As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
Clamp meter accuracy is usually published as a percentage of reading plus a number of counts, valid over a stated temperature range and within a stated jaw position. The percentage covers gain error and the counts cover offset and resolution, so a small current on a high range can carry a large relative error even when the meter meets its specification. The jaw, the conductor centring, nearby magnetic fields and the waveform all sit outside the accuracy statement. A meter that meets its published figure can still mislead you if the conductor is off centre, if the load is non-sinusoidal, or if you are reading a value below the resolution of the range you selected.
Reading the accuracy statement as a range, not a promise
Almost every clamp meter publishes accuracy in the same form: a percentage of the reading plus a number of counts, sometimes called digits. A statement such as plus or minus 2 percent plus 5 counts on the 600 A range means two separate errors are being declared. The percentage is a gain error: if the meter reads 100.0 A, two percent is 2.0 A. The counts are an offset and resolution allowance: five counts on a display showing one decimal place at 100.0 A is 0.5 A. The two add, so the honest worst case at that reading is 2.5 A, or about 2.5 percent of the reading.
The counts matter most at the bottom of a range. On the same 600 A range, a reading of 5.0 A carries the same five counts, which is now 0.5 A against a 5.0 A reading, ten percent before the percentage term is added. This is why a clamp meter that is perfectly accurate at 200 A can look wild on a small load. The fix is not a better meter; it is a lower range, a smaller jaw, or a different instrument. A digital multimeter in series with the load, or a dedicated low-current probe, will resolve milliamps that a 600 A clamp cannot.
The accuracy statement is also conditional. It applies over a stated temperature range, usually around room temperature, and outside that range a temperature coefficient applies per degree. It applies with the conductor centred in the jaw and perpendicular to it. It applies to a sinusoidal waveform at a stated frequency band. A reading taken on a cold morning, with the cable pushed against one side of the jaw, on a load driven by a variable frequency drive, is outside all three conditions. The meter may still be within its specification, but the specification no longer describes what you are measuring.
Why the jaw and the conductor decide more than the percentage
A clamp meter is a current transformer with an air gap that opens to admit a conductor. The jaw senses the magnetic field the conductor produces, and the field strength depends on the geometry. A conductor centred in the jaw produces a field the meter is calibrated for. A conductor pushed against the inner edge, or a bundle of conductors that does not fill the jaw, produces a different field distribution. Manufacturers usually publish a positioning error figure, often a small percentage of reading, but many listings do not. Where a listing is silent, treat off-centre readings as carrying an unquantified additional error and centre the conductor as carefully as you can.
Small currents in a large jaw are the hardest case. The signal is weak relative to stray fields from adjacent cables, transformers and motors. The large jaw on an industrial meter is there to accept thick conductors, not to improve sensitivity. If you need to measure a few amps on a cable that also carries a heavy load nearby, orient the jaw so the return conductor is outside it, keep the meter away from transformers, and accept that the reading is approximate. For anything below the resolution of the selected range, the display will show zero or a flickering digit, and no amount of care will recover a number the instrument cannot resolve.
Direct current is a separate problem. A Hall effect clamp measures the field from a DC conductor, and the sensor has an offset that drifts with temperature and with magnetisation of the jaw from previous high-current measurements. Many DC clamps include a zero button for this reason. Press it with the jaw closed and no conductor inside before each measurement session, and again after measuring a large current. Skipping the zero is one of the most common causes of a DC reading that is wrong by several amps while the meter is technically working.
True RMS, crest factor and the loads that break a cheap clamp
Averaging meters assume a clean sine wave. They rectify the signal and scale it so that a sine wave reads correctly, which means a distorted waveform reads low. True RMS meters compute the heating equivalent of the waveform and read correctly on distorted loads. The distinction matters for anything with a switching supply, a dimmer, a variable frequency drive or a motor on a soft starter. On those loads an averaging clamp can under-read by tens of percent while showing a stable, plausible number.
True RMS has its own limit, called crest factor. A waveform with sharp peaks relative to its RMS value, such as the current drawn by a switch-mode supply, can exceed the crest factor the meter is specified for, and the reading becomes unreliable. The listing rarely publishes a crest factor, so treat readings on heavily distorted loads as indicative. The FLIR CM276 addresses one part of this problem with a VFD mode that applies a low-pass filter, which is the right approach for the fundamental frequency but deliberately ignores the high-frequency content that a true RMS reading would include. The two readings answer different questions, and neither is wrong.
Frequency response is the other half of the same issue. A clamp specified from 10 Hz to 1 MHz, as the EX655 listing states, is unusual in claiming such a wide band, and the accuracy at the extremes will not match the accuracy at 50 or 60 Hz. The listing does not publish a separate accuracy figure for the frequency range, so the headline accuracy should be read as applying to the power frequencies where these meters are normally used. Where a listing does not publish a frequency band at all, do not assume one.
What the safety category does and does not tell you about accuracy
A CAT rating describes the transient energy the meter is built to survive, not its accuracy. A CAT IV 600 V rating, as stated for the UNI-T UT219PV, means the meter is constructed and tested for the overvoltage category at the service entrance. It says nothing about whether the current reading is within one percent or five percent. Buyers sometimes treat a high CAT rating as a proxy for quality across the board; it is not. A meter can be robustly built and only moderately accurate, or accurate and built for lower-energy circuits.
The two specifications do interact in one practical way. A meter rated for a higher category usually has a larger jaw, heavier insulation and more input protection, all of which cost money and size. That is a fair trade for anyone working on distribution equipment. For a domestic consumer unit or a HVAC service call, a lower category meter with a smaller jaw is often easier to position accurately, and positioning accuracy may matter more to the reading than the meter’s own error. The right question is not which meter has the best accuracy figure but which meter you can place correctly on the conductor you actually have.
Solar work adds a voltage dimension. The Megger DCM1500S is listed at 2000 V DC and TRMS 1500 V AC, with 1500 A AC/DC. Those are range and safety figures. The listing does not publish an accuracy statement, so a buyer comparing it with another solar clamp has no published basis for saying which reads closer to true. That is a gap in the listing, not a defect in the instrument, and it should be treated as such.
Where clamp meter accuracy stops being the right question
If you need to know whether a circuit is live before touching it, a clamp meter is the wrong instrument for the first check. A non-contact voltage detector or a proving unit and a two-pole tester answers that question. A clamp with a non-contact voltage function, such as the EX655 listing’s NCV feature, is a useful secondary check but is not a substitute for proving dead. The accuracy of the current reading is irrelevant to that decision.
If you need to detect a small leakage current, a standard clamp is usually the wrong tool. Leakage in the milliamp range is below the resolution of a 600 A or 1500 A clamp. A dedicated leakage clamp with a milliamp range, or a suitable multimeter in series, is what the measurement requires. The same applies to checking a single conductor inside a cable: a clamp around the whole cable reads the vector sum of the outgoing and returning currents, which is near zero on a healthy circuit. That is the basis of leakage measurement with the right instrument, but with a general-purpose clamp it produces a confusing near-zero reading that says nothing about accuracy.
Finally, accuracy is not the same as calibration. A meter that left the factory within its published tolerance will drift. For work where the reading is recorded or used to make a decision about a protective device, the instrument needs periodic calibration against a traceable reference, and the calibration certificate is what makes the reading defensible. No accuracy specification survives an uncalibrated meter. If the reading matters enough to argue about, it matters enough to calibrate.
The number to trust is the one you can defend
Clamp meter accuracy is a conditional statement about a well-positioned conductor, a sinusoidal load and a calibrated instrument. Read the percentage and the counts together, check the range against the current you expect, centre the conductor, zero a DC clamp before use, and treat readings on distorted loads as indicative. Where a listing does not publish an accuracy figure, as with the Megger DCM1500S in the supplied data, do not assume one; ask the manufacturer or the distributor for the published statement. The instrument that reads within its specification under your conditions is more useful than one with a tighter figure you cannot reproduce in the field.
Frequently Asked Questions
Is a clamp meter as accurate as a multimeter?
For current, a general-purpose clamp is usually less accurate at low currents because its range starts high and the counts term dominates. A multimeter in series with the load can resolve milliamps that a 600 A clamp cannot. For voltage and resistance, the clamp's built-in multimeter functions are comparable to a handheld multimeter of similar quality, but the listing must publish the accuracy for those functions before you can compare.
What does plus or minus 2 percent plus 5 counts actually mean?
It means the worst-case error is two percent of the displayed reading plus five units in the last digit. At a reading of 100.0 A with one decimal place, that is 2.0 A plus 0.5 A, so 2.5 A. The counts term stays fixed as the reading falls, so it dominates at the bottom of the range.
Why does my clamp read zero on a cable that is clearly carrying current?
Either the current is below the resolution of the selected range, or the jaw is around both the outgoing and returning conductors so the fields cancel. Clamp around one conductor only, select a lower range if the meter has one, and check that the jaw is fully closed.
Does a higher CAT rating mean better accuracy?
No. The CAT rating describes transient survival, not measurement accuracy. A CAT IV 600 V meter, such as the UNI-T UT219PV in its listing, is built for distribution work, but its accuracy is a separate published figure. Choose the category for the installation you work on and the accuracy for the measurement you need.
Can I trust a clamp meter reading on a variable frequency drive?
Only with care. A true RMS clamp will include the high-frequency content and may read higher than the fundamental current. A meter with a low-pass filter, as the FLIR CM276 listing describes for VFD mode, reads the fundamental and is the better choice for comparing against motor nameplate current. Neither reading is wrong; they answer different questions.
How often does a clamp meter need calibrating?
The listing does not state a calibration interval. Most manufacturers publish a recommended interval, commonly annual for professional use, but the right answer depends on how often the meter is used and how much the reading matters. If the reading is recorded or used to justify a decision, follow the manufacturer's interval and keep the certificate.