would be conservative ones.

The statistical summary of the ABCD ILS results is presented in table 5. The ABCD cracking temperatures of three binders were significantly lower than the PG low-temperature grade since ABCD tests evaluated the low-temperature cracking potential of asphalt binder only. The addition of mineral aggregates in the asphalt binder would raise the cracking temperature. For all three binders, the within-lab variances of the ABCD cracking temperatures were similar and independent of the cracking temperature. The between-lab variance for the medium-stiffness binder is lowest (s2 = 1.19), followed by the low-stiffness binder (s2 = 1.87) and the high-stiffness binder (s2 = 2.58). The low-stiffness binder seemed to deform easily during the trimming, transportation, and handling steps, resulting in larger variances than the medium-stiffness binder. The high-stiffness binder was extremely stiff at room temperature, making trimming unusually difficult and contributing to the large between-lab variance.

Two SBS-modified binders (PG 70-28 and PG 76-22) exhibited much larger strain jump than the unmodified high-stiffness binder (PG 88+6). The fracture strengths of the low-, medium-, and high-stiffness binders were estimated to be 6.5 MPa (943 psi), 6.3 MPa (914 psi), and 3.3 MPa (479 psi), respectively. The within-lab and the between-lab variances of the strain jump seemed larger for the binders with a larger strain jump. The SBS-modified binders, which showed about twice-larger strain jump than the unmodified binder, also had twice or more within-lab and between-lab variances.

The precision estimates of ABCD tests are given in table 6. When a single operator runs the ABCD test twice on the same sample, the acceptable range of two test results should be less than 2.69 ºC (4.84 ºF), 15.50 με, and 2.43 MPa (352.4 psi), for the ABCD cracking temperature, strain jump, and fracture strength, respectively. When two different operators at two different laboratories perform ABCD tests on the same sample, the acceptable range of two test results should be less than 3.85 ºC (6.93 ºF), 20.39 με, and 3.20 MPa (464.1 psi) for the ABCD cracking temperature, strain jump, and fracture strength, respectively.

Table 3. ABCD ILS Results: ABCD Cracking Temperature
Lab No. Cracking Temp (°C)
Average of 3 Runs Standard Deviation (°C)
PG70-28
(Lxx)
PG76-22
(Mxx)
PG88+6
(Hxx)
PG70-28
(Lxx)
PG76-22
(Mxx)
PG88+6
(Hxx)
1 -41.3 -40.0 -11.8 0.7 1.2 0.8
2 -40.3 -40.6 -11.0 1.6 1.6 2.7
3 -43.1 -40.1 -12.9 1.6 1.7 1.4
4 -40.3 -39.6 -11.7 1.0 0.6 0.8
5 -40.5 -40.4 -12.4 1.8 1.8 2.1
6 -41.7 -41.0 -14.9 2.2 2.3 1.3
7 -44.0 -42.3 -15.8 1.5 1.0 1.7
8 -42.7 -40.7 -14.7 4.2 2.6 0.8
9 -40.4 -40.0 -11.3 1.6 0.7 0.8
10 -44.8 -42.3 -13.4 1.9 1.0 0.9
11 -42.0 -40.9 -12.9 0.8 2.1 1.3
12 -41.3 -40.9 -12.0 2.0 1.2 0.9
13 -40.0 -39.4 -12.1 1.2 0.5 0.5
14 -42.4 -40.9 -11.6 2.9 1.3 0.7
15 -40.3 -39.6 -12.3 0.5 0.7 0.9
16 -40.5 -40.3 -12.2 1.2 1.5 0.7
17 -42.2 -41.3 -13.3 1.0 1.6 1.0
18 -42.7 -39.8 -11.6 1.3 1.0 1.9
19 -39.9 -38.8 -10.6 1.1 0.8 1.3
20 -39.4 -38.4 -7.9 1.2 1.7 3.6
21 -41.4 -41.0 -11.5 1.4 0.5 0.8
22 -41.0 -38.7 -12.2 2.1 1.9 0.6
23 -41.5 -42.4 -13.8 1.8 2.1 1.6
Average -41.5 -40.4 -12.3 1.6 1.4 1.3
Table 4. ABCD ILS Results: ABCD Strain Jump
Lab No. Strain Jump (με)
Average of 3 Runs Standard Deviation (με)
PG70-28
(Lxx)
PG76-22
(Mxx)
PG88+6
(Hxx)
PG70-28
(Lxx)
PG76-22
(Mxx)
PG88+6
(Hxx)
1 42.9 45.0 19.6 8.5 11.7 2.5
2 36.1 31.2 13.1 9.5 8.4 6.5
3 36.4 31.6 18.2 13.7 10.5 6.5
4 38.8 44.1 23.5 6.6 4.5 3.3
5 36.4 38.4 16.4 6.6 10.1 6.2
6 35.5 28.1 15.6 12.2 15.8 7.7
7 52.3 41.6 24.7 12.7 13.6 5.1
8 36.7 21.5 16.1 20.6 8.9 3.7
9 33.3 39.4 22.0 4.7 10.0 4.7
10 68.0 57.6 37.4 17.9 33.9 11.6
11 36.1 33.9 21.0 6.9 10.4 4.3
12 39.6 50.6 21.5 11.9 7.5 5.0
13 37.0 39.9 22.2 4.5 2.7 2.2
14 43.2 41.1 15.3 20.3 7.6 2.2
15 38.7 43.2 21.9 4.0 3.7 3.6
16 33.9 45.8 24.3 18.2 16.3 6.1
17 46.0 47.6 26.1 4.9 6.8 3.3
18 47.6 39.5 18.1 6.1 7.6 5.2
19 43.7 33.5 16.9 21.3 6.8 4.9
20 32.8 26.9 21.3 16.2 10.5 14.9
21 49.1 53.5 22.4 5.6 4.0 4.3
22 42.0 36.0 17.7 5.1 7.0 1.9
23 36.4 42.0 16.0 14.4 16.5 5.4
Average 41.0 39.6 20.5 11.0 10.2 5.3
Figure 3. Bar charts. ABCD ILS Results: Average ABCD cracking temperatures. A separate chart is shown each for low-, medium-, and high-stiffness binders. Each chart displays the average cracking temperature in degrees Celsius on a scale with 2-degree intervals for each of the 23 labs and the average of all labs. For PG 70 28 (low stiffness), values are as follows: average, about  41.5 ºC; highest, close to -45 ºC (lab 10); and lowest, about  39 ºC (lab 20). For PG 76-22 (medium stiffness): average, about  40.5 ºC; highest about  42.5 ºC (labs 7, 10, and 23); and lowest, about  38.5 ºC (lab 20). For PG 88+6 (high stiffness): average, about  12.5 ºC; highest, nearly  16 ºC (lab 7); and lowest, nearly  8 ºC (lab 20). There appears to be more variability among labs for the low-stiffness binder.

Figure 3. ABCD ILS Results: Average ABCD cracking temperatures.

Figure 4. Bar charts. ABCD ILS Results: Average strain jump. A separate chart is shown each for low-, medium-, and high-stiffness binders. Each chart displays the average strain jump in microstrains on a scale of 0 to 80 with intervals of 20 microstrains for each of the 23 labs and the average of all labs. For low-stiffness binder, values are as follows: average, a little over 40 microstrains; highest, close to 70 microstrains (lab 10); and lowest, about 30 microstrains (labs 16 and 20); however, most of the labs show numbers below and very near 40 microstrains. For medium-stiffness binder: average, 40 microstrains; highest, close to 60 microstrains (lab 10); and lowest, just over 20 microstrains (lab 20). For high-stiffness binder: average, 20 microstrains; highest, around 38 microstrains (lab 10); and lowest, around 15 microstrains (lab 2).

Figure 4. ABCD ILS Results: Average strain jump.

Figure 5. Graphs. ABCD ILS Results: Within-lab variance of ABCD cracking temperature. A separate chart each for low-, medium-, and high-stiffness binders displays the variance in degrees Celsius on a scale of 0 to 6 with 1-degree intervals for each of the 23 labs and a horizontal line showing the mean for all labs. For low-stiffness binders, the variance values are as follows: mean, a hair under 1 degree (16 of the labs had a variance less than 1 degree) and highest, about 5 degrees (lab 8). For medium-stiffness binders, the mean is slightly under 1 degree, with 18 labs equal to or less than 1 degree; outliers at between 3 and 4 degrees were labs 8, 17, and 22. For high-stiffness binders, the mean is also just under 1 degree, with 17 labs between 0 and 1 degree and 6 of those with no variance. The outliers at 3 or more degrees were labs 8, 10, 20, and 23.

Figure 5. ABCD ILS Results: Within-lab variance of ABCD cracking temperature.

Figure 6. Graphs. ABCD ILS Results: Within-lab variance of strain jump. A separate chart each for low-, medium-, and high-stiffness binders displays the variance on a scale of 0 to 350 in 50-microstrain intervals for low-stiffness binders and 0 to 140 in 20-microstrain intervals for medium- and high-stiffness binders, as well as a horizontal line showing the mean for all labs. For low-stiffness binders, the mean is just under 50 microstrains, with 17 labs at or below the mean. The extreme outlier is lab 19 at about 300 microstrains; the remaining are at 100 or lower. For medium-stiffness binders, the mean is at about 32 microstrains, with 18 labs below 40. The outlier at nearly 120 is lab 10, with labs 20 and 23 at 70 and 60, respectively. For high-stiffness binders, the mean is 20 microstrains, with 19 labs at or below the mean. The outliers are lab 11, at about 110, and 20 and 23, at about 70 and 60 respectively.

Figure 6. ABCD ILS Results: Within-lab variance of strain jump.

Table 5. Statistical Summary of ABCD ILS Results
  Cracking Temperature (°C) Strain Jump (με)
PG70-28
(Lxx)
PG76-22
(Mxx)
PG88+6
(Hxx)
PG70-28
(Lxx)
PG76-22
(Mxx)
PG88+6
(Hxx)
Average -41.5 -40.4 12.3 41.0 39.6 20.5
Minimum -44.8 -42.4 -15.8 32.8 21.5 13.1
Maximum -39.4 -38.4 -7.9 68.0 57.6 37.4
 
Within-lab variance 0.94 0.86 0.91 45.0 31.1 17.3
Minimum 0.03 0.02 0.05 2.3 1.9 1.9
Maximum 5.00 3.77 3.71 98.4 130.5 112.0
 
Between-lab variance 1.87 1.19 2.58 62.4 75.1 25.7
Table 6. Precision Estimates for ABCD Test
Condition Standard Deviation (1S)a Acceptable Range of Two Test Results (D2S)a
Single-operator precision:
Cracking temperature (°C) 0.95 2.69
Strain jump (µε) 5.48 15.50
Facture stress (MPa) 0.86 2.43
     
Multilaboratory precision:
Cracking temperature (°C) 1.36 3.85
Strain jump (µε) 7.21 20.39
Facture stress (MPa) 1.13 3.20

aThese values represent the 1S and D2S limits described in ASTM Practice C 670.

PDF files can be viewed with the Acrobat® Reader®

More Information

Events

Contact

Julie Zirlin
Highways for LIFE
202-366-9105
julie.zirlin@dot.gov

Updated: 09/02/2011

FHWA
United States Department of Transportation - Federal Highway Administration