Over the past decade, I’ve had my air conditioners fail multiple times due to refrigerant leaks, bad thermostatic expansion valves (TXVs), failed condenser units, bad start capacitors, and other issues. These repairs are often expensive, and the failures always seem to happen during the hottest days of summer. Therefore, I’d like to catch problems as early as possible.
This past June, I had my third case of a refrigerant leak from an AC coil inside an air handler. While troubleshooting the issue, I noticed that the outdoor unit was drawing about half its normal current: approximately 4 A instead of 8 A on each leg of the 240 V circuit for my upstairs unit. The HVAC company confirmed that the compressor was under very little load because virtually all the refrigerant had leaked out, which explained the significantly lower power draw.
I’m working on an ESPHome setup to monitor temperatures and pressures before and after the air handler, cycle length and frequency, and other parameters to help determine the overall health of each system. However, I wanted to implement something immediately using data I already collect, so I put together the cards below. Both units are currently working correctly:
(Note: The representation of amperage shown above is incorrect. See below.)
For the 240 V outdoor unit, summing the amperage measured on both legs is incorrect because the same current flows through both legs. For example, 10 A on L1 and 10 A on L2 is still a 10 A, 240 V load, not a 20 A load.
Summing the power , however, is correct. The complication is that I’m trying to represent the combined load of a 240 V outdoor unit and a separate 120 V indoor air handler.
Using the numbers above:
Outdoor unit:
10 A × 240 V = 2,400 W
Air handler:
5.4 A × 120 V = 648 W
Total:
2,400 W + 648 W = 3,048 W
Looking at the outdoor unit as the sum of the power measured on its two legs gives essentially the same result:
(10 A + 9.9 A) × 120 V + (5.4 A × 120 V) = 3,036 W
The small difference is due to L1 and L2 not measuring exactly the same current, along with timing differences between sensor samples and screen updates.
However, for amperage I obviously can’t just add the 240 V outdoor unit’s current to the 120 V air handler’s current and call that the total, since the loads operate at different voltages.
As a side note, the Emporia Vue 3 energy meter in my outdoor electrical panel provides V, W, A, and other data for the outdoor units. The Jasco switches powering the indoor air handlers only provide W, so I use a template to estimate A using the corresponding L1 or L2 V reported by the Vue 3. The numbers may not always line up perfectly due to timing differences between sensor samples and screen updates, but they are typically close.
Q1: Is there a better and electrically correct way to represent the total power draw of a combined HVAC system consisting of both a 240 V outdoor unit and a 120 V air handler? Is the best solution simply to show the A of each load separately and use summed W as the overall system load?
Q2: What is the best way to characterize normal power draw and flag abnormal deviations in HA?
Q3 especially for the HVAC experts: How much should I expect power draw to vary during normal operation? These are basic fixed-speed, on/off compressors with no inverter or variable-speed control. I realize operating conditions affect compressor load, but should I expect power consumption to remain within a relatively narrow range under similar conditions, or can it vary substantially even when the system is working correctly?
Looking for examples of what others have done and/or suggestions! Thanks!
