Using a Home Energy Monitor to Find a Problem Load

Why this matters

"My bill went up and I don't know why" is a call every trade gets, and most shops answer it by checking their own equipment and leaving. A home energy monitor turns that call into something you can actually solve: it shows what the house draws at 3am when nothing should run, which loads cycle and how often, and which circuit holds the surprise. This procedure works one real house through each step, so the numbers carry from the first reading to the fix. It assumes a monitor is already installed; putting one in is work inside the panel for an electrician.

Step 1: Confirm what the monitor sees before you read it

Find out whether you are looking at one whole-house measurement or at individual circuits, and check that the configuration is believable. The quickest check: run a known resistive 240 V load, such as an electric water heater or a dryer on high heat, and confirm the monitor shows close to its nameplate wattage. Half or double means a configuration error, and every later number in this procedure inherits it. See related: Circuit-Level Monitoring and Smart Panels, which owns why that happens and how far to trust a small reading.

In the case house: a whole-house monitor at the service plus four circuit sensors, one of them on the electric water heater. The water heater's upper element is rated 4.5 kW on its label, and the monitor showed 4.5 kW when it fired. The configuration was sound. The customer's complaint: daily use about 30 kWh in a mild month when the house had run nearer 22 kWh the same month the year before.

Skip this step and you can spend an hour hunting a load that exists only in a misconfigured app.

Step 2: Read the overnight floor

Open the history for the hours between about 1am and 4am on a night nobody was up. Two things are visible. The floor is the lowest level the line returns to and holds, which is the house's always-on draw, also called standby or baseload. The spikes above the floor are loads that cycle: refrigerators, freezers, a furnace or air handler blower, a sump pump, the water heater.

Read the floor across three nights, not one. One night can include a dehumidifier or a guest.

In the case house: the floor held at about 0.38 kW on all three nights. Over 24 hours, 0.38 kW is 0.38 times 24, which is about 9.1 kWh a day before a single light goes on. For a house using 30 kWh a day, that is close to a third of the total, and high: a modest house with a couple of refrigerators, a network and some electronics on standby more commonly sits between about 0.1 and 0.25 kW. Treat that range as a starting comparison, not a standard. A house with a well pump, an aquarium or medical equipment will legitimately sit higher.

Step 3: Find the standby loads, one breaker at a time

With the house quiet, switch breakers off one at a time and watch the whole-house reading drop. Wait about a minute at each so cycling loads do not confuse the result, then switch it back on before moving to the next.

Hazards at this step: operate breakers only with the panel's inner cover (the dead front) in place, and never remove it. If a breaker is hot to the touch, discoloured, buzzing or smells of burning, stop, leave it alone and refer the panel to an electrician; that is a separate problem and a bigger one than the bill. Before you start, ask what in the house must not lose power: medical equipment, a sump pump in a wet season, an alarm panel, a freezer for more than a few minutes, the network if someone is working from home. Skip those circuits or warn the customer first. And note that switching off a circuit with a computer or a network device on it without warning can lose someone's work.

In the case house: the drops were about 0.07 kW on the home office and television circuit, 0.03 kW on the circuit feeding the network equipment and alarm panel, and 0.15 kW on an outdoor circuit that, as far as the customer knew, fed only porch lights that were off. The remaining 0.13 kW was spread across a dozen circuits in amounts too small to separate. Those four figures add to 0.07 plus 0.03 plus 0.15 plus 0.13, which is 0.38 kW, the floor from step 2.

The outdoor circuit was the find. Following it led to an exterior receptacle by the back steps with a roof heat-trace cable plugged into it: a heating cable that runs along gutters to prevent ice dams, installed by the previous owner and left plugged in through spring and summer. Many heat-trace cables have no thermostat and draw continuously whenever they are plugged in. The customer unplugged it from the receptacle at ground level; nobody needed to climb.

At 0.15 kW all day, that cable alone was 0.15 times 24, which is 3.6 kWh a day.

Step 4: Look at the cycling loads for a pattern that changed

Standby is half the story. Now look at the spikes in the same overnight window, circuit by circuit, and ask whether any load cycles more often or runs longer than it should when the house is quiet.

A tank water heater with nobody drawing hot water should fire only occasionally overnight, to recover the heat its tank loses through its insulation. How often depends on the tank, its insulation, its temperature setting and the room it sits in; on a reasonably insulated tank, a single recovery in a four-hour quiet window, or none, is a common pattern. An element firing every half hour or so with no hot water being used means hot water, or the heat in it, is leaving the tank somewhere.

In the case house: between 1am and 5am the water heater fired about every 40 minutes, for about 6 minutes each time. In 240 minutes that is 240 divided by 40, which is 6 cycles. Six cycles of 6 minutes is 36 minutes, or 0.6 hours, of element time. At 4.5 kW, 0.6 hours is 0.6 times 4.5, which is 2.7 kWh in four quiet hours.

Step 5: Find what is pulling the heat out

The monitor tells you something is wrong with the water heater's pattern. It does not tell you what. Common causes of overnight cycling with no draw, cheapest to check first:

• A hot-water recirculation pump running around the clock, sending hot water through a loop of pipe that loses heat the whole way.
• A temperature and pressure relief valve weeping, with hot water leaving down the discharge pipe.
• A hot-water leak, including one under a slab or in a wall.
• Thermosiphoning through a recirculation loop or a failed check valve, which moves heat without a pump running.

Hot water and its piping can scald: read pipe temperatures with an infrared thermometer rather than by hand, and treat the relief valve discharge pipe as hot.

In the case house: an infrared thermometer at 3am read the hot line at the farthest bathroom close to tank temperature, and the return line beside the water heater was warm too. A recirculation pump on the return was running; its plug-in timer had been set to "on" permanently, overriding the morning and evening schedule it was installed to run. Resetting the timer to two windows, early morning and evening, stopped the loop running all night.

If the pump had been off and the far line still hot, the next suspects would have been thermosiphoning or a leak, and a leak goes to a plumber with a pressure test, not to more monitor reading.

Step 6: Put the finds in the customer's units

Convert each find to kWh a day and as a share of their use, and be explicit about what you measured and what you are estimating.

In the case house:

• Heat-trace cable: measured as a steady 0.15 kW, so 3.6 kWh a day. Against the 30 kWh daily total, 3.6 divided by 30 is 12 percent.
• Recirculation loop: a week later, with the timer reset, the same 1am to 5am window showed one water heater cycle of about 6 minutes, which is 0.1 hours at 4.5 kW, or 0.45 kWh. Before, the same window used 2.7 kWh. The difference is 2.7 minus 0.45, which is 2.25 kWh, in four quiet hours alone. The loop also lost heat during the rest of the night and during the day, but daytime hot-water use masks it, so the honest statement is "at least about 2.25 kWh a day, likely more". At least 2.25 out of 30 is at least 7.5 percent.

Together the two finds account for at least 3.6 plus 2.25, which is 5.85 kWh a day, against an increase of 30 minus 22, which is 8 kWh a day over last year's same month. So the finds explain most of the increase, and not all of it. Say so. The rest may be weather, occupancy or something not yet found, and pretending the two finds explain the whole bill sets up the next argument. See related: Reading Energy Data Without Misleading the Customer, which owns the comparison against last year and weather normalization.

Step 7: Confirm with a second instrument before anything is replaced

Nothing in this house needed a part, but in a house where a find points at a failing appliance, confirm it with a separate measurement first: a plug-in watt meter for a plug-in load, or a clamp meter reading at the appliance, taken by someone qualified to work at that equipment. The monitor locates. A second instrument confirms.

How to verify you got this right

Read the overnight floor again across three nights after the fixes. In the case house it should drop by about the heat-trace load, from 0.38 kW to about 0.38 minus 0.15, which is 0.23 kW, and it did, landing between 0.22 and 0.24 on three nights. The low end of that band is the pump's own small draw, which had been hiding in the 0.13 kW remainder, now switched off overnight by its timer. The water heater should show one cycle or none in the quiet window, and it showed one on two nights and none on the third. Then check the daily totals over the next two weeks against the same weeks last year, weather-normalized as the sibling card describes. If the floor did not move by the amount you predicted, a second load shares that circuit, and step 3 needs repeating on it.

References

• Manufacturer documentation for the energy monitor, including sampling interval and configuration of 240 V circuits
• Manufacturer installation manual for the water heater and recirculation pump, including timer and control settings
• See related: Circuit-Level Monitoring and Smart Panels
• See related: Reading Energy Data Without Misleading the Customer