The pull out capacity of an anchor is defined as the point where the anchor continues to move in the direction of the anchor pull without any further increase in applied load. A load-deflection curve, on the other hand, is a complete record of how the anchor head moves during the application of load. According to NFPA 501A-1975 [69] all that is required is that an anchor be able to hold 4725 lb (21 kN) at a deflection of two inches (51 mm) or less.

Regulatory agencies in many states presently require to validate the adequacy of anchors by testing. Mobile home anchor manufacturers must submit their anchors to (an independent agency) for testing. The agency then goes out and finds typical soil deposits according to Class and perform pull-out tests. The pull-out tests are summarized and a letter report is written to the appropriate state agency showing conformance of an individual's hardware to the state's requirement for loads and displacements.

Unfortunately, most of the reports on load tests, many of which are reported in the unpublished literature, do not give adequate information on the engineering soil properties of the soil such as the shear strength or the in-place unit weight. Thus, only in a very general way can the data that are available be used to further engineering knowledge of the behavior of soil and rock mobile home anchors under load. It is suggested that the procedure used in future pull-out load tests be standardized so that the information may advance the state of the art and benefit mobile home owners. Available test data are discussed in the following section.

The helix anchor was chosen to be discussed first because it tends to produce smooth load deflection curves. The displacement of the anchor head, herein referred to as uplift deflection, observed in a typical test is shown in figure 7.1. The test data presented (source of this information has not been identified at the request of the supplier) in

Figure 7.1 Pull-out Load versus Uplift Deflection for Helix Anchors. (A) 4-inch Anchors at 3-foot depth; and (B) 6-inch Anchors at 4-foot depth

figure 7.1 are shown for four inch and six inch single helix anchors as well as for four inch and six inch twin helix anchors. The upper curve shows load-deflection curves for an anchor with an embedment depth of three feet while the lower curve is for anchors installed at a depth of four feet. The number in parentheses adjacent to the test order number is the Soil Test Probe (STP) reading in inch-pounds at the depth of the (bottom) of the helix. Using NFPA 501A-1975 failure criteria [69] only tests 4, 7, 2, 1, and 8 meet the requirements. Actual failure load is indicated with the exception of tests numbers 1 and 7. The failure loads for tests 1 and 7 are undefined because the anchors were not loaded to the pull-out load.

Unfortunately, when these tests were conducted, no further information on soil properties was obtained other than classifying the soil according to its Soil Test Probe reading. Of all the field pull-out test data that were supplied by numerous helpful anchor manufacturers, there are only about 6 tests with adequate information concerning pull-out capacity versus load-displacement data and their correlation with soil properties that can be used to establish a rational design approach. However these data are so scattered that no empirical relationships are apparent. Realizing that the relationship between pull-out capacity and soil classification will be very general, a plot of pull-out capacity versus Soil Test Probe reading can be constructed as shown in figure 7.2. Data from four, six, and eight inch (0.10, 0.15 and 0.20 m) helix tests are presented in this graph for tests in which all the anchors were embedded four feet (1.2 m). One may be tempted to draw a curve from points A to B on this graph for a six-inch single helix for a depth of four feet. However, due to insufficient data, it is felt that no such curve is justified. No additional data could be located.

Many times load tests are conducted with a load cell or dynamometer installed between the anchor and the pull-out device. The pull-out device is sometimes a backhoe or drilling machine as opposed to equipment that can carefully control the load. No data could be located on cyclic loading of mobile home anchors.