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


<< PreviousContentsNext >>

Use of Magnetic Tomography Technology to Evaluate Dowel Placement

Guidelines for Evaluating Dowel Alignment Using MIT Scan-2


The placement accuracy of dowel bars can be evaluated very efficiently using MIT Scan-2. MIT Scan-2 is a state-of-the-art nondestructive testing (NDT) device that was developed specifically for detecting dowel bars placed in concrete pavements. The technology employed is magnetic tomography. Presented in the following are guidelines for using MIT Scan-2 to monitor construction quality of dowel alignment.


MIT Scan-2 is simple to operate, is efficient, and provides accurate, real-time results in the field. The device was intended for use on dowel bars placed using a dowel bar inserter (DBI), but it can also be used to scan bars placed in a basket, if the following conditions are met:

  • The bars are epoxy-coated or painted to insulate the bars from the basket.
  • The transport ties on the basket are cut.

If these conditions are met, good results can be obtained for dowel bars placed in baskets. Dowel baskets can be calibrated to obtain the same level of accuracy as bare bars. MIT Scan-2 results are not affected by changing moisture conditions in the concrete, so the testing can be conducted at any concrete age, including over fresh concrete. The test results are also not affected by the presence of water on the pavement surface. The operating temperature range is - 5 °C to 50 °C (23 °F to 122 °F).


For inserted bars, MIT Scan-2 can provide an accuracy of +5 mm (+0.20 in.) with 95 percent reliability on horizontal and vertical misalignment. The accuracy on the lateral bar position (side shift) is +8 mm (0.16 in.), and that on depth is +4 mm (0.08 in.). These measurement tolerances are valid for the following conditions:

  • Dowel bar depth 150 ± 40 mm (4.3 to 7.5 in.)
  • Horizontal and vertical misalignment less than +40 mm (1.6 in.)
  • Lateral shift (side shift) less than +80 mm (+3.2 in.)

The measurement error is higher for more severely misaligned bars.

For dowel bars placed in a dowel basket, good approximate results can be obtained without any special consideration for the dowel basket. With basket-type-specific calibration, a similar level of accuracy can be obtained for dowels placed in baskets as bare bars.

The presence of metal objects such as tie bars, nails, reflectors, or other objects within the detection range of the Scan-2 can affect the results, effectively invalidating the results for the affected bars. Such objects are usually easy to detect on the graphical output of the Windows-based software accompanying MIT Scan-2 (MagnoProof). When scan results indicate significant misalignments, a close inspection using MagnoProof is highly recommended to verify that the results are not affected by the presence of extraneous metal objects.


After setup, testing takes 1 minute or less per joint, depending on the number of lanes tested. Up to three lanes can be scanned together in a single pass, and testing multiple lanes together does not significantly slow the rate of testing. The device setup takes about 20 minutes, but marking out the test area with joint numbers can take an hour or longer, depending on the length of the section to be tested. To ensure that the test data are accurately correlated to the correct joint, marking joint numbers on the pavement is highly recommended.

Peak productivity is about 100 joints per hour when testing a single lane, but this rate of testing is difficult to sustain throughout a workday. For planning purposes, a good estimate of average daily productivity is about 70 joints per hour for single lanes (assuming a crew of 2) and moderately less (e.g., 60 joints per hour) for 2 lanes on continuous testing. If the test areas are scattered, the time required to move from one location to another (which may involve disassembling and reassembling the rail) should be taken into consideration in planning.

The battery life of the sensor unit may be a factor limiting daily productivity. The device is designed for a minimum of 8 hours of continuous testing on one charge, and under most conditions this is easily achieved. On single-lane testing, up to 12 hours of continuous testing may be possible on one charge. However, the battery life may be greatly reduced (e.g., 30 to 40 percent reduction) when testing under low temperatures (e.g., 30°F [2°C] or lower).


The principal limitation of MIT Scan-2 is that the presence of foreign metal affects the results. Although the presence of such metal objects is easily detected, the loss of information for the affected bars is a limitation. The most problematic of such objects is tie bars in close proximity to the joint being evaluated. The inability to obtain accurate results for dowel bars influenced by tie bars is a limitation of MIT Scan-2. However, the fact that a tie bar is within the influence region of the scanned joint may already be an indication of a problem, because most States require tie bars to be placed at least 500 mm (20 in.) away from the joints.


For evaluating the quality of dowel alignment in a pavement section, testing a random sample of 50 consecutive joints is recommended. Depending on the consistency of the testing results and quality of construction, the rate of testing may be on per day's paving, per mile, or per section of project basis.

For construction using a DBI, any systematic problems due to equipment adjustment could be determined from a fewer number of joints (e.g., 20 joints); however, the portland cement concrete (PCC) mix has a significant effect on the alignment of inserted dowel bars. Testing over the full 50-joint sample is recommended to capture the effects of any batch-to-batch variations in PCC mix consistency, and to detect any problems resulting from segregation.

On projects constructed using dowel baskets, problems with dowel misalignment are usually the result of inadequate anchoring of the baskets (i.e., inadequate number or size of anchoring pins). Inadequate anchoring causes the baskets to burst open, deform, or move during paving, which in turn, causes severe dowel misalignment. If the anchoring procedure is grossly inadequate, the problem will be prevalent and readily apparent in even a relatively small sample. Even if adequate anchoring is specified, occasional problems may arise due to poor work quality. The baskets are also subject to damage during construction. These types of problems are more random and less frequent. To ensure that a representative sample of random problems is captured in the test results, testing 50 consecutive joints is recommended.


The details of the operation of MIT Scan-2 are provided in Appendix A, MIT Scan-2 Operations Guide. Following is a summary of the key steps:

  • Preparation - fully charge the battery on the sensor unit and the onboard computer. A full charge provides up to 8 hours of testing.
  • Setup
    • Connect the onboard computer and the sensor unit.
    • Switch on both the sensor unit and the onboard computer. Check the memory status to ensure that both units have adequate power for the amount of testing planned. The sensor unit requires 5 minutes to warm up. If measurements are taken while the unit is not warmed up, additional errors may be introduced. The readiness of the unit can be verified by setting up and then testing the same joint three times without moving the rail. If the unit is properly warmed up, the maximum difference in measurements for any dowel should be 2 mm (0.08 in.) or less.
    • Assemble the rail system.

Before testing begins, number and mark the joints sequentially to keep track of those being tested. In the field, joint numbers are a simpler way to keep track than station numbers. To speed up the marking process, a paint mark (a dot) may be placed every 5th joint and the joint number on every 10th joint. During marking, any station numbers or mileposts found in the area should be recorded in the field notes. By recording the begin station and the station number for any joints located close to the station numbers stamped on the pavement, the station number for all joints can be determined. MIT Scan-2 automatically keeps track of both joint and station numbers.

As with any field survey, maintaining good field notes is important. The following information should be recorded at the beginning of each pavement section tested:

  • Route number and direction
  • Begin station and milepost
  • Lane number(s)
  • Direction
    • Direction of scanning (e.g., from outside shoulder to centerline joint)
    • Direction of survey
  • Time of scanning the first joint - MIT Scan-2 uses the date and time as the data file name. The record of the date and time of testing provides additional reference information for the pavement section tested.

During testing, the presence of any metal objects in the scan area that can interfere with MIT Scan-2 results should be recorded in the field notes. Examples of such objects include reflectors and drainage inlet covers.


The MIT Scan-2 results can be used directly to check for compliance with specification requirements. Dowel placement tolerances are typically specified in terms of the following:

  • Horizontal and vertical misalignment - Typical tolerances range from 5 mm (0.18 in.) to 14 mm (0.56 in.) for both horizontal and vertical misalignment for 457-mm (18-in.) dowel bars. The most common standard is 10 mm (0.375 in.).
  • Lateral displacement (side shift) - Typical tolerances range from 25 to 50 mm (1 to 2 in.); 50 mm (2 in.) is the most common.
  • Depth deviation - Typical specifications call for the bars to be placed within 25 mm (1 in.) of the slab middepth.

The evaluation can be based on either field results (produced by MagnoNorm) or the results obtained using the Windows-based software (MagnoProof) accompanying MIT Scan-2. The field results are accurate for the following conditions:

  • Mean dowel depth 150 + 40 mm (4.3 to 7.5 in.)
  • Horizontal and vertical misalignment +20 mm (0.8 in.)
  • Maximum lateral position error (side shift) <+50 mm (2 in.)

For other conditions, MagnoProof can be used to conduct a more comprehensive analysis. MagnoProof incorporates a more robust solution algorithm and allows more manual control of the analysis process to provide more accurate results. The presence of foreign metal is easily identified on the signal intensity plot provided on the MagnoProof screen. As mentioned earlier, MagnoProof analysis is highly recommended for any joints showing significant misalignment in the field results to ensure that results are not affected by the presence of foreign metal.

Limitations of Existing Standards

Most agencies have fairly strict tolerances on dowel placement accuracy, but those standards are based on limited laboratory and field data. In some cases, the manufacturing tolerances for dowel baskets are adopted directly as the dowel placement tolerance. The actual dowel bar alignment needed to ensure good pavement performance is largely unknown.

A recent study showed that many well-performing pavement sections contain at least a few joints that are potentially locked due to dowel misalignment (Yu 2005). Projects with a significant number of misaligned bars performed well without showing any signs of distress after 8 or more years of service under heavy traffic. The following conclusions were drawn based on the field observations:

  • In a joint that is already locked, additional misaligned bars have no further adverse effect.
  • On short jointed concrete pavements, the presence of a few occasional locked joints is not likely to have a significant adverse impact on pavement performance, as long as the locked joints are separated by working joints.

The above observations suggest that a joint-by-joint evaluation may be more appropriate for the evaluation of dowel alignment from a pavement performance perspective. Further research is also needed to establish the critical level of misalignment, considering all relevant factors.

General Quality of Dowel Alignment

The general quality of dowel bar alignment may be evaluated in terms of the percentage of bars at various levels of misalignment, as shown in Figure B1. The effects of horizontal and vertical misalignment are combined in this figure simply by taking the maximum value of misalignment for each bar, either horizontal or vertical. Another approach to combining the effects of horizontal and vertical alignment is to use the resultant misalignment, but the dowel placement tolerances are defined in terms of horizontal and vertical misalignment, and there are no standards for the resultant misalignment.

Figure B1. Example distribution of dowel misalignment (maximum horizontal or vertical) by range of misalignment (Yu 2005).
Example distribution of dowel misalignment (maximum horizontal or vertical) by range of misalignment (Yu 2005).

Figure B1 shows that the frequency distribution plot could be used to compare the quality of dowel alignment of different projects. Both the magnitude and number of misaligned bars affect proper functioning of pavement joints. The bars that are more severely misaligned (e.g., more than 20 mm [0.79 in.]) are much more critical than bars that are more moderately misaligned (e.g., 10 < d < 15 mm [0.39 < d < 0.59 in]). In the example given in Figure B1, the dowel alignment of IN1 is better than IN2, both in terms of total percentage of misaligned bars and the percentage of bars with more severe misalignment.

Joint-by-Joint Evaluation

Although the distribution of dowel misalignments shown in Figure B1 reflects the general quality of dowel alignment, it does not describe how the misaligned bars are distributed within a project, which may be important to pavement performance. For example, 10 badly misaligned bars in 1 joint affect the performance of that single joint, but the same number of badly misaligned bars evenly distributed over 10 joints (i.e., 1 bad bar per joint) affects the performance of 10 joints. From a pavement performance perspective, the latter case is much more critical.

The Joint Score and Rolling Average Joint Score introduced in the recent report (Yu 2005) could be used to perform a joint-by-joint evaluation. The Joint Score is a measure of the combined effects of misaligned dowel bars at a joint. Joint Score is determined by adding 1 to the sum of the product of the weights (given in Table B1) and the number of bars in each misalignment category. For example, if a joint has four misaligned bars in the range 15 to 20 mm (0.6 to 0.8 in.), the joint score is 9; if a joint has one misaligned bar in the range 15 to 20 mm (0.6 to 0.8 in.) and one bar in the 25- to 38-mm (1- to 1.5-in.) range, the score is 8. Further research is needed to refine and verify Joint Score, but the weighting factors listed in Figure B1 may be used as an interim measure.

Table B1. Weighting Factors Used to Determine Joint Score
Range of Misalignment, mmWeight
10 < d < 150
15 < d < 202
20 < d < 254
25 < d < 385
38 < d10

The Joint Score has the following meaning:

  • Joint Score s < 5 indicates very low risk of joint locking
  • Joint Score 5 < s < 10 indicates low risk of joint locking
  • Joint Score 10 < s < 15 indicates moderate risk of joint locking
  • Joint Score 15 < s indicates high risk of joint locking

In general, a Joint Score of 10 or higher indicates a significant potential for joint locking.

Field experience indicates that a few randomly distributed locked joints do not adversely affect pavement performance (Yu 2005). However, consecutive locked joints are not desirable. The risk of joint problems due to clusters of locked joints can be identified by determining the Running Average Joint Scores as follows:

  • Cap Joint Scores - If a joint has a score greater than 10, assign a Joint Score of 10; otherwise, use the actual Joint Score.
  • Determine Running Average Joint Score - The Running Average Joint Score is the maximum of the average of the capped Joint Scores for two joints ahead and two joints behind the current joint. A value of 10 indicates 2 or more consecutive locked joints and a high risk of developing distress due to poor dowel alignment.

Example plots for Joint Score and Running Average Joint Score are shown in Figures B2 and B3. In this example, the project has four joints with high potential for joint locking (Joint Score > 10), but there are no clusters of potentially locked joints. The Running Average Joint Score is less than 10 for all joints in this project (Figure B3).

Figure B2. Example Joint Score plot. A score greater than 10 indicates a high potential for joint locking (Yu 2005).
Example Joint Score plot. A score greater than 10 indicates a high potential for joint locking (Yu 2005).
Figure B3. Running Average Joint Score plot corresponding to Figure B2, indicating low risk of any performance problems due to potentially locked joints (Yu 2005).
Running Average Joint Score plot corresponding to Figure B2, indicating low risk of any performance problems due to potentially locked joints (Yu 2005).
<< PreviousContentsNext >>
Updated: 04/07/2011

United States Department of Transportation - Federal Highway Administration