Permanent set in springs: why a spring loses length and load

The worst surprise in a spring design shows up after a few cycles of use: the part comes back shorter than it left the factory and delivers less force at the same installed height. It is not wear or corrosion — it is permanent deformation, known in the trade as set. The spring was stressed beyond the elastic limit of the material and part of the wire simply did not spring back.

Understanding set means understanding the boundary between elastic behaviour — predictable and reversible — and plastic behaviour, which permanently steals free length and load. In this guide we walk through what permanent set is, the physics behind it, the difference between immediate set and slow relaxation, the deliberate use of set in the presetting process, and finally the design stress limits that keep a spring stable over its life.

What permanent set is

Every spring works by storing elastic energy: you apply a load, the wire deforms, and when the load is removed the spring gives the energy back and returns to exactly its original free length. That back-and-forth is only perfect while the stress in the material stays below the elastic limit. Past that limit, a fraction of the wire yields — deforms plastically — and does not come back. The spring takes on a new free length, shorter than designed. That permanent loss is permanent set.

The practical effect has two faces. First, the spring loses free length: a part designed at 55 mm may settle at 52 or 51 mm after a single overload. Second, and more important for the assembly, at the same installed height the spring now delivers less load, because part of the travel that would generate force has been consumed by plastic deformation. In a valve, a latch or a return mechanism, that means insufficient force and functional failure, even with no visible crack.

It is important to separate set from other failure modes. There is no loss of material, no fracture and often no external sign of damage. The spring has simply become shorter. That is why set is frequently mistaken for a tired spring, when it is in fact the direct result of a design stress that was too high for the chosen material.

The physics: why the wire yields in torsion

Although the spring carries an axial load, the wire works neither in pure tension nor in pure compression: it works essentially in torsion. When a helical spring is compressed or extended, each section of the wire is twisted about its own axis. And in the torsion of a circular section the shear stress is not uniform: it is zero at the centre of the wire and maximum at the outer surface. It is the outermost fibres of the wire that feel the highest stress.

The nominal shear stress (without the Wahl curvature correction) is given by the relation below, where F is the load, D is the mean coil diameter and d is the wire diameter.

Yielding of the outer layer of the wire

As long as this stress stays below the material's elastic limit in shear, the whole section deforms elastically and recovers. When the load grows and the surface stress passes the torsional elastic limit, the outer layer of the wire yields first, while the less-stressed core is still elastic. On unloading, the core tries to return to its original position, but the already plastically deformed outer layer prevents full recovery. The result is a residual deformation — the set — locked into the wire as a small permanent twist spread over every coil.

This is why set always starts at the wire surface, and why the wire diameter is so decisive: because stress grows with the inverse cube of the diameter, a slightly thinner wire raises the surface stress quickly and brings yielding forward. A spring at the edge of set usually reveals the problem on the very first severe compression, not gradually.

Immediate set versus relaxation (creep)

Two different phenomena reduce a spring's force over time, and confusing them leads to wrong diagnoses. Set is immediate: it happens on the first load that exceeds the elastic limit. A single overstress event — compressing the spring solid once, for example — is enough for permanent deformation to set in. After that, if the spring goes back to working at lower stresses, it is stable at the new length and loses nothing more.

Relaxation (or creep) is the opposite in pace: it is a slow, progressive loss of load that occurs even with the stress below the short-term elastic limit, as long as the spring stays loaded for a long time. Relaxation depends strongly on time and, above all, on temperature. A spring held compressed at room temperature relaxes very little; the same spring held compressed at 150 or 200 °C can lose a significant share of its load in weeks or months. It is the dominant mode in engine springs, appliances and any application that combines sustained load with heat.

The main causes of permanent deformation are:

• Working or solid stress above the material's elastic limit (immediate set).
• Accidental compression to solid height in a spring not prepared for it.
• Sustained load over long periods, especially combined with elevated temperature (relaxation).
• A material whose elastic limit is too low for the design stress level.
• No stress-relief treatment after coiling.

Presetting (scragging): using set to your advantage

Set is not always a defect — used on purpose, it is one of the most powerful tools in the spring designer's kit. The process is called presetting, or scragging. The idea is simple and clever: the spring is deliberately made a little longer than its final free length and, just once, is compressed beyond its service point, often all the way to solid height.

In that first controlled compression, the wire surface yields exactly as we described — the spring shortens and settles at the desired free length. The gain is in what stays behind: the outer layer, having been plastically deformed and then unloaded, retains residual stresses of the opposite sign to the working stresses. In service, those residual stresses subtract from the applied stress, lowering the net stress at the wire surface.

The effect is remarkable. A preset spring has already taken all the set it was going to take; in normal service it no longer yields, even working at stress levels that would make an unprepared spring take set. In practice, presetting lets the designer use higher service stresses — and therefore smaller, lighter springs for the same load — safely. That is why presetting is routine in high-performance springs.

From a manufacturing standpoint, presetting is done after coiling and heat treatment. The final free length is set, and where needed fractions of a coil are removed to trim the load, all accounting for the expected shortening. Keep in mind that the benefit is directional: a spring preset in one direction should not be loaded in the opposite direction, because there the residual stresses would add instead of subtract.

Design stress limits that prevent set

When a spring cannot or will not go through presetting, the only way to guarantee dimensional stability is to keep the design stress below the set limit. The most widely used design rule is conservative and direct: for springs that must take no permanent deformation, the nominal shear stress (without the Wahl correction) in the worst condition — usually solid height — should stay below about 45% of the material's tensile strength.

Applying the limit in practice

Rm is the tensile strength of the wire, which for high-carbon spring steel depends strongly on diameter. Below that stress level, the wire surface stays elastic even with the spring fully compressed, and the free length is preserved. Above it, you must either reduce the stress — by thickening the wire, increasing the mean diameter or the number of coils — or specify presetting to tolerate the higher stress. To avoid unwanted set, the design should:

• Keep the nominal stress at solid height below roughly 45% of the tensile strength when there is no presetting.
• Specify presetting (scragging) when higher service stresses are required.
• Choose materials with an adequate elastic limit and apply stress relief after coiling.
• Limit the operating temperature to contain relaxation under sustained load.
• Prevent the spring from accidentally bottoming out to solid height during the working stroke.

The role of heat and temperature

Temperature acts on permanent deformation in two opposite ways, and it is worth telling them apart. In manufacturing, heat is an ally: the stress-relief treatment, done at a few hundred degrees for a controlled time after coiling, removes unwanted residual stresses from forming and stabilises the spring. In preset springs, a heat treatment after presetting can further lock in the beneficial residual stresses, making the effect more durable.

In service, though, temperature is an adversary. The material's elastic limit and modulus drop as temperature rises, and relaxation accelerates sharply. A design stress that is perfectly safe at 20 °C can lead to unacceptable load loss at 180 °C held for months. That is why hot applications demand a larger stress margin and often specific materials — such as chrome-silicon steel or heat-resistant alloys — that preserve the elastic limit at temperature.

The rule of thumb is clear: the higher the working temperature and the longer the time under load, the lower the design stress must be to keep the same dimensional stability. Ignoring this is the number-one cause of springs that “lose force” in service with no visible sign of damage.

Worked example: when set shows up

Let us put numbers to it. Consider a music-wire compression spring with wire diameter d = 3 mm, outer diameter 22 mm — so mean diameter D = 19 mm — free length 55 mm and 7 total coils (5 active, closed and ground ends). The tensile strength of this wire is on the order of Rm ≈ 1900 MPa, and the resulting spring rate is about 24 N/mm.

In the service condition, the spring is compressed 18 mm, to a working height of 37 mm, requiring a load of about 430 N. The nominal stress in that condition is:

Reading the numbers of the example

A stress of 770 MPa corresponds to about 40% of the tensile strength — below the 45% limit — so set is negligible and the spring keeps its 55 mm indefinitely. It is a design that is safe against permanent deformation.

Now picture the same spring accidentally compressed to solid height, with a deflection of 34 mm and a load of about 815 N. The nominal stress jumps to roughly 1460 MPa, or 77% of the tensile strength — well above the set limit. The wire surface yields and, on unloading, the spring does not return to 55 mm: it settles a few millimetres shorter, typically 2 to 4 mm less free length, with the corresponding loss of load in service.

The example teaches two things. If this spring really must reach solid in service, it should be preset — and then the 1460 MPa is no longer a problem, because the set was already consumed in manufacturing. If it must take no set at all and will not be preset, the design has to lower the solid stress, typically by thickening the wire or adding coils, until the solid stress falls below 45% of the tensile strength.

Designing springs that hold their length

The good news is that all of this control is systematic, not a matter of luck. On molas.app.br, every configured spring has its nominal stress computed at the working height and at solid height and compared with the limits of the selected material. When the solid stress approaches the range where set appears, the design is flagged, so you adjust the wire diameter, the mean diameter or the number of coils — or choose to specify presetting — before manufacturing. That way the spring leaves with its design stress inside the safe margin and holds its free length and load throughout its service life.