Over the past few weeks I have been validating the energy entities used in my Home Assistant installation before adding them to my permanent energy measurement layer.
The goal is to determine how accurately the energy reported by Home Assistant matches a calibrated reference meter.
Test environment
- Home Assistant
- ZHA
- Same electrical load
- Same measurement period
- Same calibrated Kopp reference meter
- Energy compared in kWh
- Percentage deviation calculated
Test methodology
Each device was measured using the same approach and compared against the same Kopp reference meter.
For the Innr SP240 devices I compared the Home Assistant Riemann Sum Integral Helper, because that is the energy entity I actually use in my installation.
For the Shelly Plug Gen3 and Shelly 1PM Mini Gen4 I used their native energy entities.
The purpose of this comparison is to validate the energy entities that are actually used in my Home Assistant energy monitoring.
Current results
| Device | Energy entity used | Test period | Deviation |
|---|---|---|---|
| Innr SP240 + Riemann Integral Helper (Cooker) | Riemann Helper | 43 h | +3.2% |
| Innr SP240 + Riemann Integral Helper (Kitchen refrigerator) | Riemann Helper | 44 h | +0.02%* |
| Innr SP240 + Riemann Integral Helper (CV refrigerator) | Riemann Helper | 44 h | +0.02%* |
| Innr SP240 + Riemann Integral Helper (Combi oven) | Riemann Helper | Test cycle | +12.6% |
| Shelly Plug Gen3 | Native energy entity | 44 h | +0.02% |
| Shelly 1PM Mini Gen4 | Native energy entity | 44 h | +0.02% |
- The Kopp reference meter displays energy with a resolution of 0.01 kWh, while Home Assistant stores values with 0.001 kWh precision. The percentage deviations are therefore calculated from the complete measurement dataset rather than the rounded values shown on the Kopp display.
Based on these measurements I have started replacing the Innr SP240 devices in my Home Assistant energy measurement layer with Shelly devices where measurement accuracy is important.
Next test
The next device to be validated is the Shelly EM Mini Gen4.
Configuration:
- Firmware: 2.0.0 (20260710-101117 / 2.0.0-g87fbfa4)
- Zigbee
- Home Assistant
- ZHA
I will use exactly the same test methodology and update this topic with:
- Test duration
- Reference energy (Kopp)
- Measured energy (Shelly EM Mini Gen4)
- Percentage deviation
If anyone has performed similar comparisons against a calibrated reference meter, I’d be interested in seeing the results.