Reliable Power Meters have "trigger on current variation" on the drawing board, but a release date is not yet known. Presented here is a method of accomplishing current event captures until the software boffs come up with a 'real' solution (in both PAS and Scenario please guys!).
The Power Recorder is designed to trigger when the variation in voltage between two successive cycles exceeds 2% (up to version 4.1.106 this was a fixed 2V). Locating nuisance tripping or other related faults can sometimes be missed if the voltage variation is too small to 'trigger' a capture (especially if the source impedance is very low).
If the current variation is causing a breaker to trip then the usual suggestion is to wire the recorder after the breaker so that the trip causes an event to be captured. This will allow up to 2 'pre-trip' cycles to be examined. If it is suspected that the trip was caused through sharp current increases then this will suffice. If it is believed that long duration issues are at stake, or the 'knock-on' effects on the voltage by the sudden release in current is important, then another solution is called for.
One method is to somehow increase the source impedance (not a smart idea!). What we are proposing is using the voltage applied to the current channel and apportion some of this as part of the voltage. This gives us a recorder that captures 'x%' variations on current, regardless of the current being drawn i.e. again no thresholds to set (which we believe is still a far smarter idea!).
The principle is simple and, the fact the Power Recorder is designed the way it is, also simple to implement. The heart of the circuit is a transformer that takes the output of the 'flexi' current sensor amplifier and increase it to a reasonable voltage.
The input voltage is fed through the transformer output such that it negatively sums the input voltage with that from the current. This is done as the voltage would normally decrease with an increase in current. The circuit is simply accentuating this action.
When sufficient change occurs the recorder captures this as a voltage event. As an added advantage this mode captures both serious changes in voltage and/or current.
The first issue is to calculate the change of current required to cause a trigger (the actual current is the percentage multiplied by the full-scale current of the flexi used). The formula is simply
DI% = Vnom(Tsec/Tpri)
Strangely enough, the amount of percent change is directly related to the voltage of the secondary if the primary is the same as the working voltage. Example, should the transformer be a 230:5 then the DI is 5%. In our unit we used a 115+115 to 5+5 transformer giving us 5, 10, and 20% variation settings.
A simple adapter is required that brings out the input to the current channel (shown below).
It is best to test the transformer and this is done using the power recorder in "split phase", feeding the one phase with the direct input, and the second via the transformer and then compare the results.
The voltage as shown on the events and RMS trend graphs is not the true voltage and is why the Neutral-Ground has been employed to measure the voltage generated across the transformer output. To get the real voltage (or as near real as possible) the trends for both phase and Neutral must be exported and added to each other (using Excel and then creating a new graph would be one method).
A similar thing can be done using the current RMS trend, as long as you know the ratio by which the current change influences the voltage. Once the circuit is built it would prove beneficial to do tests in single phase and make notes of such changes, and then keep these handy for when real recordings are done.
Figures that are not to be used are harmonics, THD, flicker, and power flow as these are affected by the pseudo-feed impedance that has been created. Current changes will influence the flicker and voltage harmonics will be a product of both current and voltage values.
Note: This web page does not form part of any official documentation.
© M.T.P. - 12.07.02