As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
A drive fed circuit needs a true RMS meter because the alternative, an average responding meter, is not computing the quantity you believe it is. It rectifies the waveform, takes the mean of the result, and multiplies by a fixed constant of about 1.1107 that is correct only for an undistorted sine. A variable frequency drive puts out a train of switched pulses, so that constant no longer applies, and the resulting error is not a nudge of a percent or two. Depending on the duty cycle it runs from roughly eleven percent high to around two thirds low. A true RMS converter computes the heating value of whatever shape actually arrives, which is the number a motor winding and a cable respond to.
Why the averaging meter was calibrated for a sine it will never see
An average responding instrument has no way to square anything. Its input stage rectifies the alternating waveform so every half cycle points the same way, then a simple filter produces the mean of that rectified signal. The mean of a rectified sine of peak value Vp is 2Vp divided by pi, or about 0.6366 Vp. The root mean square of the same sine is Vp divided by the square root of two, or about 0.7071 Vp. Divide one by the other and you get 1.1107. That single number is baked into the scaling of every average responding meter ever built, and the display is marked in RMS even though nothing inside has computed an RMS.
The arrangement works perfectly as long as the input really is a clean sine. It was a sensible compromise when almost everything on a distribution system was a transformer, a heater or an induction motor fed straight from the supply. The moment a switching converter sits between the supply and the load, the ratio of mean to RMS changes, the constant stops describing the waveform, and the instrument carries on multiplying by 1.1107 regardless. The longer explanation of what true RMS means on a multimeter covers how the newer converters get around this.
What a pulse width modulated output looks like to an input stage
A drive does not generate a sine. It rectifies the incoming supply into a direct current bus, then reconnects that bus to each motor terminal through solid state switches, alternating the polarity thousands of times a second. The instantaneous voltage between two output terminals is almost always sitting at one of three values: the full bus voltage, zero, or the full bus voltage reversed. The illusion of a sine comes from varying how long each pulse lasts across the cycle, wide near the peaks and narrow near the zero crossings.
A motor winding does not mind, because its inductance integrates the pulse train into a current that is reasonably smooth. An averaging rectifier does mind, badly. The mean of a rectified pulse train scales with the duty cycle, while the RMS scales with the square root of the duty cycle. Those two quantities part company immediately, and the gap widens as the pulses get narrower. The instrument is not broken and it is not drifting. It is faithfully reporting 1.1107 times a mean, on a waveform where that product has no relationship to the heating value.
The arithmetic of a chopped waveform, worked through
Take a bipolar pulse train of peak amplitude Vp with a duty cycle D, meaning the waveform sits at full amplitude for a fraction D of the time and at zero for the rest. The mean of the rectified signal is Vp multiplied by D. The true RMS is Vp multiplied by the square root of D. An average responding instrument displays 1.1107 times the mean, so the ratio between what it shows and the truth is 1.1107 times the square root of D. Run that through a few duty cycles and the picture is stark.
| Duty cycle | True RMS as a fraction of peak | What an averaging meter displays | Error against the true value |
|---|---|---|---|
| 1.00, a square wave | 1.000 | 1.111 | About 11 percent high |
| 0.50 | 0.707 | 0.555 | About 21 percent low |
| 0.25 | 0.500 | 0.278 | About 44 percent low |
| 0.10 | 0.316 | 0.111 | About 65 percent low |
| 0.05 | 0.224 | 0.056 | About 75 percent low |
Notice that the sign of the error reverses between the top row and the rest. That is what makes an averaging reading unusable on a drive rather than merely imprecise. If every reading were eleven percent high you could allow for it. Here the correction depends on a duty cycle that is itself changing with motor speed and load, which means there is no number you can carry in your head.
Crest factor and bandwidth, the two limits a true RMS badge does not cover
Buying a true RMS instrument removes the sine assumption but introduces two new specifications that matter just as much. The first is crest factor, the ratio of peak to RMS. An undistorted sine sits at 1.414, and the rectifier currents drawn by a drive on its supply side routinely run at three or above. A true RMS converter has a stated crest factor limit, and a peak sharper than that limit clips inside the instrument and produces a reading that is low with nothing on the display to say so.
The second is bandwidth. The converter integrates everything it can see, so on a drive output it dutifully includes the switching carrier along with the fundamental. Some instruments publish a rejection or filtering feature to deal with this. The KPS DCM8500PV states high frequency rejection intended to improve performance in electrically noisy surroundings, alongside true RMS measurement of resistance, capacitance, frequency, temperature, continuity, diode and low current signals. The FLIR CM78-PV states true RMS readings together with a drive mode and a low impedance mode, and includes test leads and connectors rated for its stated category and voltage. Neither bullet set states the corner frequency of its filter or a crest factor limit.
What these four listings publish about drive work
| Instrument | True RMS stated | Filtering or rejection stated | Display behaviour stated | Bandwidth or crest factor |
|---|---|---|---|---|
| KPS DCM8500PV clamp meter | Yes | High frequency rejection | Bluetooth recording through the maker app | Not stated in the listing bullets |
| FLIR CM78-PV clamp meter | Yes | Drive mode and low impedance mode | Wireless sharing of field findings | Not stated in the listing bullets |
| Fluke 177 multimeter | Yes | Not stated in the listing bullets | 6000 counts, four digital updates a second, a 33 segment bar graph at forty a second | Not stated in the listing bullets |
| OWON XDM1241 bench multimeter | Yes | Not stated in the listing bullets | 55000 counts, internal memory, remote control over a computer link | Not stated in the listing bullets |
The bar graph on the Fluke 177 is worth a sentence of its own, because it updates ten times as often as the digits and that difference is useful on a drive. The digits settle into something readable while the analogue segments still show you that the quantity underneath is moving. A fast strip of segments jumping around under a calm numeric display is itself diagnostic information. For the separate question of how finely a reading is subdivided, the explanation of resolution in a digital meter covers what counts actually buy you.
Where to put the probes on a drive fed motor
Work the three zones in order and do not carry one assumption into the next. On the supply conductors ahead of the drive the voltage is usually a recognisable sine, but the current is not, so a true RMS clamp is the right instrument there even though a voltage reading from an older meter would look believable. On the direct current bus you need an instrument that measures direct voltage at the level involved and you need to respect the stored energy in the bus capacitors after isolation. On the output conductors you need true RMS and a filtering mode together, and you should note down both the filtered and unfiltered numbers rather than only the one that looks sensible.
Identifying which conductors belong to which zone is often the harder half of the job in a crowded panel, and tracing a run back to its origin before clipping anything is time well spent. The circuit tracer explainer covers the instruments built for that, and the clamp meter write up goes through what a jaw measurement does and does not disturb. The order matters more than the instrument: confirm what the conductor is, confirm the instrument is rated for the place you are standing, then worry about whether the number is a true RMS one.
Frequently Asked Questions
How large can the error from an average responding meter actually be?
It depends entirely on the shape, which is why no single correction factor exists. On a symmetrical square wave an average responding instrument reads about eleven percent high. On a bipolar pulse train at a quarter duty it reads roughly forty five percent low, and at a tenth duty it reads about sixty five percent low. The sign of the error flips as the shape changes, so you cannot even apply a mental fudge in one consistent direction.
Does true RMS matter on the supply side of the drive as well?
Yes, though for a different reason. The drive presents a rectifier and a capacitor to the incoming supply, which draws current in short peaks near the voltage crest rather than smoothly. That current waveform is far from sinusoidal even though the voltage usually still looks clean, so a current measurement on the supply conductors needs the same true RMS treatment as anything on the motor side.
Is true RMS on its own enough for drive output voltage?
Usually not. A true RMS converter will faithfully include every bit of energy inside its bandwidth, and on a drive output a large share of that energy sits at the switching carrier rather than at the fundamental the motor turns on. That is why instruments intended for this work add a filtering mode. Without it you get an honest number for a quantity you probably did not want.
What is crest factor and why does it appear in the specification?
Crest factor is the peak value divided by the RMS value. An undistorted sine sits at about 1.414. Rectifier input currents and some drive waveforms run at three or more, and a true RMS converter has a stated crest factor it can handle before its own internal stages clip. Feed it a sharper peak than that and the converter under reports, quietly, with no warning on the display.
Does a higher count display help with a distorted waveform?
Count resolution and conversion accuracy are separate things. A high count display subdivides the range more finely, which is useful when you are watching for a small change, but it does nothing about a converter that is computing the wrong quantity. Resolution tells you how finely the instrument can express an answer, not whether the answer is right.
Can I use a clamp on the drive output instead of probing voltage?
You can, and current is often the easier measurement to take safely because the jaws never open the circuit. The same two conditions still apply: the current path has to be read by a true RMS converter, and on a drive output you generally want the filtering mode engaged so the reading reflects the fundamental the motor responds to rather than the switching energy riding on top of it.