Nail Penetration vs Crush Test: Which Battery Abuse Test Do You Need?

The nail penetration vs crush test question comes up in almost every battery abuse equipment inquiry we handle: which one do we actually need? The two are often quoted in the same sentence, sometimes by the same salesperson, as if one machine could cover both requirements. They test different failure modes, they are governed by different clauses in different standards, and the machines differ in force capacity by a factor of five or more.

We build both, and we also build combined machines where a customer genuinely needs both tests on one frame. This comparison is based on what we see in orders, factory acceptance tests, and the questions customers ask after they have been quoted by three suppliers and gotten three different answers.

Nail Penetration vs Crush Test: What Is the Difference?

The short version: nail penetration drives a conductive spike through the cell to create a localized internal short circuit. Crush squeezes the cell or pack between flat plates to simulate distributed mechanical damage. One concentrates force on a few square millimeters, the other spreads it over the whole contact area.

If you only take one line from this article: UN 38.3 requires crush, not nail penetration. We have written separately about how the eight UN 38.3 tests map to machines, and T6 is a crush test. Nail penetration appears in other standards and in customer specifications, not in UN 38.3.

What Does Each Test Actually Simulate?

Nail penetration: a single point of intrusion

A needle goes into the cell, shorting the layers it passes through. The current path is small, the resistance is low, and the heat is concentrated exactly where the needle is. In a cell that fails, the sequence is usually fast: local heating near the puncture, electrolyte vaporization, venting, and in a bad case ignition of the vented gas.

In the field this maps to a dropped tool during assembly, a metal fragment inside a pack, debris penetrating a enclosure in a crash, or a deliberate abuse test for a customer who wants to know how a cell behaves when a single point shorts internally. It is also the test most closely associated with the phrase "internal short circuit" in battery literature.

What makes nail penetration useful is also what makes it hard to reproduce. The result depends on needle diameter, tip geometry, penetration speed, penetration depth, and where on the cell the needle enters. Change the needle from 3 mm to 8 mm and a cell that passed can fail, or the reverse. This is why the test method matters as much as the machine, and why a report that says "nail penetration passed" without stating the needle and speed tells you very little.

Crush test: distributed mechanical damage

A flat platen presses on the cell or pack until a force or deformation limit is reached, then holds. The load spreads across the face of the cell, compressing the electrode stack and the separator over a wide area. Failure usually comes from separator compression or puncture at the edges, or from the casing deforming enough to bring the layers into contact.

In the field this maps to a vehicle crash where the pack is squeezed, a module dropped on a corner, a stack of cells loaded in a container, or equipment rolling over a pack. Crush is the test that asks whether the mechanical design of the cell or pack can survive being squashed without going into thermal runaway.

Crush is easier to standardize than nail penetration because the geometry is simple: two flat plates, a defined force or a defined deformation, a defined hold time. That is why the standards bodies use it, and why UN 38.3 picked it for T6 rather than a needle test.

How Are the Two Tests Run?

The nail penetration method

A typical procedure fixes the cell on a fixture, positions a steel needle perpendicular to the largest face, and drives it through at a controlled speed. Common parameters:

• Needle diameter: 3 mm, 5 mm or 8 mm, sometimes with a 45 to 60 degree conical tip

• Speed: 0.1 to 100 mm/s; 8 mm/s and 10 mm/s are common in customer specifications

• Depth: through the cell, or to a defined depth for partial penetration studies

• Force capacity: 5 kN covers most small cells; 20 kN is the practical standard for prismatic and large-format cells

• Instrumentation: cell surface thermocouple, voltage tap, sometimes a camera behind a protective window

The machine has to do two things well that are easy to get wrong. It must hold the speed constant as the needle meets resistance, and it must record force and displacement so the report can show what happened. A pneumatic press that punches through in a fraction of a second is not a nail penetration tester in any meaningful sense, even if it has a needle on the end.

Our own battery nail penetration test machine guide covers the specification side in more detail, including why the speed range matters more than the headline force.

The crush method

Two flat steel platens close on the cell or pack. The procedure defines the force, the rate, the hold time and the pass criteria. For UN 38.3 T6 the force is 13 kN, held for 10 minutes, with the cell observed for six hours afterwards; a deformation limit is used as an alternative acceptance route in some procedures. For electric vehicle packs, GB 38031 asks for a much higher force, up to 100 kN, or a defined percentage of deformation, whichever comes first.

Which Standards Require Which Test?

This table is the part most quotation requests get wrong. Read the "required" column carefully, because several popular claims in the industry do not match the current text of the standards.

The GB 38031 mistake worth knowing

Many supplier pages still say nail penetration is mandatory under GB 38031. That was true of the 2015 edition. The 2020 revision removed the cell-level nail penetration requirement and replaced the emphasis with a system-level thermal propagation test, which requires a warning before the fire spreads and gives occupants time to exit.

This matters to buyers in two ways. If a Chinese customer asks you to run nail penetration and cites GB 38031, ask which edition they mean, because the answer changes what you have to buy. And if you are buying a machine mainly to satisfy "the Chinese EV standard", a nail tester may not be the machine that clause actually requires.

What Does Each Test Reveal That the Other Misses?

Nail penetration answers a narrow question very well: what happens when one small area of the cell is shorted through the separator. It stresses the separator's puncture resistance, the current collection path around the needle, and the cell's ability to survive a hot spot without cascading. It is a good screen for cell chemistry and winding or stacking design.

Crush answers a different question: what happens when the whole cell or pack is loaded mechanically. It stresses the casing, the electrode stack compression behavior, the weld and tab design, and the pack structure. A cell can pass a 13 kN crush and still fail a needle test, and the reverse also happens. They are not substitutes.

Neither test predicts thermal propagation in a finished pack. A single-cell nail test tells you the cell vents and possibly ignites; it does not tell you whether the module design contains that event or lets it spread to the next cell. That is a separate system-level test with its own equipment and its own standard, and it is where most of the difficult engineering now sits.

Can One Machine Run Both Tests?

Yes, and for many laboratories it is the sensible purchase. A servo-driven frame with a 20 kN or 50 kN capacity, a platen set and a needle fixture can run both methods, and the same controller records force and displacement for either. What you cannot do is buy a small nail tester and add a crush platen rated for pack-level force later.

Three configurations cover most of the market:

Integration with a chamber

Both tests are frequently run at elevated temperature, and sometimes inside an environmental chamber so the cell starts at a defined condition. That changes the machine design: the frame has to fit inside or through the chamber wall, the load cell has to survive the temperature, and the cables have to pass through a sealed port. It also changes the safety case, because a chamber full of vented electrolyte is a different risk from a chamber full of air.

We build abuse machines that sit inside our environmental chambers, and the usual failure in competitor designs is thermal drift in the load cell. A 20 kN cell that reads 2% high at 60 °C will make a passing test look like a failing one, or the reverse. Ask for the temperature compensation specification, not just the accuracy at 23 °C.

Safety equipment that is not optional

Both tests can produce fire, vented gas and flying debris. A serious installation includes an extraction hood or duct, a fire suppression option, a reinforced viewing window, and remote operation so nobody stands next to the cell during the test. On pack-level crush systems, the fixture design and the exclusion zone matter as much as the press. If a quotation includes none of this, it is a quotation for a press, not for a battery abuse test system.

Which Test Should You Buy First?

The answer depends on what you sell and who audits you. Some practical patterns from our order book:

Cells for export under UN 38.3

You need crush. A 20 kN servo frame with flat platens covers T6, and the same frame will run nail penetration later if a customer asks. Buy the 20 kN frame rather than a 5 kN nail-only press, because the marginal cost is small and the flexibility is not.

Cells for Chinese customers or domestic certification

Crush plus nail penetration, because customer specifications and the older GB/T 31485 habit still drive requests even where GB 38031 no longer requires the cell nail test. Expect the nail parameters to be specified by the customer, and expect them to differ between customers.

EV packs and modules

Crush at 100 kN or above, with pack-sized fixtures and a proper exclusion zone. A cell-level nail tester is not the machine for this work, and a 20 kN frame will simply not reach the force the standard asks for.

Third-party laboratories

Both, plus the ability to change fixtures quickly. A lab that cannot switch from a needle to a platen in under an hour will lose bookings, and throughput is the whole business model. Look at fixture changeover time as a specification, not an afterthought.

What Belongs in the RFQ?

Send these items and you will get comparable quotations instead of three different machines described in similar words:

1. Test methods and clause numbers you must satisfy, with the edition year of each standard

2. Maximum force you need now, and the largest product you expect within three years

3. Sample range: cell formats, dimensions, pack sizes and weights

4. Speed range and control: constant speed under load, recording rate, force and displacement curves

5. Temperature condition: ambient only, or inside a chamber at a defined temperature

6. Instrumentation: thermocouples, voltage taps, camera, data export format

7. Safety scope: extraction, suppression, window rating, remote operation, interlock

8. Acceptance: factory test protocol, calibration certificates, witness test by video or in person

Two questions separate a real abuse machine builder from an assembler. Ask what happens to the force reading when the frame warms up under a chamber, and ask for the fixture drawing for your specific cell. A supplier who cannot answer both is selling a generic press with a needle attached.

Nail Penetration vs Crush Test FAQ

Does UN 38.3 require nail penetration testing?

No. UN 38.3 requires a crush test as T6, at 13 kN, and does not include a nail penetration clause. Suppliers who list nail penetration as a UN 38.3 requirement are usually confusing it with customer specifications, SAE J2464, or the older Chinese cell standards.

Can a crush test replace a nail penetration test?

No, they produce different failure modes. Crush loads the whole face of the cell and tends to fail at the edges or through separator compression. A needle creates a small, hot, low-resistance short at one point. A cell can survive one and fail the other, so a crush result says nothing useful about nail performance.

What force capacity should I buy for both tests?

For cells, 20 kN covers nail penetration and UN 38.3 crush with margin. For modules and packs, plan for 100 kN or more, and buy that capacity once rather than upgrading later, because the frame, load cell and drive package cannot be scaled up in the field.

What nail diameter and speed should the machine support?

Support 3, 5 and 8 mm needles with a controlled speed from 0.1 to at least 20 mm/s, ideally to 100 mm/s if you serve automotive customers. The speed must stay constant as the needle meets resistance; a machine that only quotes a maximum speed without stating speed stability under load is not giving you the number that matters.

Do the tests have to run inside a temperature chamber?

Not always, but often. UN 38.3 T6 runs at ambient. Many automotive customer specifications require the cell to be soaked at a defined temperature first. If you expect that work, specify the chamber interface now: frame dimensions, load cell temperature compensation and cable ports are all designed in, not added later. The same reasoning applies when you are comparing nail penetration vs crush test options for a new lab, because chamber integration is usually the line item that gets cut from the first quotation and added back at full price later.

How long does a test take?

The mechanical part is fast: nail penetration is seconds, crush is a few minutes including the hold. The observation period is what takes time, typically six hours for UN 38.3 T6 and up to 24 hours for some customer procedures. Plan bench space and ventilation for the observation, not just for the test.

What safety equipment is required?

At minimum, an extraction system for vented gas, a reinforced viewing window or remote camera, and remote operation. For pack-level work, add fire suppression and a defined exclusion zone. These are not accessories in our quotations; they are part of the system, because the tests are designed to make cells fail.