There always seem to be questions about hi-pot testing. Maybe I can present some of those questions and their answers.
Some Definitions
What is hi-pot? This is an acronym for high-potential.
What is dielectric withstand? This is a shorthand phrase for dielectric withstand (or withstanding) voltage. It is a voltage which a dielectric material (insulator) will withstand without breaking down.
What is electric strength? This is nearly the same as dielectric withstand. It is the highest voltage at which the specific insulation will not break down.
What is breakdown? This is the failure of an insulator due to the voltage impressed across it.
What is the hi-pot test?
The hi-pot test applies a relatively high voltage between the primary circuits and the protective grounding conductor. For two-wire products, the high voltage is applied between the primary circuits and accessible conductive parts (or foil wrapped around accessible non-conductive parts).
What is the safety purpose of this test?
To answer this question, we need to (1) identify what part fails when the product fails the test and (2) identify the safety consequences of that part failure.
Since we are applying a voltage between the primary circuits and the grounding circuit (or accessible conductive parts), the part we are testing is insulation. The insulation between any point of the primary circuit and the grounding circuit is either solid or air, or both solid and air in series.
In the event of a hi-pot failure, there is a failure of either the solid insulation or the air insulation.
If the failure is solid insulation, then a conducting path is impressed upon the surface or through the solid insulation, and the insulation is destroyed catastrophically. The material is no longer an insulation, but is now a resistor of indeterminate value. The resistance may be sufficiently low value to allow an electric shock to occur from grounded parts (in the absence of ground) or from accessible conductive parts.
If the failure is air insulation, then a conducting path exists for the duration of the test. When the high voltage is turned off, the system returns to normal because air is a renewable insulation. A shock could exist for the duration of a primary-circuit overvoltage.
(Interestingly, safety standards authorities and certification house authorities commonly do not permit breakdown of air insulation even though, in practice, the breakdown only exists for the duration of the overvoltage. Following the arc event, the air in the clearance is renewed and the construction is returned to its pre-breakdown state.)
The purpose of the dielectric withstand (hi-pot) test is to determine whether the insulation from the primary circuit to grounded or accessible parts has sufficient electric strength to withstand the normal overvoltages which could occur in service.
Why is the test voltage so high, i.e., more than 10 times the rated input voltage?
Electric current in an inductor creates a magnetic field. When the current is switched off, the magnetic field collapses, generating a current in the opposite direction. This current can generate a very high voltage impulse into the power distribution system. Such impulses are entirely normal (e.g., during the starting process of an electric motor).
Because high-voltage impulses are impressed upon the power line, all insulations on a power distribution system (including product internal insulations) must have sufficient electric strength to withstand not only the normal system operating voltage, but also the normal system overvoltages, i.e., the high voltage impulses. Consequently, product primary-to-ground insulations must be tested with a high voltage to confirm that the insulations will not break down under normal conditions. Note that normal conditions include high-voltage impulses.
Many studies of power line overvoltages have been published. As a general rule, overvoltages on 120-volt systems are less than 1000 volts peak. Overvoltages on 230-volt systems are less than 1500 volts peak.
How does voltage cause insulation to fail?
Air is the culprit behind virtually all insulation failures.
Air is either an insulator or a conductor, depending on the voltage impressed across a specific distance. In very round numbers, and for the sake of discussion, we will assume that air can withstand about 1000 volts per millimeter. At voltages above 1000 volts per millimeter, air will break down, and an arc will occur.
Likewise solid materials are either insulators or conductors, depending on the voltage impressed across their thickness. For the sake of discussion, we will assume that solid insulation can withstand about 10,000 volts per millimeter. However, the breakdown of solid insulation usually involves the breakdown of air.
Consider a solid insulation 1 millimeter thick with 10,000 volts impressed across it. If air was trapped in the solid insulation, then it is possible that the air would be subject to more than 1000 volts per millimeter. If this occurs then the air within the solid insulation will break down and an arc will occur. The temperature of the arc could cause a carbon path in the void where the air was trapped. This effectively shorts out a small part of the insulation, thus increasing the volts per millimeter of the remaining insulation. This could cascade to other sites where air is trapped, and a catastrophic insulation failure occurs.
A good solid insulator is characterized by having virtually no air trapped within the solid material.
Air entrapment within a solid material is not the only solid insulation failure mechanism. If we provide several layers of thin insulation such as in a transformer or a capacitor, air will be trapped between the layers. The same failure sequence can occur.
How do you determine if a failure has occurred?
A hi-pot failure occurs when there is an arc in a clearance, or an arc which damages solid insulation.
Various third-party safety certification houses have made many pronouncements as to what constitutes a hi-pot failure. There are two popular ones: (1) the voltmeter does not increase linearly as the voltage is increased, and (2) the ammeter increases non-linearly as the voltage is increased.
These are both reasonable indicators that the insulator is not behaving properly. Such meter behavior can be due to the air within an insulation system breaking down in small pockets, but not yet resulting in a complete breakdown of the system.
We should not rely on the meter readings as always representing an insulation breakdown. Always find the point of arcing and the point of breakdown. This may mean disassembling the product or component, and repeating the test on the remaining assembly and on the individual parts.
If the hi-pot tester is the collapsing field type, and the unit under test requires a high test current, the tester can indicate a non-linear change in voltage or current yet the unit under test may not have incurred a breakdown.
The best way to tell if a failure has occurred is to look for the arc, or for evidence of arcing. Look and listen for the ZAP!
If part of the insulation breaks down due to failure of a small number of air pockets, does this constitute a breakdown? Clearly some of the insulation has been damaged. However, if the insulation still passes the hi-pot test, then the system is okay. (This scenario
can occur!)
Which is better, AC or DC?
Some say because ac more readily ionizes air than dc, ac is a more sensitive test. Some say because ac more readily ionizes air than dc, dc is a less stressful test.
At the common test voltages used for electronic products, in the range of 1000 volts to 3500 volts rms or so, there is no conclusive evidence that either is the better test.
At what current should I set the hi-pot trip point?
It doesn’t matter.
In the old days, hi-pot testers had no current trip. The only way you could tell a failure (other than hearing the ZAP!) was that the voltmeter would fail to advance to full voltage. Some hi-pot testers also had ammeters, in which case you knew you had a failure when the ammeter needle pinned to the high end of the scale.
Having said that, there is a minimum value of current that the test requires to avoid nuisance tripping.
The minimum value of hi-pot test current is proportional to measured leakage current.
If you are doing an ac hi-pot test, then the hi-pot current will be proportional to the ac leakage current. If you are doing a dc hi-pot test, then the hi-pot current will be proportional to the dc leakage current. Specifically:
You can also calculate the hi-pot current from the capacitive reactance of all the Y-capacitors in the primary circuit. You will need to add some capacitance to account for the primary-to-ground capacitance of the mains circuits and transformer.
The hi-pot trip current must be set at some value greater than the actual current required for the hi-pot test. The value must account for the component and manufacturing tolerances contributing to leakage current (the tolerances of the Y capacitors and the variation in primary-to-ground wiring capacitances).
For example, if the 250-volt leakage current is 2.0 milliamperes, then the hi-pot current for a 1500-volt hi-pot test would be:
You would set your hi-pot trip current at, say, 15 milliamperes to account for component and manufacturing tolerances.
Some people set the trip current at a value to find wrong-value EMC filter capacitors. I don’t believe it is very useful to use the hi-pot test to discriminate against excessive primary-to-ground capacitance.
Having said all this, I now say that the actual value of the trip current is not important.
In manufacturing, the purpose of the hi-pot test is to find gross manufacturing errors (which are catastrophic hi-pot failures). Gross manufacturing errors are indeed gross, and usually show up at much lower voltages and much higher currents than those required for the test.
Except for BABT, the various safety certification houses do not specify hi-pot trip current limits.
What are the certification house requirements for hi-pot testers?
Some certification houses require the hi-pot tester to include a manual reset following a hi-pot failure. Some certification houses require visual, audible, or both indications of a failure.
These requirements negate automated hi-pot testing on an otherwise automated production line.
What is the effect of humidity?
In some cases, a required hi-pot test follows humidity treatment. The purpose of this sequence is to find hygroscopic (water absorbent) insulating materials.
Supposedly, the humidity treatment does not include dew-point conditions. If this is the case, then the air trapped within the insulating materials can be assumed to have relatively high humidity as a result of the humidity treatment. The assumption is that humid air has less dielectric strength than dry air.
This is a false assumption. Humid air actually has equal or greater electric strength than dry air! Water vapor is a gas, and does not have the same properties as water in the liquid state.
Copyright 1996 by Richard Nute Originally published in the Product Safety Newsletter, Vol. 9, No. 4, October – Dcecember, 1996
