How to Choose an Impulse Voltage Generator: A 6-Step Buying Guide

Sep 11, 2026

Leave a message

ElectriCore
ElectriCore
Senior electrical engineer with 15 years in substation automation and power quality. Author of technical articles on design codes, troubleshooting, and smart grid. Committed to practical knowledge sharing.

1. Introduction

An impulse voltage generator (commonly built on the Marx generator topology) is one of the most expensive and technically demanding pieces of equipment in any high-voltage test laboratory. Buy the wrong rating and you either waste budget on unused headroom or, worse, fail to meet the waveform requirements of IEC 60060 / GB/T 16927 for your test objects.

This guide walks through the six decisions that actually matter when you specify or buy an impulse voltage test system - drawn from more than a decade of configuring systems from 20 kV teaching units up to 2400 kV UHV type-test systems.

 

2. Step 1 - Match Generator Voltage to Your Test Object's BIL

The single most common mistake is sizing the generator to the rated system voltage of the test object (e.g. 35 kV), when you actually need to size it to the object's Basic Insulation Level (BIL).

BIL is the lightning impulse withstand voltage an item must survive, and it is always much higher than its rated voltage:

 

Test object rated voltage Typical BIL Minimum generator voltage
10 kV class 75–95 kV 200 kV
35 kV class 200 kV 400 kV
66 kV class 325–350 kV 600 kV
110 kV class 450–550 kV 800 kV
220 kV class 950 kV 1200 kV

 

As a rule of thumb, add ~20% headroom above the BIL to account for the generator's no-load efficiency (typically 85–90%) and waveform tolerances.

If you are testing distribution-class transformers, bushings, surge arresters, switchgear or cables up to 35 kV class, a 400 kV / 30 kV impulse voltage generator is the standard choice.

 

3. Step 2 - Size Energy (kJ) to the Test Object's Capacitance

Energy determines whether you can deliver the full rated voltage into a large, capacitive test object without the waveform collapsing.

Rough guideline:

Below 2000 pF load (instrument transformers, small bushings, MV switchgear): 10–30 kJ is sufficient.

2000–4000 pF load (power transformers, GIS bays, long cable lengths): 30–60 kJ recommended.

Above 4000 pF load (EHV power transformers, large generator step-up units): 120 kJ and up.

The 400 kV / 30 kV class maintains ≥90% voltage utilization up to 2000 pF load and ≥85% up to 4000 pF - which covers the vast majority of distribution-class apparatus.

If your test objects are heavy-duty 35 kV-class transformers or reactors, step up to the 400 kV / 60 kJ version rather than over-specifying voltage.

 

4. Step 3 - Do You Need Chopped-Wave Capability?

A standard 1.2/50 µs full-wave test stresses the longitudinal insulation of windings. A chopped wave - terminated artificially at 2–6 µs - additionally stresses inter-turn and inter-section insulation, which is why power transformers and instrument transformers almost always require both.

If your test standard (IEC 60076-4 / GB 1094) calls for chopped-wave tests, budget for a multi-gap chopping device (typically rated at 75% of generator voltage, e.g. 300 kV chopper on a 400 kV generator). Skipping it means you cannot complete the type-test report.

 

5. Step 4 - Oil-Filled vs Dry-Type Capacitors

This is an environmental and maintenance decision, not a performance one:

Oil-filled pulse capacitors: lower cost, higher energy density, but require periodic oil leak checks and are unsuitable for cleanrooms, indoor labs with strict hygiene rules, or climate-controlled research facilities.

Dry-type fully insulated capacitors: zero oil leakage, cleaner operation, slightly higher unit cost. Recommended for third-party labs, university environments, and any site where oil spillage is unacceptable.

Both types produce identical waveforms - the choice is about how you want to run the facility.

 

6. Step 5 - Manual, Semi-Automatic, or Fully Automatic?

Manual is fine for an R&D lab firing a few shots a day.

Semi-automatic (PLC + manual trigger) is the sweet spot for most type-test floors - auto charging, auto grounding, but the operator presses fire.

Fully automatic (configurable shot count, interval time, auto-trigger and report export) pays for itself when you are running production batches or continuous routine tests. Look for a system that stores waveforms and exports PDFs automatically - manual data entry is where human error creeps in.

 

7. Step 6 - Plan for Expansion

Generators are built from modular 100 kV stages. The same charging supply, control console and measurement chain that drives a 4-stage (400 kV) unit can usually be extended on-site to 6, 8, 12 or even 24 stages by adding tower sections - without scrapping the base system.

If you anticipate testing higher-voltage apparatus within 5 years, buy the charging and control architecture of a larger system today and add tower stages later. It is almost always cheaper than replacing the whole rig.

 

8. Quick Selection Cheat-Sheet

If you are testing… Start with…
10 kV switchgear, instrument transformers 200 kV / 5–10 kV
35 kV distribution transformers, bushings 400 kV / 30 kV
35 kV heavy-duty transformers and reactors 400 kV / 60 kV
66–110 kV GIS and substations 600–800 kV
220 kV EHV power transformers 1200 kV / 120 kJ
UHV type-test laboratory 1600–2400 kV

See the full WGJ series range from 20 kV teaching units to 2400 kV UHV systems for detailed per-model configurations.

 

9. Bonus - Don't Forget the Companion Equipment

Surge arrester type tests require both an impulse voltage generator (1.2/50 µs) and an impulse current generator (8/20 µs). They are two different machines measuring two different parameters. If arrester testing is on your roadmap, specify both at the same time to avoid integration headaches later.

 

10. Conclusion

Choosing an impulse voltage generator comes down to five numbers:

Generator voltage ≥ 1.2 × your test object BIL

Energy (kJ) sized to the largest test object capacitance you will see

Chopped-wave device if transformers are in scope

Capacitor type (oil vs dry) based on your lab environment

Automation level matched to your daily shot count

Get those right and the rest - divider accuracy, fiber-optic isolation, PLC protection - is table stakes that any serious supplier should deliver as standard.

Not sure which model fits your test objects? Send us your equipment list and system voltage class, and we will recommend a specific configuration.

Send Inquiry