Why Is My PD Detector Reading High? Common Causes and Fixes

The Short Answer: Verify Before You Act

High readings on a partial discharge detector fall into three buckets: real PD (internal voids, surface tracking, corona), electrical noise (telecommunications, nearby switching, grounding issues, LED drivers), and measurement errors (wrong sensor position, poor coupling, loose contact). Only one of the three requires a repair. That is why the golden rule of PD testing is: never schedule an outage on a single reading — always verify with a second method and a pattern analysis first.

Key principle: A PD detector is a screening tool. It tells you where to look and how loud the activity is. Confirmation — deciding whether that activity is dangerous PD or harmless noise — comes from PRPD pattern analysis, a second sensor method, and a thermal cross-check.

What dBmV and pC Readings Actually Mean

Handheld PD detectors like the HPD200 report two types of values:

• dBmV (TEV channel): the amplitude of the transient earth voltage pulse on the switchgear enclosure. Typical readings range from 0 to 60 dBmV. It is a relative, location-dependent number — it depends on sensor placement, panel size, and grounding.
• dB or pC (ultrasonic / HFCT channels): the acoustic intensity in dB, or the apparent charge in picocoulombs for directly coupled measurements. pC is closer to a physical quantity but still requires calibration against a known reference.

The single most important habit in PD testing is trending: a value that jumps from 8 dBmV to 25 dBmV on the same panel, measured the same way, matters far more than an absolute number. Instruments that store per-asset history, like the HPD500, make this comparison automatic.

Seven Common Causes of High Readings

Notice that only the first five rows mean real asset damage. Noise and measurement errors are the two most common reasons for a “high reading” that turns out to be nothing.

Real PD vs Noise: How to Tell Them Apart

The most reliable discriminator is the PRPD pattern (Phase-Resolved Partial Discharge). When you display the pulses against the AC sine wave:

• Internal PD clusters in stable phase windows, typically before the voltage peaks (30–90 degrees and 210–270 degrees), with a consistent shape.
• Corona appears right before the zero crossing in a narrow band, usually symmetrical in both half-cycles.
• Surface discharge forms elongated patterns that widen as humidity rises.
• Noise fills random phase positions, has no repeatable shape, and changes when nearby equipment switches on or off.

Entry-level detectors report numbers only; instruments with PRPD/PRPS capability — such as the SC Series and the HPD500 — let you see the pattern directly on the screen and store it for review. If your detector cannot show a pattern, treat any high reading as unconfirmed until a second method agrees.

A Six-Step Verification Workflow

When a reading looks high, run this sequence before deciding anything:

1. Repeat the measurement correctly. Clean the panel surface, press the TEV sensor firmly, hold for 10–15 seconds, and record the reading at the same point as the baseline.
2. Cross-check with the second sensor. If TEV is high, listen with the ultrasonic sensor at the same location. Real PD in switchgear usually produces both; noise often produces only one.
3. Check adjacent panels. If neighboring panels show the same level, the signal is likely coupled through the busbar or coming from outside — not from a single defect.
4. Analyze the PRPD pattern. Classify it as internal PD, surface, corona, floating, or noise (see the table above).
5. Thermal cross-check. Scan the suspect area with a thermal camera. A hotspot confirms a resistive component of the problem; a completely cold panel points to pure insulation PD.
6. Retest over time. Record the reading, set a follow-up interval (1–3 months), and compare. A rising trend confirms real progression; a stable or falling reading suggests interference or a transient condition.

For hard-to-reach or ambiguous sources, an acoustic imaging camera like the GSW Series can visualize the discharge location directly, removing all doubt about whether the reading belongs to the asset you are testing.

Setting Sensible Thresholds for Your Plant

Absolute thresholds vary by equipment type, manufacturer, and environment, so publish your own. A practical starting framework used by many LatAm utilities:

• Green (normal): TEV below 10 dBmV with no pattern — continue routine schedule.
• Yellow (caution): TEV 10–20 dBmV or a clear pattern — retest within 1–3 months and trend.
• Red (action): TEV above 20 dBmV with a confirmed PD pattern, or a rising trend of more than 10 dB over two tests — plan repair and shorten the interval.

Baseline your own plant first. A quiet rural substation will have different background levels than an industrial plant with VFDs and welding loads. The HPD200 and HPD500 both store per-asset baselines, so your thresholds become data-driven instead of guesswork.

The Bottom Line

A high PD reading is an invitation to investigate, not a verdict. Real PD, electrical noise, and measurement errors all produce elevated numbers — and only one of them needs a repair crew. Verify every alarm with a second sensor, a PRPD pattern, and a thermal cross-check before scheduling an outage. Build baselines, trend everything, and set thresholds that reflect your own environment. With that discipline, a high reading becomes a precise, actionable finding instead of a false alarm or a missed defect. If your current detector shows numbers but no patterns, upgrading to an instrument with PRPD analysis is the fastest way to stop guessing.