Hipot vs Insulation Resistance: What Each Proves

Two tests get quoted interchangeably in fan qualification documents, and they
are not interchangeable. One asks whether the insulation system survived being
built. The other asks how clean and how dry the insulation path is right now.
Both produce a number. Only one of them can be repeated on the same unit without
changing the answer.

That distinction matters more for a fan than for most components, because a
fan packs a winding, a lead bundle, a metal frame and a rotating shaft into a
volume the size of a hand — and then puts all of it inside an enclosure that a
customer may later wash, pot, coat or seal. The insulation question is not
academic. It is the one failure mode that presents as a safety event rather than
a performance complaint.

1. Two tests, two different questions

Insulation resistance is a measurement. A known DC voltage is applied across
an insulation path and the resulting current is read, from which a resistance is
derived. The test is non-destructive by design: the voltage is high enough to
polarise the path and expose leakage, but not high enough to break it down. Run
it twice on the same unit, on the same day, and you should get the same answer
back.

Dielectric strength — the hipot test — is not a measurement in that sense. It
is an overstress. A substantially higher voltage is applied for a defined
interval, and the question asked is binary: does the insulation hold, or does it
break down? If it holds, you have learned that the path survived a stress well
above what it will see in service. If it breaks down, you have learned that it
would have failed eventually anyway — and you have also destroyed the unit that
told you so.

So the two tests sit on opposite sides of a line. Insulation resistance
describes a state and can be trended. Dielectric strength proves a margin and
consumes a sample. Treating the second as if it were the first is where most of
the confusion in a qualification package comes from.

2. Insulation resistance: how clean and how dry the path is

What insulation resistance actually responds to is contamination and
moisture, not geometry. Two units built from identical drawings will read
differently if one was handled with bare hands and the other with gloves, if one
sat in a humid warehouse over a weekend, if one was washed and not fully dried,
or if the potting resin in one did not fully cure.

This is why the reading is best understood as a cleanliness indicator rather
than a quality grade. A path that is geometrically sound can read poorly because
of a fingerprint. A path that is geometrically marginal can read well because
the sample happened to be dry. Neither reading is wrong; both are answering the
question that was asked, which is about the state of the surface and the bulk at
the moment of measurement.

The practical consequence: insulation resistance is a process probe. It is
cheap, repeatable and non-destructive, which makes it the right instrument for
watching a production line drift — a change in cleaning chemistry, a change in
resin batch, a change in cure profile, a change in ambient humidity in the
assembly area. A pass stamp tells you one unit was acceptable. A trend tells you
whether the process is still the process you qualified.

3. Dielectric strength: whether the system survives an overstress

The hipot test is aimed at a different set of defects — the ones that are
structural rather than environmental. A nick in the enamel, a wire pinched
against the frame during assembly, a creepage distance that looked adequate on
the drawing but was eaten by a connector body, a void in potting that bridged
two conductors. These do not show up as gradual leakage. They show up as a
breakdown under stress, or they do not show up at all until the field does the
testing.

Because the test is pass-or-fail at a stress level, it has to be calibrated
against the insulation scheme rather than against a habit. Basic insulation,
supplementary insulation and reinforced insulation are three different
constructions, and a voltage chosen for one is not a voltage chosen for another.
The same applies across working-voltage classes: a part specified for a
low-voltage DC bus and a part specified for line AC are not the same test
subject, even if they share a housing and a bearing.

This is the point where a fan catalogue and a fan qualification diverge. The
catalogue describes a family. The test has to be specified per construction,
per insulation scheme, and per the standard the buyer is actually being held to.

4. The paths, not the terminals

The most common defect in a fan test specification is not the voltage. It is
the choice of where to connect the leads.

A fan contains at least three distinct insulation paths, and they fail for
different reasons. Winding to core tests the slot liner and the enamel. Winding
to lead tests the internal splice and the lead exit. Lead to frame tests the
grommet, the bushing and the point where the cable passes the housing — which is
also the point most likely to be disturbed during installation. Potting, where
present, changes all three, but not equally: it reinforces the lead exit far more
than it reinforces the slot.

A test that clips onto whichever two points are easiest to reach measures
whichever path happens to lie between them. If that path is not the path the
application stresses, the result is not evidence — it is a reading. The
specification has to name the path, and the fixture has to reach it.

5. Why the order matters

Insulation resistance is taken before the hipot test. This is not
convention; it is a consequence of what the hipot test can do to a sample.

A unit that passes a hipot test has survived a stress, but surviving is not
the same as being unchanged. Partial discharge, local heating at a thin spot,
and progressive damage in a void can all leave a path that still holds at the
test voltage while reading worse afterwards. If insulation resistance is taken
after the hipot, the second reading is contaminated by the first test, and a
drift that came from the process becomes indistinguishable from a drift that came
from the instrument.

Run in the other order, the two readings stay independent: insulation
resistance describes the unit as built, and dielectric strength describes what it
can survive. Run in the wrong order, you have one number and a story.

6. Type-test voltage is not routine-test voltage

Qualification documents routinely quote a hipot voltage without saying which
kind of test it belongs to, and the two kinds are not the same test at different
sample sizes. They are different tests with different purposes.

A type test establishes that a design meets its insulation requirement. It is
performed on a sample, at a stress that demonstrates margin, and it is allowed to
be severe because the design is expected to survive it and the sample is not
expected to go back into the line.

A routine test is performed on every unit leaving the line. Its voltage is set
deliberately below the type-test level, because its job is to catch a
manufacturing escape — a mis-assembled unit, a wrong component, a damaged lead —
not to re-prove the design. Every unit that passes has still been stressed. If a
buyer writes the type-test voltage into the routine-test requirement, the line is
no longer screening units; it is ageing them, one at a time, and shipping the
ones that survived.

The distinction is worth insisting on in writing, because it is invisible in
a purchase order that says only “hipot tested”.

7. What changes the reading

Both tests are sensitive to conditions, and the conditions are part of the
test rather than noise around it.

Humidity is the largest single factor for insulation resistance, because
surface moisture turns a clean path into a leaky one without changing anything
about the part. Temperature shifts both the resistivity of the insulation and the
reading of the instrument. Surface contamination — flux residue, mould release,
handling oils — dominates the result at the lead exit and around the grommet.
And the interval between the last handling step and the measurement matters: a
unit measured immediately after washing and a unit measured the next morning are
not the same specimen.

For dielectric strength the sensitivity is different but no smaller. The shape
of the applied waveform, the rate at which the voltage is raised, the dwell at
the test level, and the leakage-current threshold that defines failure all change
the outcome. A slowly raised AC stress finds defects that a fast DC step misses,
and a threshold set too loosely passes units that a tighter threshold would have
caught.

None of this makes the tests unreliable. It makes them conditional, and a
conditional test is only reproducible if the conditions travel with the result.

8. What a test report has to carry

A test record that says “passed” is not transferable, because the reader
cannot tell whether it describes the same test they would run. For these two
tests, a record is only usable if it carries the conditions that produced it.

For insulation resistance: the DC voltage applied, the path tested, the
settling time before reading, the ambient temperature and humidity at the time of
measurement, and whether the unit was measured dry or after a defined conditioning
step. For dielectric strength: the test voltage, whether it was AC or DC and at
what frequency, the raise profile and dwell, the leakage-current threshold used to
declare failure, the path or paths energised, and whether the unit was destroyed
by the test or returned to stock.

That last field is the one most often missing, and it is the one that tells a
reader whether they are looking at a type test or a routine test. Without it, the
voltage figure has no meaning at all.

9. Seven questions before you approve a fan’s insulation claim

These are the questions that separate a qualification from a claim. None of
them requires a laboratory to ask; all of them require an answer in writing.

One. Which insulation paths were tested, and does the list
cover winding-to-core, winding-to-lead and lead-to-frame, or only the two most
convenient terminals?

Two. Is the stated hipot voltage a type test or a routine
test, and if it is a type test, what does the routine test use?

Three. Was insulation resistance measured before or after the
hipot, and was the same unit used for both?

Four. What leakage-current threshold defines failure, and was
it chosen for this construction or inherited from another product?

Five. Under what humidity and temperature was the sample
conditioned, and for how long?

Six. Is insulation resistance recorded per batch as a trend,
or only as a pass stamp at final inspection?

Seven. If the part is potted, coated or sealed downstream,
does the insulation test happen before or after that operation — and does the
answer change which defects the test can still find?

A supplier who can answer all seven is describing a process. A supplier who
answers only the voltage is quoting a number.

10. Why this matters beyond the certificate

Insulation failure in a fan rarely announces itself as an insulation failure.
It appears as a nuisance trip on a drive, a ground-fault indication that will not
reproduce, a controller that resets when the enclosure is washed, or a unit that
passed incoming inspection and failed in the second year. By the time the symptom
is visible, the test that would have found it was months in the past and its
conditions are no longer recoverable.

That is the argument for treating these two tests as process instruments
rather than certificates. Dielectric strength tells you the design has margin.
Insulation resistance tells you the line is still building the design you
qualified. Neither one substitutes for the other, and a qualification file that
carries only one of them is describing half a process.

If you are specifying a fan for an enclosure that will be washed, coated,
sealed, or installed where a ground fault is expensive to investigate, ask for
both readings — and ask for the conditions that produced them. The number is the
easy part. The conditions are what make it a fact.