Loop Configuration Evaluation

Written by Don Wood, County of Santa Clara

An evaluation of the effectiveness of various forms of vehicle loop installations and configurations. In 1990, the County of Santa Clara made an evaluation of the effectiveness of various loop detector configurations for the detection of bicycles. It was desirable to provide bicycle detection for each lane at each intersection. There were all sorts of claims regarding the effectiveness of various loop configurations being made. This evaluation conducted by the County of Santa Clara was not extremely detailed nor exhaustive. It did provide an indication of how these types of loop installations would work as they are typically installed by the County.

Our standard loops being used at that time were Type “A” (6×6′ hex loop). There were a few locations with Type “Q” loops installed as the front loop of an array of one Type “D” and two Type “A” loops. Our experience with that configuration was not proving successful.

We prepared test loop fixtures of type “A” loops, type “D” loops and a type “E” (Round) loops and a type “B” (diamond) loop. Identical tests were performed on each loop in “solo” and as part of an array including the test loop and two Type “A” loops connected in series. Measurements of the sensitivity of the various configurations when presented with the influence of lightweight bicycle wheel were recorded. In addition, a measurement of the height of the detection field was made by suspending a metal object over the loops and raising the object until the actuation was dropped by the detector.

There were three basic questions that we were seeking to answer:

1. What was the sensitivity of the various loop types to small metal objects (bicycles) within the loops detection area?

2. What was the effect of combinations of various loop configurations on the sensitivity of various loops?

3. What was the range of detection height of the various loop configurations.

The County typically uses area detection as part of it’s signal installations. These are usually three six by six foot Type “A” loops centered in the traffic lane starting three feet behind the stopbar and spaced ten feet between the loops (16 foot center to center). They were normally all three turn loops except the head loop of the left turn lanes which were normally four turn loops. We were trying to improve our detection without requiring the replacement of all loops. Our intent was to try to install new front loops that were bicycle sensitive while still maintaining proper detection of larger vehicles.

Four loop configurations were tested.

Configuration

Description

Turns

Type “A”

Hex

3

Square

3

Type “B’

Diamond

3 and 5

Type “D”

3 and 5

Type “E”

Round

3 and 4

Test #1

A bicycle was rolled through the loop area at three different positions. At the center, at a point approximately 18 inches from the left edge (45 degree), and at the left edge. At each position, the highest variation in loop frequency was noted and the percentage of deviation was calculated. This test was performed twice, once on the loop connected to the amplifier in “solo” and a second time with the loop connected in series with two Type “A” (standard 6×6) loops.

The results of this test indicate that the type “A” loop, when connected in solo has the greatest amount of deviation. In combination, the Square 3 turn loop provided the greatest deviation, however, the Type “A”, the Diamond, and the Round 4 turn combinations were close.

Test #2

A metal object was placed over the loop area and the deviation in loop frequency was noted at various heights and the point of loss of detection was noted. Loss was indicated in feet above the pavement.

From this test it is apparent that the Type “B” (diamond) 5 turn loop performed best overall followed closely by the round 4 turn loop. The type “D” 5 turn loop performed the best at the first level (0 feet), but demonstrated low sensitivity at all other heights.

Another factor in loop performance is loop life. Loops can fail prematurely due to pavement failure around the loop. The extent of damage to the roadway can be reduced by reducing the number and extent of cuts to the roadway in particular eliminating closely spaced cuts. Using preformed loops or paving over the newly installed loops are also effective means of reducing the pavement failure around loop installations, however, this is not always practical especially in a replacement situation.

Considering the results of both tests, and the methods of installation, the performance of the type “E” (round) loop is very good and the roadway damage very minor. The performance of the type “B” is very similar to the round loop with possibly a more even sensitivity across the entire width of the loop. The type “A” loop is also very sensitive and has good detection height, however, they do not provide even detection across the entire width of the loop. They do require more pavement cuts and patching and therefore cause more damage to the roadway. The type “D” loop configuration’s main advantage would be the generally even sensitivity throughout the area of detection. However, it is not as sensitive nor does it provide good detection above one foot in height. The damage caused to the roadway during installation of the Type “D” loop is the greatest of any configuration.

The depth and number of cuts involved in installing each type of loop is extremely different.

The depth required for a round loop is only that required to cover the four loop conductor turns plus the required filler height. Two and one half inches is sufficient. The Type “D” loop requires approximately three inches throughout the loops pattern and approximately five inches at the point of intersection of the homerun. In addition, the type “D” presents the danger of large pieces of AC breaking out due to the intense cut pattern. Even with the intense cut pattern, there are still some severe angles presented to the loop conductor path. These angles may lead to early loop failure. The type “E” (round) loop has only one intersection of the cuts and is cored to improve the turn radius at the exit in order to reduce the potential of loop failure. The type “B” (diamond) loop requires the next least number of cuts when it is cut with square corners that are cored.

The height of detection is important due to the passage of high axle vehicles and the fact our signals rely on good detection in order to extend the timing of the phase. If a truck or 4×4 vehicle were to pass slowly through the lane, the phase might terminate prematurely. The loops with large center areas have higher detection height (over four feet high for the round loop). Loops with closely spaced runs have lower detection heights. Roughly one half the distance between sides of the loop is an indication of the reliable detection height.

From the above tests we determined that the type “E” (round) or type “B” (diamond) loops provide the best detection capabilities for the vehicle and bicycle detection needs of the County. They have excellent sensitivity, a very good detection pattern including the detection height and bicycle sensitivity. And the potential for pavement damage is reduced significantly by use of coring.

Both the round and diamond loops may also present less adjacent lane pickup even with higher sensitivities due to the edge alignment of the loop being 45 degrees to the adjacent lane. More of the loops most sensitive areas are located farther away from the edge of the lane. The diamond loop may be best in this regard due to its straight edge at 45 degrees.

The original data from these loop tests is no longer available. This article is taken from a report that summarized the test results. The report was prepared by Don Wood using data collected by Tony Rucker and Stuart Leven. Our standard loop installation was revised as a result of these tests to be either round or diamond loops in an array of three loops spaced 16 feet center to center with the front loop having 4 turns and the middle and rear loops having 3 turns, all connected in series. The type”B” loops are installed without corner crosscuts using coring instead. We have found this configuration to meet our requirements for vehicle and bicycle detection.