2026-11-19
A working routine for hipot sets (high-potential withstand testers), VLF (very low frequency) generators, megohmmeters, relay testers and cable fault locators - written for the people who own the equipment, not only for the lab that calibrates it.
A high-voltage test set rarely fails with a bang. It drifts. A unit that reads 2% low still switches on, still prints a curve, and still says PASS. The first sign of trouble is usually a failed joint that should have passed, or a good cable that was rejected on site. Maintenance exists to stop that from happening.
First, a plain definition. A high-voltage test set applies a controlled voltage above the normal working level of a piece of equipment, then measures how much current leaks through its insulation. That single idea covers withstand tests, leakage tests and most diagnostic tests that follow.
For a multimeter, maintenance mostly means "keep it working". For high-voltage test equipment, three things are at stake at the same time, and they fail in different ways.
A useful way to think about it: a torque wrench that reads 5% high tightens every bolt too far, and nothing about the wrench looks wrong. High-voltage test sets behave the same way. The reading is the product, so the instrument has to be maintained as a measuring device, not just as a machine.
In service experience, the high-voltage generator is usually the healthiest part of the kit. The parts that fail are the ones that get coiled, dragged, stepped on and thrown into a van: output cables, ground leads, clamps, adapters, discharge rods and dividers.
Most field failures are caught by a short, repeated check rather than by an annual service. The routine below is deliberately boring, which is the point.
DC tests charge the specimen like a capacitor, and a large part of that energy is still there after the set is switched off. The amount is easy to underestimate.
E = 0.5 x C x U^2
For a 5 µF cable at 60 kV: E = 0.5 x 5e-6 x (60,000)^2 = 9,000 J (9 kJ). That is a dangerous amount of energy, and it explains why discharge is a procedure with steps, not a button press.
What happens after the test depends on the type of test and on the specimen. The three cases below cover most field work, but they are general guidance only. The instrument manual and your site safety procedure define the exact steps.
Switching off the output, or watching the indicated voltage fall to zero, does not necessarily mean the specimen is fully discharged. Cable insulation, long test leads and the specimen's own capacitance can hold charge that the display no longer shows. A practical sequence:
An insulation resistance test can leave dielectric absorption, or polarization-related charge, in the specimen even after the terminal voltage has fallen to a low value. This is most noticeable on large windings and long cables. Discharge to earth and keep the connection for the period your procedure specifies.
IEEE 43-2013 gives a minimum discharge-time recommendation of four times the voltage application time. That recommendation applies to the rotating-machine winding insulation tests within the scope of that standard. It is not a universal requirement for every megohmmeter test, for cables, or for other insulation systems. Where a manufacturer's instruction or a product-specific test procedure gives a different time, that instruction takes priority.
A VLF test charges a cable in a different way from a DC test, but the ending deserves the same discipline. A common sequence is:
IEEE 400.2-2024 can be referenced for the general principle of slowly reducing the voltage, discharging the cable and maintaining grounding. It should not be read as a universal instruction to hold a specific voltage-reduction period for every cable and every set. Follow the edition of the standard, the cable manufacturer's guidance and your own procedure for the actual timings.
Zero indicated voltage does not mean zero stored charge. The display shows what the instrument can measure at its terminals, not what remains inside a cable or a winding.
Discharge and grounding are two separate steps. Verification of residual voltage is necessary, but it has to be interpreted together with the discharge and grounding procedure, not instead of it. Never touch, short or disconnect a test circuit merely because the instrument reads zero.
Treat this section as general guidance, not as a universal operating procedure. Exact voltages, discharge device ratings, waiting times and grounding arrangements vary by instrument, specimen and standard. The equipment manual and the site safety procedure are the controlling documents.
Moisture is the most common environmental cause of inaccurate readings and flashovers. The mechanism is simple: water films on insulation surfaces create leakage paths, and leakage current adds to the current you are trying to measure.
A megohmmeter check on the high-voltage circuit of the set, at a voltage the insulation can take, gives an early warning that moisture or contamination is building up. One reading proves little. The trend across months is what tells you something is changing.
The instrument's own high-voltage circuit is one thing; its internal electronics are another. An external megohmmeter should only be connected to test points and circuits that the equipment manufacturer permits users to test, and the test voltage must suit that specific circuit. Do not apply insulation resistance test voltage indiscriminately to internal electronic circuits, measurement modules or protective components. If the manufacturer's documentation does not describe an approved test method, consult the manufacturer or qualified service personnel rather than improvising one.
PI = R(10 min) / R(1 min)
PI compares resistance after one minute with resistance after ten. A flat ratio suggests the insulation is stable; a ratio that falls over successive checks suggests moisture or contamination. Use the limits from the applicable standard and from the equipment manufacturer, and treat a single abnormal result as a reason to investigate, not as a final verdict.
These two activities get mixed up, and the confusion causes most of the arguments about "calibration intervals".
Calibration is a comparison against a reference that is traceable to a national standard, performed by a competent laboratory, and reported with measurement uncertainty. It tells you how far the instrument is from the truth, and it does not necessarily adjust anything.
Verification is a quick in-house check between calibrations, using a divider, a reference meter or a known resistance. It tells you whether the instrument has moved since the last calibration. It does not replace calibration, but it catches drift early and it is cheap.
Divider-based measurement is the normal way to read high voltage, and the ratio is where errors enter. The correction is simple once the purpose is clear: you are converting what the instrument displays into what is actually at the specimen.
k = U(reference) / U(indicated) and U(true) = U(indicated) x k
If a 1000:1 divider and its meter indicate 60.0 kV while the reference reads 60.6 kV, then k = 1.010 and every future reading should be multiplied by 1.010. Record k with the divider, not with the person who measured it. When k moves outside the instrument's specified accuracy - about 1% for many high-voltage sets, wider for resistance ranges - the divider or meter needs attention rather than a new correction factor.
| Activity | Typical interval | Notes |
|---|---|---|
| Visual and functional check | Every use | Leads, ground, self-test, battery |
| In-house verification | Monthly, or before critical work | Divider ratio, current reading, timer |
| Traceable calibration | 12 months | Shorten for heavy use, safety-critical work or harsh environments |
| Insulation check of the set | 6 to 12 months | Compare with the previous record |
| Battery capacity check | 6 months | Replace when runtime falls below about 70% of new |
These are reasonable starting points, not universal rules. Your quality system, the manufacturer's instructions and how hard the set works all move the interval. A set used daily on a production line deserves a shorter cycle than one used twice a year.
High-voltage sets concentrate loss in a small package, and both batteries and insulation dislike heat. Two habits prevent most of the damage.
High-voltage generators are rated for a working pattern, not for continuous operation. For example, the XHZG series DC hipot sets are rated for 30 minutes at rated load and 10 minutes at 1.1 times rated voltage. That is a product-specific working duty for that series, not a universal duty cycle for all high-voltage generators. The actual operating and rest periods for any instrument must follow its rated duty and the manufacturer's instructions. Exceeding the rating heats the transformer, the multiplier and the insulation, and heat shows up first as increased ripple (the small AC component riding on the DC output) and drift, then as failure.
Keep vents and fans clear, clean dust filters, and do not operate the set inside a closed vehicle or a sealed cabinet. As a reference example, the XHZG series is specified for an ambient range of about -10 °C to +40 °C, relative humidity up to 85% and altitudes below 1500 m. Those figures describe typical reference conditions for that product family, not universal ratings for every instrument. At high altitude, thinner air reduces both cooling and withstand margin, so derating may be necessary; check the nameplate and manual for the equipment you actually have.
A maintenance programme without records becomes a set of opinions. Three records carry most of the value:
Trending turns a single number into a warning. A leakage current that rises steadily suggests the set's insulation is absorbing moisture or contamination, even while it still passes a go/no-go check.
Delta I(%) = (I(current) - I(baseline)) / I(baseline) x 100
Compare like with like, or the trend means nothing. A baseline is only useful if it was taken under conditions you can reproduce: the same test voltage, the same warm-up time, the same ambient temperature and humidity, the same lead arrangement and connections, and the same test configuration. Establish the baseline when the equipment is known to be in good condition - preferably after calibration or a satisfactory verification. Take at least three consecutive measurements under those consistent conditions and use a stable representative value, rather than a single reading.
As an in-house guide, a rise beyond roughly 20% is worth investigating and beyond 50% is a reason to stop using the set for critical work until it is checked. Those two numbers are internal maintenance policy examples. They are not IEC or IEEE requirements, and an abnormal trend is not automatic proof that the insulation has failed. Repeat the measurement, confirm the conditions, and investigate before drawing a conclusion.
Instrument memory is not a backup. Many sets hold a limited number of records - ninety is typical on some VLF units - and then overwrite the oldest. Export results monthly and keep the calibration certificate with the asset record.
| Equipment | Maintenance focus |
|---|---|
| DC hipot / withstand set | Output cable and multiplier condition, ripple, voltage and current accuracy, duty cycle, discharge function |
| VLF generator | Waveform quality, frequency stability, cable and connector condition, cooling, stored records |
| High-voltage divider | Ratio stability after impact or transport, surface cleanliness, stable mounting and ground |
| Insulation oil BDV (breakdown voltage) tester | Electrode condition and gap, oil cup cleanliness and dryness, calibration of the voltage ramp |
| Megohmmeter | Lead insulation, guard terminal, battery, resistance accuracy at the top of each range |
| Relay protection tester | Output accuracy on current and voltage channels, channel balance, firmware, connector wear |
| Circuit breaker analyser | Contact timing channels, travel transducer calibration, clamp condition, timebase accuracy |
| CT / PT analyser | Ratio and excitation accuracy, lead sets, grounding practice, software version |
| Cable fault locator and surge generator | Capacitor bank health, discharge gap, earthing, surge cable condition, safe discharge after use |
| Symptom | Likely cause | First action |
|---|---|---|
| No high voltage output | Interlock open, ground not connected, supply or battery fault | Check grounding and interlocks before suspecting the generator |
| Ripple higher than normal | Overload, overheating, failing filter capacitor | Stop, cool, retest at reduced load; compare with baseline |
| Reading lower than expected | Wrong divider ratio, contaminated leads, poor ground | Verify ratio and clean connectors; check with reference divider |
| Trips immediately on a cable | Genuine specimen fault, or a fault in the lead set | Disconnect the specimen and test the set alone to separate the two |
| Erratic current display | Loose connector, moisture, EMI from nearby equipment | Dry and reseat connections; move away from noise sources |
| Over-temperature shutdown | Blocked vent, high ambient, exceeded duty cycle | Clean vents, allow cooling, review the test plan |
| Short battery runtime | Aged cells, cold conditions, charger fault | Capacity test; replace pack if below about 70% of rated runtime |
| Software will not connect | Cable, driver, firmware or port conflict | Swap cable and port first; then check firmware and driver versions |
| Frequency | Tasks |
|---|---|
| Every use | Visual check, ground first, self-test, duty planning, discharge and residual voltage check |
| Monthly | In-house verification, lead and clamp inspection, data export, battery top-up |
| Quarterly | Deep clean, fan and filter check, insulation resistance of the set, record review |
| Annually | Traceable calibration, full functional test, accessory replacement review, firmware update |
| After any impact or fault | Verification before return to service; investigate rather than reset and continue |
Repair usually wins when the fault is in a lead, a connector, a battery or a fan, and when spare parts and service are available. Replacement becomes the better decision when calibration cannot be brought back within specification, when high-voltage insulation is damaged, when the unit is unsafe, or when the repair cost approaches half the price of a new instrument. Age alone is not a reason; a well-maintained 15-year-old hipot set can be more trustworthy than a new one with no records.
Twelve months is the common default, but it is a starting point rather than a rule. Shorten the interval if the set is used daily, if results are safety-critical, or if it works in heat, humidity or vibration. Lengthen it only with evidence from verification records, and only if your quality system allows it.
For non-critical, internal checks you may choose to continue with a documented risk decision, supported by an in-house verification. For acceptance testing, customer witness tests or safety-related work, an expired certificate is usually not acceptable. The decision belongs to your quality procedure, not to the person holding the test lead.
The three usual causes are a wrong divider ratio, contamination on the high-voltage leads creating a leakage path, and a poor ground connection. Check those before touching any internal adjustment. A set that reads low because of a ratio error will read consistently low, while a contamination problem often varies with humidity.
It is common but not ideal. Temperature swings cause condensation, and condensation on high-voltage insulation is exactly what you do not want. If outdoor storage is unavoidable, use a sealed case with fresh desiccant, allow the set to reach room temperature before use, and check insulation resistance more often.
Measure runtime under a realistic load, not idle time. When usable runtime falls below roughly 70% of the original figure, or the pack swells or becomes unreliable under load, replace it. A weak pack can cause intermittent shutdowns that look like instrument faults.
Replacing a lead normally needs only a verification, not a full recalibration. Replacing or repairing a divider changes the measurement chain, so verify the ratio against a reference before returning the set to service, and record the new correction factor.
These are examples of the instrument types covered above, with the specifications that matter for maintenance planning.
Tell us the instrument type, the highest test voltage and how often it is used, and we will help you set a practical verification and calibration plan - including the lead set, divider and accessories that usually get forgotten.
Scope note: this guide describes general good practice for high-voltage test equipment. Always follow the manufacturer's instructions and your site's safety procedures. Intervals and thresholds quoted here are typical starting points, not mandatory requirements.
Send your inquiry directly to us