Direct Verification of Threaded Fastener Locking Compounds and Adhesives
This article explores the feasibility of a method to directly verify that thread lock compound or adhesive has cured sufficiently to provide secondary locking. The application of a torque after assembly and cure time was found to provide a viable verification test. Fastener material and coatings were found to significantly affect cure. For example, medium-strength anaerobic compound used with fasteners made of inactive material such as stainless steel had curing issues which prevented determination of a useful verification test torque. However, modifications such as higher-strength compound or primer can be introduced making this method viable even for fasteners made with inactive materials and coatings. This article outlines a process to implement this method in practice. This requires sample tests with representative product to determine a test verification torque. This process is particularly useful in identifying curing and locking performance issues, and provides guidance for modification so that the method can be successfully implemented in practice. Test results show that application of verification test torque with or without standard vibration test exposure does not degrade the locking performance.
KeywordsFastener Thread lock compound Loosening Thread adhesive Prevailing torque Secondary locking Verification method Locknut
Introduction and Background
Threaded fasteners remain as widely used components in machinery and structures. Fasteners literally hold systems together. The friction in the thread, bolt head, and nut face provide an inherent resistance to (or locking against) loosening moments induced from the joint preload itself and external loads.
Mechanical features such as lock wire or cotter pin with castle nut;
Prevailing torque devices such as lock nuts with distorted threads or nylon strips; and
Adhesives such as anaerobic compounds and epoxies.
An important aspect of the assembly process with a secondary locking feature is a means of verifying the locking feature during or after assembly. Mechanical features such as lock wire and cotter pins can be visually verified after assembly. Prevailing torque in lock nuts can be verified with a torque measurement during assembly. Locking from fastener adhesives and compounds is generally not directly verified. In some cases, such as in the aerospace industry, for example, an indirect method with sample coupons with representative product is utilized, in which samples are destructively tested with a removal break-away torque measurement. The direct measurements of the sample coupons are used as an indirect verification of the actual hardware.
Issues with fastener thread lock compound curing can result in insufficient locking. Such issues result in restrictions on use of thread lock compound. This has occurred recently in aerospace applications . The basis for these restrictions is instances of insufficient curing and lack of direct verification. The purpose of this article is to present a means of direct verification for such product which assesses curing and locking function.
Ideally, thread lock compound is applied to the threads before assembly, completely fills the space between the threads upon assembly, cures, and thereby prevents or limits slip within threads. Tests show that threaded fastener locking compounds can provide excellent locking of fasteners in bolted joints provided the compound cures sufficiently [2, 3]. However, cure is not guaranteed in all cases. For example, even though it is widely known that anaerobic compounds require the absence of oxygen to cure, it is sometimes used in blind hole applications where air tends to get trapped between mating threads and inhibit cure. In addition, its ability to cure depends on the chemistry of the fastener material or coating . Curing occurs faster when used with fasteners made from active metal such as plain steel, whereas cure with inactive materials such as stainless steel requires more time. and the compound does not reach full strength .
This article presents the results of a series of tests focused on assessing the viability of a direct verification method for thread locking compounds. Different fastener materials, coatings, compounds, and loading environments are considered. It is found that direct verification is viable, but sensitive to fastener materials and coatings. Assessment of locking compound degradation from application of the verification test and dynamic loading is performed.
The main objective of this study was to identify a method for implementing direct verification testing of threaded fasteners with thread lock compound. This was accomplished by performing a series of tests for different fastener configurations and analyzing the results.
The tests performed include torque–tension tests, break-away torque tests, verification torque tests, and dynamic tests. Torque–tension tests were performed to establish the torque–tension relationship for each configuration.
Break-away torque measurements were performed on samples of uncured and cured product. These data were statistically analyzed and used to determine and specify a verification of cure test torque.
Tests were performed to assess for possible locking compound degradation resulting from the application of a verification torque. Tests were also performed to assess for degradation from dynamic tests alone and in combination with application of verification torque.
Test Specimens and Fixtures
Plain steel and zinc-coated steel 0.25-28 UNF thread fasteners were used in this study. These included 2.5-in.-long, grade 8 hex head bolts with 0.75-in. thread length with mating grade 8 hex nuts, and washers.
The thread lock compounds used were Loctite 242 and 271 which are medium- and high-strength anaerobic compounds, respectively. Loctite 7471 primer was also used.
Fastener preparation consisted of cleaning, drying, priming, and applying thread lock compound. All fasteners were cleaned in an ultrasonic bath of methyl ethyl ketone for 5 min to remove contaminants and residual oils from manufacturing, and then allowed to dry for 15 min. Fastener threads were sprayed with primer and allowed to dry for 15 min. Locking compound was applied to both internal and external threads. One to two drops was sufficient to visually coat the threads. The fixture spools were assembled with the test fasteners employing a tightening torque of 205 in.-lb. This torque was applied slowly and held for 5 s with a dial-type torque wrench. This torque provided a preload of 65–75% yield for the fasteners tested.
Torque–tension tests were performed to assess the relation between tightening torque and preload for the fasteners studied. A separate fixture with a load cell was used for this test. A tightening torque of 205 in.-lb was gradually applied with a dial-type torque wrench and held for 5 s. The resulting preload was measured. Six samples for each fastener combination were tested. The mean preload for the plain steel was 3610 lb with a standard deviation of 122 lb. The statistics for the zinc-coated component was a mean of 2960 lb and standard deviation of 200 lb. These data reveal that the uncertainty or spread in the torque–tension relationship at 205 in.-lb is within plus or minus 15%. This is not uncommon for fasteners with threads coated with lubricant or liquid thread lock compound.
Break-away Torque Tests
Break-away torque is defined as the torque required to initiate motion or movement of a fastener, specifically, a nut or screw head. Break-away torque can be measured in either the loosening or tightening direction.
In this study, break-away torque is measured in the loosening direction of an assembled fastener. Measurements are obtained for sets of fastener samples with uncured and with cured thread lock compounds. These data are used to assess the viability of the direct verification method for a given fastener and compound configuration, and to identify a test verification torque value.
The overlap of the distributions in Fig. 5 is the result of a lower strength of the locking compound due to the zinc coating and insufficient cure in some specimens. This was confirmed by visual inspection upon disassembly post testing by the presence of uncured compound in the threads. Thread lock compound manufacturer's data sheets include data highlighting differences in cure strength as a function of fastener material and coatings. Clearly, the medium-strength compound used with zinc-coated steel fasteners does not provide as much locking as when used with plain steel fasteners. Increasing the time for cure from 24–48 h did not alter the results significantly. Using a higher-strength thread lock compound notably increases the measured break-away torque values of cured specimens, resulting in separation of the uncured and cured distribution plots.
The break-away torque measurement data presented in Figs. 4 and 5 are representative of other fastener material and thread lock compound configurations tested but not reported in this article. Specifically, the measured break-away torque distributions of uncured and cured specimens exhibit either overlap or separation.
In this study, verification torque is defined as the torque value applied to a fastener to assess for sufficient cure or verify locking. This torque is applied in the loosening direction after a specified cure time, nominally 24 h. It is applied slowly and held for 5 s.
The distributions of measured break-away torque for the uncured and cured samples are used to assess the viability of the direct verification process for a given fastener and lock compound configuration, and to identify a test verification torque value.
The distribution plots in Figs. 4 and 5 are representative for all fastener materials and compound strengths tested. The distributions of uncured and cured samples will either have separation as shown in Fig. 4 or exhibit overlap as shown in Fig. 5. In general, higher-strength compounds and active materials will provide larger separation in these torque distributions. Lower-strength compounds and inactive materials will result in less separation or overlap.
Separation of the uncured and cured break-away torque distributions is necessary to implement the direct verification method. A verification torque value is selected near the center of the separation range. For example, for the plain steel fastener with medium-strength compound samples in Fig. 4, the center is about 172 in.-lb. Specifying this value for a verification torque for this fastener and compound configuration is reasonable. It is at 2.5 standard deviations above the uncured sample mean, and at about 2.5 standard deviations below the cured sample mean. Instances of approval of uncured specimens and rejection of cured specimens have an equal probability. Specifying a verification torque value above 172 in.-lb decreases the probability of approving uncured specimens and increases the probability of rejecting cured specimens. This is more conservative and reduces the probability of approval of uncured specimens. On this basis, a more conservative value for a verification torque for this configuration is 174 in.-lb.
Eighteen plain steel fasteners with medium-strength thread lock compound were assembled, given 24-h cure time, and then subjected to an application of the 174 in.-lb verification torque for 5 s. One of the specimens failed during the application of this torque. The other seventeen specimens passed the verification test. Break-away torque measurements were then taken from these seventeen specimens. These ranged from 180 to 200 in.-lb with a mean of 191 in.-lb and a standard deviation of 5.5 in.-lb. These data are consistent with the break-away data collected from specimens without a verification torque test presented in the previous section. This indicates that a single application of the verification torque test does not degrade the performance of the thread lock compound, but does identify specimens (one in this case) with insufficient cure.
In cases where overlap of uncured and cured break-away torque distributions occurs, statistics indicate the probabilities of passing uncured specimen and failing cured specimens are too high. However, from a practical perspective, the amount of secondary locking offered by the compound is quite low.
An example of such overlap was found with break-away torque measurements of zinc-coated steel with medium-strength compound presented in Fig. 5. In such cases, secondary locking is minimal and application of direct verification torque is not practical. Modification of the fastener material, coating or lock compound is required before the direct verification process can be used. Recommended modifications include using higher-strength compound or more active fastener material or coating. Such modifications have been found to increase cured break-away torque measurements thereby providing separation between the uncured and cured break-away torque measurements.
Dynamic Loading Tests
Tests were performed to assess the effect of dynamic loading and combined verification torque application and dynamic loading on break-away torque measurements. Dynamic loading is provided with the NASM1312-7 test . This test in this study consists of 30,000 cycles of repeated impacts between the spools and slotted frame at a rate of 30 Hz.
Eighteen plain steel fasteners with medium-strength compound were used to assess the effect of dynamic loading. After assembly and cure time, specimens were subjected to the dynamic loading environment and then break-away torque values measured. These values had a range from 180 to 205 in.-lb with a mean of 193 in.-lb and a standard deviation of 8.0 in.-lb. These data are consistent with the data obtained without dynamic loading suggesting the dynamic loading does not degrade the lock compound.
Additional tests assessed the effect of combined verification torque application and dynamic loading. After assembly and cure time, specimens were subjected to a single application of verification torque followed by a dynamic loading environment and then break-away torque values measured. Using an additional 18 plain steel fasteners with medium-strength compound, a verification torque value of 174 in.-lb was used, and a dynamic loading provided using the NASM1312-7 fixture. None of the specimens failed the verification test. Measured break-away torque values ranged from 175 to 206 in.-lb with a mean of 190 in.-lb and a standard deviation of 10.5 in.-lb. These data are consistent with the data obtained with and without dynamic loading indicating the combined application of a direct verification torque test and dynamic loading does not degrade the lock compound.
Process for Implementation
Thread lock compounds should not be used in critical applications without establishing a well-defined process to verify performance. An objective of this study was to provide a method for implementing direct verification of such compounds in practice. The application of a torque after assembly and cure time is found to provide a viable direct verification test. A step-by-step process to facilitate the selection of a conservative value for verification torque is outlined below. This is based on sample break-away torque measurements with cured and uncured lock compound performed on a representative joint configuration. This process identifies joint configurations that are viable for the direct verification method. Such cases coincide with good secondary locking. The process provides practical modification options to improve secondary locking for joint configurations where locking and the method are not viable.
Obtaining samples of representative product (fasteners, compound, and assembly components);
Defining assembly procedure (cleaning, application of compound, tooling, preload, and cure time);
Assembling and measuring break-away torque from samples (10–30) with no cure time;
Assembling and measuring break-away torque from samples (10–30) with defined cure time; and
Plot distributions of samples from steps “c” and “d”
- Separation between distributions?
- a.Yes: Select test verification torque between distributions
Higher torque is more conservative (less uncured pass, more cured fail)
Lower torque is less conservative (more uncured pass, less cured fail)
No: Modification required (e.g., use primer, higher-strength compound, different fastener materials or coatings, increase of cure time) and start over at step “a”
This article presented a method for direct verification of thread lock compounds and adhesives. It is based on the application of a verification torque in the loosening direction after a specified cure time. A process was provided to facilitate the selection of a conservative verification torque. This is based on sample break-away torque measurements with cured and uncured lock compound performed on a representative joint configuration (e.g., specific fasteners, lock compound, joint components, tooling, and preload). This process identifies joint configurations that are viable for the direct verification method. Such cases coincide with good secondary locking. The process provides practical modification options to improve secondary locking for joint configurations where locking and the method are not viable. These modifications include use of higher-strength lock compound, use of primer, increased cure time, and use of more active fastener material or coating. To illustrate the method and potential issues, this article presented torque measurements and statistical analyses for two different sample joint configurations. The application of direct verification torque did not degrade the locking performance of thread lock compound. The application of a direct verification torque provides a nondestructive assessment of thread compound cure and locking performance using standard tooling.
- 1.CxP 70135: Constellation Program Structural Design and Verification Requirements, NASA (2008)Google Scholar
- 2.Light, D.J.: Vibration loosening of threaded fastenings. CME 30, 56–60 (1983)Google Scholar
- 4.Haviland, G.S.: Machinery Adhesives for Locking, Retaining and Sealing. Dekker, New York (1986)Google Scholar
- 5.NASM1312-7: Fastener Test Methods, Method 7, Vibration, National Aerospace Standard (1997)Google Scholar