As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
Accuracy on a multimeter is published as a tolerance, usually a percentage of the reading plus a number of least significant digits, and it only applies within a stated temperature range, after a stated warm up, and for a stated period after calibration. The percentage covers gain error and the digits cover offset and quantisation. Neither covers the test leads, the contact resistance at the probe tips, or the waveform you are measuring. A meter that is accurate to 0.5 percent on a bench in a warm room can be several percent out on a cold rooftop with a tired lead, and the display will not tell you.
What the accuracy figure is actually promising
Every multimeter specification has the same shape. It gives a percentage of the reading, a number of digits, and a set of conditions. The percentage is a gain error: it scales with the size of the measurement, so a 1 percent error on a 230 V reading is 2.3 V. The digits are an offset error expressed in the smallest unit the display can show at that range, so on a 6000 count meter set to the 6 V range the least significant digit is 0.001 V, and two digits means 0.002 V of uncertainty regardless of what you are measuring.
The two parts matter differently at different points on the range. At the top of a range the percentage dominates. Near the bottom the digits dominate, which is why a 6000 count meter reading 0.5 V on the 6 V range is far less certain in relative terms than the same meter reading 5 V. This is the reason a meter with more counts is useful: it lets you put the reading higher up a range, where the percentage error is smaller relative to the number you are reading.
The conditions are not fine print to be skipped. Accuracy is normally quoted at a temperature around 18 to 28 degrees Celsius, after a warm up period, and for a defined interval since the last calibration, often one year. Outside that temperature band the manufacturer usually publishes a temperature coefficient, in percent per degree, and you add it to the base figure. A meter used in a cold plant room or a hot ceiling void is outside the band its accuracy was quoted for, and the honest reading of the specification is that the number is less certain than the brochure suggests.
Counts, digits and the arithmetic you can do yourself
The count figure describes the resolution of the display, not its accuracy. A 6000 count meter can show 5.999 on a range, and the last digit is worth 0.001. The number of digits in the accuracy specification is expressed in those same units, so the two are linked: more counts means each digit is a smaller step, which makes the offset term smaller in absolute terms.
Take a meter quoting plus or minus 0.5 percent of reading plus 2 digits, reading 12.00 V on a range where the last digit is 0.01 V. The percentage term is 0.5 percent of 12.00, which is 0.06 V. The digit term is 2 times 0.01, which is 0.02 V. The total tolerance is 0.08 V, so the true value lies somewhere between 11.92 V and 12.08 V. That is the whole calculation. It is worth doing once for the measurement you care about, because it tells you whether the difference you are chasing is real or is inside the meter’s own uncertainty.
This is where a lot of troubleshooting goes wrong. Two meters disagree by 0.3 V on a 12 V rail, and the technician concludes one is faulty. If each meter carries plus or minus 0.08 V, the two readings can differ by 0.16 V purely from specification, and a little more from lead and contact variation. The disagreement is not evidence of a fault. It is the specification doing what it said it would.
The Fluke 177 is a 6000 count instrument, and the listing states the count figure and the display update rates rather than an accuracy number, so the tolerance for a given range is Not stated in the listing bullets. The same is true of the Fluke 116 and the KPS DMM4500PV: the listings describe functions and ranges, not accuracy classes. That is a normal omission in consumer-facing copy, and it means you cannot compare these three on accuracy from the bullets alone.
True RMS is an accuracy question in disguise
An average-responding meter measures the rectified mean of the waveform and scales it by a factor that is only correct for a pure sine wave. Feed it a distorted current from a variable speed drive or a switched-mode supply and the number is wrong, not by a fraction of a percent but by tens of percent, because the form factor of the waveform has changed. A true RMS meter squares the signal, averages it and takes the square root, which is the definition of RMS and is correct for any waveform within its bandwidth.
This matters when you are reading current on a circuit that is not a clean sine. The explanation of true RMS is worth reading before you trust an AC current reading on anything with a motor, a drive or a power supply in front of it. The KPS DMM4500PV is described as true RMS and as measuring AC voltage up to 1500 V, which is a photovoltaic-oriented range; the listing does not publish a crest factor limit or a bandwidth, so how far that true RMS claim extends to distorted waveforms is Not stated in the listing bullets.
True RMS also has a bandwidth limit that the accuracy specification usually does not mention. Above a few kilohertz most handheld meters roll off, so a reading on a high-frequency component is lower than the true RMS value. The specification is quoted for a sine wave at a stated frequency, often 50 or 60 Hz. If your waveform has significant energy above that, the accuracy figure does not apply.
What the meter cannot include in its accuracy figure
The accuracy specification describes the instrument’s own error. It says nothing about the test leads, and the leads are frequently the largest error in a low-resistance or low-voltage measurement. A lead with 0.05 ohm of resistance adds 0.05 ohm to every resistance reading, which is enormous when you are checking a winding that should read 0.2 ohm. The meter’s own accuracy may be a fraction of a percent; the lead error is 25 percent. Good practice is to short the leads together, read the residual, and subtract it, or use a meter with a relative or zero function.
Contact resistance at the probe tips is the same problem in a different place. A probe held against an oxidised terminal, or a crocodile clip biting into paint, adds resistance and can also add thermal EMF in a DC measurement. The meter cannot know. Neither can it know that the circuit is live when you thought it was dead, which is why the voltage test before touching anything is a safety step and not a measurement step. The distinction between a voltage tester and a multimeter is relevant here: a tester is built to answer the safety question quickly, a multimeter is built to answer the measurement question carefully, and they are not substitutes.
Burden voltage is the third hidden error. Every current range has a shunt resistance, and the meter drops a small voltage across it while measuring. In a low-voltage circuit that drop can change the current you are trying to measure. The listing rarely publishes the burden voltage, so for a precise current measurement the figure you need is Not stated in the listing bullets.
Where accuracy stops being the right question
For a great many jobs the limiting factor is not the meter’s accuracy but the stability of what you are measuring. Mains voltage moves by several volts over a day. A battery under load sags. A sensor output drifts with temperature. Chasing a 0.1 percent meter for these measurements is wasted effort, because the quantity itself is not stable to that level. The right instrument is the one whose accuracy is comfortably better than the variation you are trying to detect, and no better.
There are jobs where the accuracy specification genuinely is the point: calibrating a reference, checking a shunt, verifying a 4 to 20 mA loop, or matching a flame sensor current. The Fluke 116 includes a microamp range for flame sensors, and that is a measurement where the current is small and the meter’s offset term matters; the listing states the function but the accuracy on that range is Not stated in the listing bullets. Where a job needs a stated accuracy, buy an instrument that publishes one for the range you will use, and check the calibration interval.
For current in a cable, a clamp is often the only practical instrument, and its accuracy is usually worse than a series meter because the jaw and the core dominate. The comparison of multimeter and clamp meter sets out the trade. The POROMETISTO CM06 is a clamp meter quoting 6000 counts and 600 A AC/DC, with an inrush function that the listing says measures for about 100 ms and applies to AC only; those are useful facts, and the accuracy for each range is Not stated in the listing bullets.
Finally, accuracy is not the same as suitability. A meter can be accurate and still be the wrong instrument: not rated for the category of the circuit, not safe at the voltage present, or not capable of the waveform. The guide to using a multimeter covers the safety and category ratings that decide whether the accuracy figure is even relevant to the job in front of you.
A worked comparison of three meters on the same reading
The table below sets out what the supplied listings do and do not publish for the three handheld meters, so you can see how much of an accuracy comparison you can actually make from the bullets. Where a cell says Not stated, the listing is silent and no figure should be inferred.
| Feature | Fluke 116 | KPS DMM4500PV | Fluke 177 |
|---|---|---|---|
| Display counts | Not stated | Not stated | 6000 |
| True RMS | Not stated | Stated | Stated |
| DC voltage range | Not stated | Up to 2000 V | Not stated |
| AC voltage range | Not stated | Up to 1500 V | Not stated |
| Current range | Microamps for flame sensors | Up to 10 A AC/DC | Not stated |
| Temperature | -40 to 400 C | Stated as a function | Not stated |
| Accuracy figure | Not stated | Not stated | Not stated |
| Bluetooth logging | No | Yes, KPS Link App | No |
The lesson from the table is that the listings are written for buyers choosing on function, not on accuracy. If accuracy is the deciding criterion for your work, you need the manufacturer’s full specification document, not the bullet list, and you need to read the conditions attached to the figure. The difference between a voltmeter and a multimeter is also worth keeping in mind: a dedicated instrument can publish a tighter figure for its one job than a general purpose meter can for all of them.
The number is a claim, and the claim has conditions
Multimeter accuracy is a tolerance with a temperature band, a calibration interval and a waveform assumption attached. Read it as a promise about error, do the arithmetic for the reading you care about, and treat the leads, the contacts and the stability of the circuit as part of the measurement. When the listing does not publish the figure, the honest position is that you do not know it, and the right next step is the manufacturer’s specification document rather than an assumption.
Frequently Asked Questions
Is a more expensive multimeter always more accurate?
No. Cost tends to buy safety ratings, build quality, serviceability and calibration support as much as accuracy. A meter with a modest accuracy figure but a genuine CAT rating and a traceable calibration is often the better buy for electrical work than a tighter meter without those things. The listings here do not publish accuracy figures, so cost cannot be used to rank them on accuracy at all.
How often does a multimeter need calibrating?
The accuracy specification is normally quoted for a stated interval since calibration, often one year. After that the figure is not guaranteed, though the meter may still be well within it. For hobby and general troubleshooting work, many people never calibrate; for work where the reading is used to make a decision that matters, follow the interval the manufacturer publishes.
Can I check my meter's accuracy at home?
Only against something more accurate than the meter, such as a voltage reference or a calibrated source. Comparing two multimeters tells you they disagree, not which one is right. A fresh battery and a known-good reference are the practical checks; anything else is an assumption.
Why do two meters read different voltages on the same point?
Because each has its own tolerance, its own lead resistance and its own contact quality, and the true value lies somewhere inside both tolerances. If the difference is within the combined tolerances, neither meter is proven wrong. If it is outside, check the leads and the contacts before concluding either instrument is faulty.
Does true RMS make a meter more accurate?
It makes it accurate for non-sinusoidal waveforms, where an average-responding meter is systematically wrong. On a clean sine wave at the calibration frequency, both types can be equally accurate. True RMS is a correction for waveform, not a general accuracy upgrade.
What accuracy do I need for automotive or HVAC work?
Usually less than the meter's own tolerance allows, because the quantities are not stable to that level. The functions matter more: microamps for flame sensors, temperature for HVAC, and a DC range that covers the system voltage. The Fluke 116 is built around those HVAC functions, and its accuracy figures are Not stated in the listing bullets.