CT Clamp sensor interference on ESP32 ADC

Hardware
1

I built a 4-channel home energy monitor using an esp32 and some CT clamps.

signal-2022-07-30-21-34-13-728
signal-2022-07-30-21-34-13-728576×1024 72.5 KB

I based my circuit on this thread: Esphome ct clamp 30A/1V and created four of these on a protoboard. I kept the burden resistor as the two larger clamps I used for my split-phase mains wires did not include them. These are the clamps: https://m.aliexpress.com/item/2251832596888355.html

I hooked these up to two of the inputs, and added a 100A SCT013-030 I had lying around to monitor one more circuit.

It appeared to work well for a while, but today when I was doing calibration with a high amperage heater, I realized that the SCT013-030 clamp was interfering with readings on the other sensors.

For example, if I connect the SCT013-030 up to a circuit pulling 12 amps (a1320 watt heater), suddenly the other CT clamps jump and read 20-30 amps, with nothing connected. I believe this could be noise on my board, but I am unsure how to correct this issue. Please note I have omitted a burden resistor for the sct013-030 input, as that clamp already has one. I am using 10k resistors for my voltage dividers as well as a capacitor to ground. I put a multimeter on the ESP32 ADC pin that is reading the larger CT clamp, and it reads 1.64V but then jumps to over 2V periodically with the third clamp attached.

2

Updating as I have a solution to my particular problem. The issue lies in a distinction between the different versions of the common YHDC clamp sensor.

Most low-amp sensors with the model number SCT-013-xxx come with a built-in burden resistor (SCT-013-030, SCT-013-050). However, the 100A version of this sensor, SCT-013-000, does NOT include a burden resistor!

I had removed the resistor on the bottom-right of my circuit board as I believed the sensor I connected was a voltage-output type. However, as can be seen in this documentation, the 100A SCT-013-000 is of a current output type (with a turn ratio of 100A:50mA, or 2000:1).

I simply moved the sensor to the screw terminals above which still had a burden resistor, re-calibrated, and now everything is working fine.

I learned a lot about current sensing and wanted to share some helpful resources I came across in my hunt for a solution:

Resources on selecting a proper burden resistor

Burden resistor calculator

For reference, I am currently using 10ohm burden resistors on all three active inputs. Per the calculator, a 100A sensor in a 3.3V system with a turn ratio of 2000:1 (SCT-013-000) should use a burden resistor of 23 ohms for maximum accuracy. A 100A sensor in a 3.3V system with a turn ratio of 1000:1 (PZCT-02) should use a burden resistor of 12 ohms. However, resistors are not always commonly available at these resistances.

Per the first link above, the burden resistor should be the “smaller value, or the maximum load current will create a voltage higher than AREF”. Therefore, the 10 ohm burden resistors are ideal for the PZCT-02 CT sensors, but this value is a little low for the SCT-013-000 and this could lead to accuracy issues. In my testing and for my purposes, this is not of concern, however I could change the SCT’s burden resistor to a 22 ohm.

3

Hello, I have on issue, when I use the burden resistor if I turn off the load I still getting values like 0.8A instead 0V, and when a turn on one load with 4.3A the result are very precisely. If I remove the burden resistor I can value more approximated to zero like 0.01A(caused by the noise probability).
Any idea what causes the issue ?