A thermal camera will put a number on every hot spot. Turning that number into a grade the client can act on, repair now, repair soon or keep an eye on it, is the part of the job that has to hold up when someone asks why.
- Grade by temperature difference (ΔT), not by absolute temperature.
- Compare with a similar component under similar load first; use ambient air only when you have to.
- NETA Table 100.18 is the common reference: over 15 °C above a similar component means repair immediately.
- Record the load, weigh the grade against criticality, and keep the same scale every year.
Why ΔT and not temperature
A breaker terminal at 45 °C means little on its own. In a warm electrical room beside identical terminals at 42 °C, it is normal. Beside identical terminals at 25 °C, carrying the same current, something is wrong with that connection. The difference is the finding.
Two ways to measure it
Against a similar component under similar load
The preferred method. Compare phase A's lug with phase B's and C's on the same breaker, or one breaker's terminal with its neighbour carrying a similar load. Because both parts share the same room, the same load pattern and usually the same surface, most sources of error cancel out. Make sure the reference really is similar: same component type, similar current, similar surface.
Against ambient air
Used when there is nothing comparable to measure, such as a single-phase component or a lone connection. It is less forgiving: the reading now carries the full effect of load, emissivity and the air temperature you chose as ambient. Measure ambient near the equipment, not at the thermostat, and note in the report that the grade is against ambient.

NETA Table 100.18
The ANSI/NETA Maintenance Testing Specifications give the temperature-rise guidance most electrical thermographers in North America work from. It has a column for each comparison method:
| ΔT vs. a similar component, similar load | ΔT vs. ambient air | Suggested action |
|---|---|---|
| 1 °C to 3 °C | 1 °C to 10 °C | Possible deficiency; warrants investigation |
| 4 °C to 15 °C | 11 °C to 20 °C | Probable deficiency; repair as time permits |
| — | 21 °C to 40 °C | Monitor until corrective measures can be accomplished |
| Over 15 °C | Over 40 °C | Major discrepancy; repair immediately |
Summarized from ANSI/NETA MTS Table 100.18. Wording varies slightly between editions; check the edition your program or client specifies.
Note the third row: the ambient column has a monitor band (21–40 °C) that the similar-component column does not. Many reports collapse the table into the familiar three comparative bands, which is fine as long as the report says which column a finding was graded on.
Load changes the reading
The heat made at a resistive connection is the current squared times the resistance, so doubling the current makes four times the heat. A loose lug that is a mild finding at 30% load can be a serious one at the afternoon peak. That is why load matters as much as the camera:
- Scan at normal operating load where you can. NFPA 70B guidance is no less than 40% of normal circuit loading.
- Record the load with the finding, ideally amps on each phase.
- Treat a reading taken at light load as a minimum. If the finding matters, re-scan at a higher load rather than estimating.
Severity is not the same as priority
A ΔT grade says how bad the connection is. It does not say how much the client loses if it fails. A probable deficiency on the feeder to an operating room, a process line or a fire pump is scheduled sooner than the same reading on a spare circuit. Good reports keep the two apart, a severity from the measurement and a criticality from the asset, and rank the repair list on both.

Pick a scale, publish it, keep it
Some programs use the NETA bands as they are. Others use a three-level traffic light, or a five-level scale like the one in our guide to writing an electrical IR inspection report. Any of them works if three things hold:
- The criteria are printed in the report's methodology, so the client and the insurer can see how a finding was graded.
- Every finding is graded on the same scale, by every thermographer on the job.
- The scale stays the same next year, or the year-over-year comparison stops meaning anything.
How DTI Suite applies it
In DTI Suite you place spot and box markers on each radiometric photo and read the ΔT on the photo. Every edited photo is rated against your severity scale: the default, or your own, with your names for each level, the temperature-rise range for each and your colours, including a three-level traffic light. Findings are ranked for repair using NETA temperature-rise benchmarks and plotted on a severity × criticality matrix, and the report carries the same scale from one year to the next.
Questions
What is a ΔT in electrical thermography?
What ΔT is considered serious?
Should I compare with ambient or with a similar component?
How much load does a ΔT reading need?
Build the whole report in minutes
DTI Suite reads your radiometric photos, pairs them to your Excel asset list and writes the full Word report: summary, findings pages, severity matrix, re-inspection schedule and year-over-year trending. 30-day free trial, no credit card.
Start your free trial →