Measuring VFD Input Current

Evidently you can't use an analog clamp-on ammeter to measure VFD INPUT current, at least the kind I have. I suspect that it has to do with the digital nature of the VFD and the response of the meter. The ammeter agreed pretty closely with the VFD display for the OUTPUT current. It was about 3.9 amps for an unloaded 3 hp motor. But the VFD INPUT was about 0.8 amps, which is obviously wrong. When the motor was not running, the VFD INPUT current wouldn't even move the needle on the ammeter, showed 0 amps on the 6 amp range. It isn't important that I know the INPUT current, but I was curious about what was going on here.

The problem is the nature of a "rectified load". It isn't anything about the "digital nature", that would be more of an issue on the output if it were an issue anywhere.

But, especially at low currents, the rectifier pulls current in short pulses. Those short pulses are mostly made up of high frequencies (harmonics), much higher than the 60 Hz The usual analog meter like an "Amprobe" is good for low frequency, because it is intended to measure 60 Hz current, with a few harmonics.

When the current is MOSTLY harmonics, the meter does not respond to most of them, and may read completely wrong. You'd probably get a much closer match between the meter indication and actual if the VFD were powering a heavily loaded motor, because the hevy load will "spread out" the pulses of current so they have much more low frequency content.

Any decent "true rms" meter will read many more harmonics and give a much more true indication.

The meter reads closer on the "digital" output, because the motor inductance forms a filter which eliminates most of the high "digital" harmonics, leaving frequencies that the meter can read.

Reasons:

The rectifier charges up a capacitor in the power supply. The rectifier conducts current only when it has voltage difference across it in the "forward" direction. With NO load, the capacitor charges up to the peak of the line voltage, and then there is no more voltage difference, so current stops flowing.

With just a small power drain from the capacitor, the voltage does not drop much. Therefore there is no "voltage differerence" across the rectifier until the input voltage is very nearly at the maximum peak voltage. When the voltage goes above the capacitor voltage, the rectifier conducts and "refills" the capacitor, at which point current stops again. So there is a very short "pulse" of current, which is mostly rather high harmonics of the 60 hz. Analog meters often do not read those high harmonics well, if at all.

With a hevier load, the capacitor voltage drops farther, so the current pulse starts sooner, and lasts longer. That means the pulse has much more low frequency content, that an analof meter can read.

If true rms meter then it should read correctly as jst stated.

Why do u think .8 is wrong unloaded? Sound a bit HIGH to me...

Remember 3.2 a output is with low PF; it includes prob 2.5a magnetize ng (not real) current plus the .8amp real current. Only the real torque producing current comes from input. So .8amp sounds about right or a bit high. I would have guessed around 0.5 amp....

Sent from my SM-G900V using Tapatalk

A bit of both. I'm reasonably sure the meter is not reading right, although it actually might not always be low at that condition.

It's not just torque producing current, it's also losses. The motor has about half the full load current so about 1/4 the copper losses as full load. To that is added the iron losses that are more constant, and windage, which can be significant since the fan is moving a little MORE air than it does at full load (slip is less, speed is several percent faster, and fan losses are proportional to speed cubed).

You'd expect maybe 9 A full load current, and an efficiency of maybe 80%. Of the 9A, 20% or so is loss. So around 2A supplies losses. At idle, very little is really producing output torque, the bulk is losses.

Of the losses, a fair bit, but not all is copper loss, the rest is fairly constant at any load. If 2/3 the losses are copper losses, then you'd expect that about 1 1/3 A is supplying copper losses. With only 1/4 the copper losses, that would leave about 2/3 A supplying iron, friction, and fan losses, plus about 1/3 A of copper loss. Any actual utput torque to supply drive train friction is added on.

So you can make a case for about 1A of current at no load that has to be supplied by the line. He's measuring 0.6 A.

And without knowing the out speed, we don't know outvolts. So if he was running 1/2 speed, the input current, losses and torque, get cut in half yet again.

Sent from my SM-G900V using Tapatalk

I had a few minutes to play so tried my little cheap $75? "true rms" clamp on Tenma 72-7226 on 40amp scale...

My Barn fan is a 1.5hp 230v 3ph, on small hitachi vfd. My demo motor on my desk is a blackmax 1/4hp wired 480v .58amp rating.

Iin/Iout are the clamp on readings, DriveA & V are the drive parameter reported

ITEM HZ Iin Iout DriveA DriveV comment
---------------------------------------------------------------------------------
FAN 6 .7 4.1 1.8a 27v
FAN 60 8.5 5.0 3.5a 223v full motor load

DEMO 6 .22 8.0 .3a 38v no load (torque amps I sq=.1a)
DEMO 60 .36 6.2 .3a 335v no load (torque amps I sq=.1a)

So goes to JST point that these little clamp on ammeters, even marked true rms, leave quite a bit to be desired! Wish now I took the time to use a better clamp on as 2nd test!

"True RMS" is now just a marketing term in my experience, there never was an official industry definition of what that means. Fluke originally coined that term and it meant something back then (and to them still), but now just about everyone puts it on their marketing material even for the cheapest meters.

That said, I think it's more likely to be along the lines of what JST first said, that it might not be wrong. My rule-of-thumb is that if a motor is UNLOADED on a VFD, the VFD input current is only around 5-10% of the motor FLA, even though the VFD output current appears to be much higher. That's again because unloaded, motor current is mostly reactive, as in the motor having a .2 power factor, maybe lower. The reactive current is supplied by the capacitors in the drive, the active current is very very low, hence the low input current. Combine that with a cheap meter that has a low band width that might be off by 10%, and at no load it reads so little that a small error looms large.

The biggest issue with cheaper meters and "true rms", is what is called "crest factor".

That is essentially the difference in peak voltage between your actual waveform, and the peak voltage of a sine wave. The meter has a crest factor, which is the maximum peak amount larger than a sine wave that it can accept as it does its "rms conversion'.

As always, crest factor in a meter costs money.

Those short, high peak currents that a lightly loaded VFD draws tend to have a high crest factor, and cheap meters may just run out of "headroom" and fail to measure the entire wave, showing up with a low reading, or possibly with a high reading, depending on how the meter innards react to being overloaded.

Motors DO read low on the VFD input, because a VFD draws current based on the POWER output. At no load, the power output to the motor is as low as it can be for that motor and voltage. But the biggest variable power in a motor, aside from the load itself, is the copper loss. The fan, and iron losses are fairly constant as the voltage and speed are fairly constant, so they are not reduced at no load. Both are, in fact, slightly higher at no load than at full load. .

No telling. But it is a fair bet that the meter is not telling the real truth unless it is a fluke or similar quality, especially with a high crest factor current..