This report, authored by Jon Ness, PE, of Matrix Engineering Consultants and Gerald Robinson of Lawrence Berkeley National Laboratory, provides a comprehensive technical guide to the two fundamental mechanisms of bolted joint loosening in solar PV mounting systems: relaxation loosening (non-rotational) and self-loosening (rotational). The PV industry has long reported anecdotal accounts of chronic fastener loosening — joints that fail to maintain preload despite multiple retightening attempts, but until DOE-SETO-funded targeted research, there was no systematic framework for understanding why these failures occur or how to prevent them.
Loose fasteners in PV mounting systems have serious implications for structural reliability, electrical bonding, and long-term safety. DOE-SETO-funded structured interviews spanning 17,000 systems and over 94 GW of installed capacity found that nearly 44% of reported bolted joint failures were attributed to loosening, and that the vast majority occurred at loads well below design wind and snow loads. Only 13% were traced to installer error, while 37% stemmed from design-related issues. This report provides the technical foundation that mounting system designers, field engineers, and site owners need to specify, design, and install bolted joints that resist loosening over the life of the structure.
What the report covers:
- Relaxation loosening through yielding of joint components — Improperly designed or assembled joints can loosen due to permanent deformation of one or more components. Common causes include missing washers (concentrating compressive stress beneath the small bearing surface of a hex-head bolt or nut, often exceeding the yield strength of aluminum or structural steel), improperly specified washers (unhardened washers that yield under bearing stress, or thin washers that bend over oversized holes and slots), and soft joints (joints with air gaps where thin clamped members yield under pretension or service loads). The report also addresses the use of star washers as improvised bonding devices in structurally bonded joints, where field investigations have shown that bolt pretension concentrates at the small contact area of the star washer protrusions, causing compressive yield of the aluminum module frame and chronic relaxation that cannot be resolved by retightening.
- Relaxation loosening through fretting wear — When a bolted joint repeatedly slips under dynamic loading, the relative motion between clamped surfaces under high contact pressure wears down surface asperities at the faying interface. This fretting wear can remove protective coatings and base material, leading to significant pretension loss. The solution is to design and assemble joints with sufficient pretension to prevent slip at the faying surfaces under expected dynamic loads — eliminating the primary cause of fretting.
- Relaxation loosening through embedment — Even in properly designed and assembled joints, microscopic surface asperities plastically deform and flatten when tightened, causing a slight reduction in clamped thickness known as embedment. This is a normal and expected phenomenon, but it causes pretension loss that should be accounted for during design. The report provides calculations using VDI 2230 guidelines and presents a case study comparing a module top-down clamp joint (L/D ratio of 5, approximately 22% pretension loss from embedment) with a module through-bolted joint (L/D ratio of 0.5, approximately 78% pretension loss), demonstrating that short, stiff bolts are far more sensitive to embedment than longer, more flexible ones.
- Relaxation loosening through differential thermal contraction — PV mounting systems commonly join dissimilar metals with different coefficients of thermal expansion. Aluminum components (22.1 × 10⁻⁶/°C) expand and contract more than stainless steel (17.8 × 10⁻⁶/°C) or alloy steel (12.3 × 10⁻⁶/°C) fasteners. When temperatures drop below the installation temperature, pretension decreases. A case study shows that a 300-series stainless steel bolt may lose approximately 4% of pretension when the temperature drops from 32°C to -20°C, while an alloy steel bolt may lose nearly 11%. Although modest individually, these losses are cumulative with other loosening mechanisms.
- Strategies for minimizing relaxation — The report details two primary approaches to reducing bolt sensitivity to embedment and thermal effects: increasing the L/D ratio of the bolted joint (using longer clamped lengths or spacers to reduce bolt stiffness) and incorporating Belleville springs (conical disc washers that compress flat at a known spring rate and then expand slightly if pretension drops, acting as a mechanical buffer that compensates for dimensional changes). Belleville springs also serve as a visual inspection tool — when flat, proper pretension has been achieved; when visibly conical, the joint has loosened.
- Self-loosening (rotational loosening) and the role of friction — Self-loosening occurs when a bolted joint repeatedly slips in shear, momentarily reducing friction in the threads and under the bolt head or nut, allowing incremental counter-rotation that accumulates over slip cycles into dramatic pretension loss. The report traces this mechanism to the foundational research of Gerhard Junker, whose work demonstrated that transverse slip — not axial loading or vibration per se — is the root cause. Friction governs both the initial pretension achieved during tightening and the joint’s resistance to slip and self-loosening in service.
- The Junker test and comparative performance data — The Junker test subjects a pretensioned joint to repeated transverse displacement while monitoring pretension decay. A fastener or locking device is considered resistant to self-loosening if it retains at least 80% of its initial pretension after 2,000 cycles. The report presents compiled Junker test data showing that standard nuts, helical spring washers, and prevailing torque lock nuts lose pretension rapidly under slip conditions, while lockbolts, high-strength pre-applied thread adhesives, and wedge lock washers demonstrate significantly better retention.
- Devices effective at resisting self-loosening — Lockbolts (two-piece pin-and-collar fasteners that do not rely on torque measurement and are inherently immune to rotational loosening), pre-applied microencapsulated thread adhesives (factory-applied coatings that activate upon assembly and bond internal to external threads), and wedge lock washers (paired hardened washers with cam faces and serrations that lock the fastener through a wedging action). The report notes that while these devices resist self-loosening, they do not prevent joint slip itself, and joints that repeatedly slip remain susceptible to fretting wear and fatigue failure.
- Devices ineffective at resisting self-loosening — Helical spring washers (lock washers), serrated flange bolts and nuts, and prevailing torque lock nuts (both nylon-insert and all-metal types) are all widely specified in the PV industry under the assumption that they prevent loosening. Junker testing has shown all three to be ineffective at maintaining pretension when the joint repeatedly slips. Lock nuts do offer one practical benefit: while they will not keep a joint tight, they generally prevent the nut from completely detaching, reducing the risk of module detachment or structural separation. The report also addresses degradation risks specific to nylon-insert lock nuts, whose nylon inserts become brittle under UV exposure and low temperatures.
- Top 10 tips for avoiding loosening in PV bolted joints — The report concludes with a consolidated set of design and assembly recommendations: design for high pretension to prevent joint separation or slip, match hole and slot geometry to bolt size per accepted standards, specify washer hardness and thickness correctly, avoid star washers for electrical bonding in dynamically loaded joints, eliminate or reinforce soft joints, account for embedment and thermal contraction during design, use torque stripes for quality control and field monitoring, avoid reliance on helical spring washers and lock nuts for self-loosening prevention, and use thread locking devices judiciously while recognizing they do not address the root cause of joint slip.
The central message is that bolted joint loosening in PV mounting systems is not primarily caused by installer error or extreme weather — it is overwhelmingly a design and specification problem. Relaxation and self-loosening are well-understood phenomena in mechanical and structural engineering, and the tools to prevent them exist. Applying these fundamentals systematically — through proper joint geometry, fastener and washer specification, pretension design, and appropriate locking strategies — will reduce O&M costs, prevent system downtime, and improve structural safety across the growing fleet of installed PV systems.
Jon Ness
PE, PMP, NPDPJon is a Managing Governor and Principal Engineer at Matrix Engineering and has over 34 years of experience in business and engineering team leadership. His career has been focused on the development of off-highway equipment and powertrains. He has unique technical expertise in designing and validating dynamically loaded bolted joints. In his consulting role, Jon has led numerous joint failure investigations, including re-design efforts to mitigate risks to the system owners. He actively participates in ongoing research projects and has taught many classes related to Failure Modes and Effects Analysis and Bolted Joint Design and Validation. He received a Bachelor of Science in mechanical engineering from South Dakota State University. A licensed engineer in Minnesota, Jon is an active member of the UL2703 Standards Technical Panel, a contributor to the ASCE Manual of Practice for Solar PV Structures, and a Certified Fastener Specialist through the Fastener Training Institute.