0
\$\begingroup\$

To avoid XY question, I start with my motivation and then my question.

I have an Emporia Vue2 (in-panel) power sensor but it is known not to be very accurate. For me there is a discrepancy between the sum of the branch circuits and the main phases. The is a discrepancy between my Enphase solar sensor (also known not to be accurate). And there is a discrepancy with my utility power meter. My firmware (ESPHOME) supports calibrating the power sensors.

Option 1 (most obvious): Measure with right equipment. But that’s not simple: current and voltage have to be measured in parallel, multiplied and integrated. Required equipment is expensive and not DIY friendly.

Option 2: Use utility meter. While I don’t know how accurate it really is it’s not really practical because I have to power off all but one circuit. Furthermore, power is shown in kW, so load needs tk be big. And lastly, it would require a somewhat constant test load which stays constant

Option 3: Test load. Ideally constant power load. But these are not trivial either. Simplest might be a resistor. Then the voltage across the resistor could be measured via multimeter (somewhat accurate) and power determined via V^2/R. But there are some catches too: Resistor needs to handle high power and voltage, needs to be very accurate and needs to be on the order of 10…10kOhm to present a load of 1W…1kW. I could not find a proper part.

More problematic, a single resistor would only provide one datapoint. Ideally I’d measure datapoints like 0.1, 1, 5, 10, 100, 1000W and possibly do a quadratic correction (if it’s nonlinear).

What is the easiest way to present a very accurate (desired 1%) power load to a 120V AC circuit?

\$\endgroup\$
1
  • 1
    \$\begingroup\$ For option 1, there exists very good and cheap plug in type power meters which does everything for you. If the accuracy is not good enough, you can probably spend some effort and come up with a calibration factor. If still not good enough, try a Voltech PM100. \$\endgroup\$
    – winny
    Commented Jun 25, 2023 at 13:19

2 Answers 2

Reset to default
2
\$\begingroup\$

An electric kettle makes a good load for a kW or so. They are made with a resistance alloy wire, so tempco will probably be quite reasonable. Measure the resistance of one when cold, and when filled with boiling water, to see what the tempco is and whether it needs correction. Borrow a few from neighbours and put them in series to get down to a few hundred watts.

Resistors are relatively inexpensive, buy a bunch of low power wirewound resistors (in my part of the world, 5 W wire-ended is the sweet spot of best power per money) and that will get you from a few watts to 100 watts, by judicious series or parallel connection.

If that gap from the top of the resistors to the bottom of the kettles is too far to interpolate, then travel kettles generally run in the several hundred watt range.

Don't try to use filament light bulbs, motors, or anything electronic as loads, stick to simple, predictable, metal resistors.

\$\endgroup\$
3
  • \$\begingroup\$ (Incandescent lamps made for nice loads with somewhat predictable current.) \$\endgroup\$
    – greybeard
    Commented Jun 25, 2023 at 14:41
  • \$\begingroup\$ @Neil_UK This is a great answer. Do you have a suggestion of the nonlinearity of such a Current Transformer characteristics? Is it likely nonlinear and if yes in which region? I am trying to decide how many test points to choose and how to distribute them. Even 1% 5W resistors are way more expensive than I thought so I need to be very mindful with the measurement points. \$\endgroup\$
    – divB
    Commented Jun 29, 2023 at 7:29
  • 1
    \$\begingroup\$ I measured both an iron-cored and ferrite cored transformer for linearity as current transformers. They were both reasonably linear for several decades of current in the middle of their range. Both steadily declined at the top end due to (I suspect) saturation causing a varying leakage reactance. Interestingly, the iron cored one declined at the very bottom end, where the excitation was not enough to take the material round its hysteresis loop. This was a power transformer, a dedicated 'current transformer' core ought to be better. The ferrite one worked well down to low currents. \$\endgroup\$
    – Neil_UK
    Commented Jun 29, 2023 at 8:44
0
\$\begingroup\$

Get a bunch of incandescent bulbs, or boilers or something other ohmic and resistive spanning the desired Watts range.

Then measure both the voltage and current. That way you don't need precision resistors or anything like that.

You can measure the current with the same AC voltmeter, by using a single small precision shunt resistor. This resistor only needs a medium power rating.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.