View Full Version : True RMS current measurements

29th July 2014, 09:13 PM
During the recent bout of electric motor HP versus frequency tests I also recorded the currents involved at different motor loads, both the current going into and out of the VFD as well as the current displayed by the VFD itself. The idea behind this was to see if I could compare the electrical power consumption with the mechanical power being produced and hence determine the efficiency of the motor and the VFD.

One problem I knew about was that a true RMS current meter is required to accurately measure the VFD output current at different frequencies. The Protek multimeter I was using is a true RMS voltmeter but the conventional current clamp (CCC) is an unknown quantity and being a cheap chinese clamp somehow I doubted it would be up to the mark.

Just after I performed these measurements I was discussing them with a techo mate who mentioned that a place in the US was selling Fluke true RMS current meters for about 2/3 to half the going rate so I decided to bite the bullet and get the 324 model. It arrived this morning and I have been playing with it since then, and here I post some test results where I compare the readings of the Fluke with other meters.

I don't intend to perform a full review of this Fluke meter as I don't have the gear to adequately test it anyway and there is more than enough info about it on the web, however I will provide a few details.

The 324 is one of the 3 basic Fluke True RMS meters described here , see Fluke 324 True RMS Clamp Meter (http://en-us.fluke.com/products/clamp-meters/fluke-324-clamp-meter.html#techspecs)
It comes with a reasonably rugged pouch and a massive manual which turns out to be half a dozen pages of basic instructions in a zillion different languages.
Fluke do however provide additional documentation on the web, including application notes and a comprehensive calibration booklet.


The meter includes basic Voltage, resistance. capacitance and temperature measurement capability, a nice long set of rugged leads, as well a thermcouple.
While it's useful to have these features, none of them are that special or make it worth buying as it's main purpose is to measure current.
It measures 0 - 400A with a resolution of 100 mA for >40A and 10 mA for <40A.
The meter is guaranteed accurate to 1.5% between 45 and 400Hz, it will measure below 45Hz but no accuracy is given for sub 45Hz frequencies.

During testing I recorded the output current of the VFD, and the Protek multimeter connected to a conventional current clamp (CCC), and the fluke meter simultaneously, as I increasingly applied load to a 1HP motor connected to the VFD.
The other panel meter displays the current and voltage going into the VFD.


The readings had to be taken simultaneously so I did this using the same photo technique I used for measuring the Motor HP i.e. change the load and wait for it to stabilise and the take a photo.

Then look at the images one by one and enter all the data into a spreadsheet and calculated the % deviation of the VFD current being displayed, from the value shown on the fluke meter, and did the same for the CCC. I repeated this over a range of motor loads/currents and frequencies.

The % deviations are plotted below - the fact that the % deviations are all negative values mean that the currents are being recorded too low by the CCC and the VFD.
The solid lines are for the CCC and the dashed lines are for the VFD current.
The same colour of line refers to the same frequencies.
Unfortunately the graph is a bit messy but it is obvious that both the CCC and the VFD currents are being measured too low by ~ 10 and 20 % respectively.
I've just shown the error bars ( reading error for each meter) for one set of data for each meter otherwise the graph would become really cluttered.


To more easily see if there was a frequency dependence I plotted the average deviation for each frequency and it does look like the CCC has a frequency dependence, while the VFD just reads between 19 and 13% too low.
The error bars are one standard deviation of the mean of the data for each frequency.
The means the CCC and VFD effectively read the same at 30 and 40 Hz but are less likely to do so at the other frequencies.


While the currents at higher frequencies have not been explored, these results are not as poor as I thought they would be.
I will do some more testing WIGRTI.

Of course other VFDs may provide more or less accurate current measurements.

30th July 2014, 09:46 PM
I'd note that the Fluke is accurate to 1.5% plus 5 counts, which for the readings you're taking brings the accuracy to a bit over 3%, which is still very good.