- Pulse voltage
- 100 V to 3000 V on HG9813, 100 V to 5000 V on HG9815, 200 V to 6000 V on HG9816 and 300 V to 10,000 V on HG9810, all in 10 V steps at 5% ±10 V
- Pulse energy
- 0.09 J on HG9813, 0.25 J on HG9815, 0.36 J on HG9816 and 0.5 J on HG9810
- Inductance range
- ≥10 µH on HG9813, HG9815 and HG9816; ≥50 µH on HG9810
- Test speed
- 15 per second on HG9813 and HG9815, 12 on HG9816 and 8 on HG9810
- Waveform capture
- Up to 100 MSa/s in 10 steps, 8-bit resolution, 6500-byte record depth and 1 to 32 sample averages
- Display
- 7-inch 65k colour TFT, 800 × 480 dots, with a 650 × 256 waveform area and 1.5× magnification
- Judgement methods
- Area, area difference, corona discharge and phase difference
- Storage
- 300 reference waveforms and settings inside the instrument, 600 on a USB drive, and statistics for up to 20,000 measurements
- Interface
- HANDLER, RS232C, USB Device and USB Host as standard, GPIB optional
Two coils can have the same inductance and the same resistance and still behave differently under a pulse, because one of them has a weakness between two turns that no steady-state measurement reaches. This range is built to make that difference visible.
Product overview
Four models share one measuring method and differ in how hard they can drive a winding. The pulse voltage runs from 3 kV to 10 kV across the range, the energy steps up with it, and the inductance floor rises at the top model because a larger pulse is intended for a larger winding.
Around that core sit the parts that make the instrument usable on a line: a colour display large enough to compare two traces, storage for reference waveforms, statistics on the results, and the interfaces a handler or a computer needs.
How the winding insulation tester works
The instrument discharges a pulse into the winding and records the voltage across it as the oscillation decays. The shape of that decay depends on the inductance of the winding, the capacitance between its turns and the losses in its insulation, so a change in any of them shows up as a change in the trace.
A known-good winding is measured first and its trace is stored. Each part afterwards is pulsed the same way and compared against that reference. A short between adjacent turns removes turns from the resonant network and shifts the response; a weakness that has not yet shorted may leave the low-frequency shape intact but add high-frequency activity.
Corona analysis looks for exactly that activity. A partial discharge emits a short burst of high-frequency energy, and the instrument’s sampling rate and corona display are there to catch it. Because the burst is brief, the capture has to be fast and the record has to be deep enough to hold the whole event.
Several pulses can be applied in one test, with demagnetising pulses at the end, so a winding is stressed more than once before a decision is made. That matters for faults that only appear after the insulation has been worked.
How to select the right winding insulation tester
Start with the test voltage. It comes from the insulation class of the winding and from the procedure the customer already uses, not from the highest number on the datasheet. A pulse above what the insulation can take will damage good parts, and the damage may only surface later.
Then check the inductance. Each model has a floor below which the response is too small to judge, and the high-voltage model has the highest floor because it is meant for large windings. A small coil on the largest instrument is the wrong pairing.
Then look at the energy and the speed. A large winding needs more energy to reach the test voltage, and a line that tests every part needs the faster model. The two move in opposite directions across the range, so the choice is a compromise that the production rate usually settles.
Then decide how the result is used. A bench uses the display and the foot switch; a line that sorts needs the handler lines; a line that records needs the serial or USB link. Send the winding drawing, the inductance and the insulation class, and a model, a lead and a fixture can be quoted against them.
Measurement limitations and selection considerations
The test judges a part against a reference, so the reference is part of the measurement. A trace stored from a marginal winding will pass marginal parts, and a reference taken at a different voltage or in a different fixture is a different test.
The pulse is a stress. A part that passes has survived that pulse at that energy; whether it survives the insulation test that follows, or the voltage in service, is a separate question, and a test that is too severe can create the fault it is meant to find.
Corona detection is sensitive to the environment as well as to the part. Humidity, surface contamination and the capacitance of the fixture all affect what the trace shows, so a comparison between shifts or between rooms has to be made under the same conditions. The material does not state the conditions behind the repeatability figure, which is why it should be established locally before a limit is set on it.
Finally, the method finds a difference, not a cause. A trace that departs from the reference tells the operator that this winding is not like the reference winding; whether the difference is a shorted turn, a damaged core or a handling fault has to be found by looking at the part.
Every value below is taken from source documentation. Where a parameter is not stated in the available material it is marked as information required rather than estimated.
