Hipot Test: The Three Settings That Decide Pass or Fail
Home
Yiwu, Zhejiang, China
Mon – Fri 9:00 AM – 6:00 PM · Sat 9:00 AM – 4:00 PM
[email protected]

What Is a Hipot Test: Three Numbers Decide Pass or Fail

Alan Xiao, Founder of Sourcing Pioneer
Alan Xiao Founder of Sourcing Pioneer
· Updated November 9, 2026 · 8 min read

A hipot test pushes high voltage across an air gap and waits.

The gap is whatever should keep mains power away from the metal a user can touch. If the gap holds, almost nothing flows. If it leaks or breaks down, the tester trips and the unit fails.

The name is short for high potential. The same test is also called a dielectric withstand test, or a voltage proof test. All three names describe one action.

Studio still life of a power adapter with a test lead clipped to its metal chassis
The clip goes on metal a user can touch. The probe goes on the live side.

The Short Answer

A hipot test is a pass or fail check on insulation, run at a voltage far above the working voltage. A tester applies that voltage for a set time and watches the current that leaks through.

If the current stays under a limit, and no arc appears, the unit passes. If the current climbs, or the voltage collapses, the unit fails.

Three settings decide the result, and all three appear in a test plan: the voltage, the time, and the current limit. A spec that names only one of the three is not a spec.

What the test does not do is just as useful to know. It does not prove the product works. It does not measure insulation resistance, which is a separate check at a lower voltage. And it does not tell you how long the unit will last.

What the Tester Actually Does

The setup is simple. One lead goes to the live parts, often joined together. The other lead goes to the metal a user can touch, or to a foil wrapped around the case.

Then the machine raises the voltage, holds it, and measures. A tester of this class offers trip points from 0.5 mA to 25.5 mA in fine steps. Its timer runs from a fraction of a second to minutes.

The judgment has two edges, and the lower one surprises people. A unit fails if the current goes above the upper limit. It also fails if the current stays below the lower limit, because a reading near zero usually means the clip is not touching anything.

A fail looks different on the dial. The current steps up and holds, or the voltage collapses as an arc forms. Either way the tester cuts the output within a fraction of a second. The operator rarely sees the fault itself, only the red lamp.

That lower limit is the cheapest guard against a false pass in a factory. A test with a dangling lead looks clean on paper and proves nothing.

The Three Numbers, and Where They Come From

The voltage is not chosen by the factory. It comes from the safety standard that covers the product, and it changes with the type of insulation.

For household appliances, IEC 60335-1 handles this. Clause 13 covers the test at operating temperature. Clause 16 covers the same test at room temperature, and a table in the standard sets the voltage for each insulation type.

NumberWho sets itWhat happens if it is wrong
VoltageThe standard, by insulation type and rated voltageToo low passes bad units; too high damages good ones
TimeThe standard, and shorter for routine testsToo short misses slow breakdown
Current limitThe product standard, or the factory’s own processToo high hides a leak; too low rejects good units

Public IEC 60335 test reports for appliances at 230 V repeat one line. An electric strength test at 1000 V, between live parts and accessible metal parts. That figure covers basic insulation on a Class I appliance.

Two adjustments follow from that. If the barrier is reinforced rather than basic, the voltage rises. If the working voltage is higher, the voltage rises with it. Both are read from the standard’s table, not invented.

Flat vector panel showing the three settings on a hipot tester: voltage, time and current limit
Three settings, one verdict. A plan that names only the voltage is half a test.

One more setting hides in plain sight: the ramp. A tester that jumps straight to full volts will trip on the inrush of a capacitor that is perfectly healthy. A slow ramp lets the same unit settle before the reading counts.

The current limit is the number buyers ask about least and should ask about most. A tester set at 100 mA will pass a unit that a 5 mA limit would reject. Same product, same voltage, two different verdicts.

Two Tests Wear the Same Name

This is the part a quote usually hides. The same standard describes two different tests, and they answer different questions.

The type test proves the design. It runs at the full voltage for the full minute, on a small number of samples, and it is done once for a certificate.

The routine test checks production. It runs on every unit, at a much shorter time, and it exists to catch a bad batch rather than a bad design.

IEC 60335-1 keeps them apart in the document itself. The full test sits in Clause 16, and a separate annex, Annex A, is titled Routine tests.

Flat vector comparison of a type test on a sample and a routine test on every unit
One proves the design once. The other guards every unit that leaves the line.

The numbers differ. A one-minute hold on a sample is normal for the type test. A one-second hold on every unit is normal for the routine test. The voltage is often raised to keep the product of voltage and time in the same range.

A real example sits in a tester catalogue. The instrument maker prints its own routine test in the safety section. It runs 1390 V AC for 2 seconds, between the mains input and the chassis.

The raised voltage is not a trick. A shorter test needs more stress to expose the same weakness, so the two settings are chosen together rather than one at a time.

So when a supplier says “we hipot every unit”, the useful reply is a question. At what voltage, for how long, and with which current limit. A one-second test at a low limit is not the same claim as a minute at the type-test voltage.

What a Failure Means for the Lot

A failed unit is usually scrap. The test drives voltage through a fault, and the fault often gets worse on the way. Repair is possible on expensive goods, but the insulation has already been stressed once.

The harder question is what one failure says about the other units in the box. The honest answer is: not much on its own.

Three readings are worth separating.

A single failure in a large lot is often a handling fault. A pinched wire, a metal burr, a splash of solder. Inspect the failed unit, find the cause, and check whether the same cause can appear again.

A rising failure rate points at the process. Contamination on a connector, a changed plastic part with less creepage, a new operator on a sealing step. This is the reading that should stop a shipment.

A failure on the first unit off a new tool is a design signal. The gap in the drawing is too small for the process, and no amount of rework will fix the next thousand.

Write the rule down before the first lot ships. A failure rule agreed in advance turns an argument into a procedure. A supplier who knows the count is watched has a reason to report it.

One number turns all three readings into a decision. Ask for the failure rate per lot, not the pass rate. A line that passes 99 of 100 is failing one in a hundred, every day.

What you seeWhat it points atWhat to do
One failure in a large lotA handling faultInspect that unit, find the cause
A rate that keeps risingThe processHold the shipment
First unit off a new toolThe designChange the gap, not the operator

The Test Is Not Free

A hipot test applies a stress the product never meets in normal use. Most units shrug it off. A unit with a thin margin may pass the test and lose part of its life doing so.

That is why the test belongs at the end of the line, once. Repeating it to “be sure” puts the same stress on the same barrier a second time, and a second pass says less than the first.

It also explains a rule that factories sometimes bend. A unit that failed once should not go back on the tester after a quick wipe and come out clean. The first failure is the information. The second result is noise.

The Number You Read Is Not All Product

Part of the leakage current comes from the fixture. A test lead lying in the air is a small capacitor, and at a few kilovolts it passes a little current.

The instrument maker prints the size of that effect. With 350 mm leads hanging in air, the tester’s own reading is about 20 µA at 1 kV and about 100 µA at 5 kV. With the supplied accessory lead, the figures double.

That matters when a spec sets a tight limit. A 100 µA limit on a 5 kV test may be measuring the cable, not the product. Either shorten the leads, use a guarded fixture, or set a limit above the floor the fixture creates. A guarded fixture wraps the high-voltage path in a screen held at the same potential, so the stray current never reaches the meter.

The same trap appears in reverse. A limit set too high hides a real leak, because the product’s contribution is small next to the fixture’s.

AC or DC Changes the Number

Most routine tests are done with alternating current, because that is what the mains does and because a fault shows up fast.

A direct-current test is also used, often for long cables or for products with large capacitors. It has one immediate consequence for the number on the dial.

Alternating current is quoted as a root-mean-square value. The peak is about 1.41 times that. So a DC test at the same stress needs roughly 1.4 times the AC figure.

Two practical points follow. A DC test charges the product’s capacitors, so the unit has to be discharged before anyone touches it. And a unit that passes on DC may not pass on AC, because the two stress the barrier in different ways over time.

Whatever is chosen, it belongs in the order along with the voltage and the time. A supplier who swaps AC for DC to speed up the line has changed the test.

Five Lines for the Order

None of this needs a long document. Five lines cover it, and they fit inside the same one-page brief you already send for every other requirement.

Flat vector card listing five lines for a hipot test requirement in an order
Five lines. Each one turns a verbal promise into something a later claim can test.

One. Name the standard and the clause the voltage comes from, not just the voltage.

Two. Give all three settings: voltage, hold time, current limit.

Three. Say whether this is the type test or the routine test, and how many units it covers.

Four. Say what happens to a failure: quarantine, count, and a written report with the cause.

Five. Say what is recorded per unit, so a later claim can be checked against data rather than memory.

The same logic sits behind the other marks a product carries. A hipot result is one input to a safety file, not a replacement for it. This piece on what a UL mark covers explains. Nor does it replace the workmanship rules that decide whether the solder joint inside survives the stress, which is what IPC-A-610 is for.

For a product going into the EU, the result also has to sit behind a declaration of conformity. The CE and RoHS guide covers how that paperwork is built.

TEST CHECK

Send me the hipot line from your test plan, or the report you were given. I will check whether all three settings are named, whether the voltage traces back to a clause, and what the failure rule says. No cost, no obligation.

Check My Test Plan

Questions Buyers Ask

Is a hipot test the same as an insulation resistance test?

No. Insulation resistance is measured at a low voltage, often 500 V, and reported in megohms. A hipot test applies a much higher voltage and gives a pass or fail. Many benches run both in one sequence.

What voltage should my product be tested at?

It comes from the standard for your product type, and it depends on the insulation class and the working voltage. For a Class I appliance at 230 V, published IEC 60335 reports commonly show 1000 V between live parts and accessible metal.

Can the test damage a good unit?

It applies a stress the product never meets in normal use. Most units are unaffected. A unit with a thin margin can pass and still lose life, which is why the test is done once, at the end of the line.

Does every unit need the test?

The routine test is normally applied to every unit, with a short hold. The full one-minute type test is done on samples for the certificate. The two are separate requirements in the same standard.

What current limit should be set?

It should come from the product standard, and it has to sit above the leakage the fixture itself creates. On a 5 kV test, the leads alone can account for around 100 µA.

Is a DC hipot as good as an AC one?

They stress the barrier differently. DC is common on cables and on units with large capacitors, and the voltage is set about 1.4 times the AC figure to reach a similar peak. A unit that passes one may not pass the other.

Questions & Comments

We read every one and reply within 24 hours

If your test plan names a voltage and nothing else, send me that line. I will say which of the three settings is missing, and what the gap costs you at the bench.

Subscribe
Notify of
guest
0 Comments
Most Voted
Newest Oldest
0
Would love your thoughts, please comment.x
()
x