Choosing an amplifier

Power
1

Greetings! I’m working on a project and am having difficulties choosing the right components to get the output I need.

I have a YHDC brand SCT013 5A/1V class:1 current transformer and will be using it to measure the current of a 120 V AC line at 60Hz and feeding the output voltage to a microcontroller. The microcontroller only receives positive inputs up to 3.3V. My teammate who is handling the code wants the signal input into the microcontroller to have a range of 1.65Vpp centered at 1.65V so we can measure both the positive and negative magnitudes of the signal. He also wants the input signal to be an AC voltage and will be reading it using an analog input pin.

In order to accomplish this, my current understanding is that I need to take the 1V AC output of the current transformer and apply both a 1V DC offset and a 1.65 V/V gain to the CT output (assuming this gain will apply to both the AC voltage and the 1V DC offset).

Since our project is to construct a power meter we want to have as many data points as possible for our current measurements. We also want to filter out any noise from our microcontroller input signal. We also will need the device to only require 5V maximum for its input but if that isn’t feasible we can try to figure something out.

Does anyone have a component recommendation for us to meet all of these parameters? We will be needing 4 of these so low cost options are ideal, but we are open to $10-$20 price ranges (per unit).

Parameters for needed component:

Signal Input: 0V-1V AC 60Hz

Signal Output: 0Vpp-1.65Vpp with a 1.65V DC offset

Device input: 0V-5V

This is my first forum post so I apologize if it’s hard to find information or long winded. Thank you so much for taking the time to read this, and any assistance is appreciated.

2

The MCP6002 op-amp is a low-cost option and should fit within your budget. You can calculate the specific resistor and capacitor values based on the desired gain, offset, and filtering characteristics.

2 Likes
3

Thank you for the suggestion! Do you know if it operates well at low frequencies? We mostly worked with high frequencies in my classes and labs where we used op-amps so I don’t have much of an idea of how they operate at low frequencies

4

Op Amps are DC coupled and many work best at amplifying from DC to only a few tens of Hz AC.

The particular device @bidrohini suggested is a 1 MHz Op Amp that with a quick scan of the data sheet appears to work well with DC input.

2 Likes
5

Hello Shocking Truth,

Perhaps you can eliminate the op amp using a circuit like this:

Current sensor to microcontroller interface.
Current sensor to microcontroller interface.835×579 83.9 KB

As I understand the SCT013 specification, it is a passive device with no reference to ground. That allows us to float the sensor… The 1/2 VCC reference is obtained from a voltage divider. The voltage from the current sensor then adds to the DC offset. As shown in the schematic, this will give you a voltage that swings from 0.24 to 3.06 assuming the sensor provides 1 V RMS full scale.

Observe that the DC level is passed to the microcontroller on the A1 ADC input. From time to time you will need to read this voltage and then always subtract it from A0 voltage. This action will remove any drift from the power supply. It also relaxes the tolerance specifications for the voltage divider resistors.

Tech Tip: For improved accuracy, use one of the microcontroller’s internal references or better yet, purchase an external voltage reference knowing that you get what you pay for.

By default, many microcontroller use the voltage rail as the reference. That’s a poor solution as the rail voltage changes as the various loads change. A proper voltage reference is independent of the rail. An binary friendly precision external 1.024 or a 2.048 reference may be just what your project needs. Be sure to adjust the voltage divider as necessary to center the DC level for the new ADC reference.

As for filtering, I’m assuming your microcontroller is sampling at least 10 to 100 times faster than the signal of interest. If you capture, square, and then sum the signals together you are well on your way to performing an RMS calculation. It’s not perfect, but the calculation has an implicit filtering. For more accuracy, extend the integration time to perhaps 10 full cycles. However, watch out as there is a balance between accuracy and response time.

Finally, assuming you are looking for a zero crossing to measure phase shift (reactive power). You could capture the samples immediately below and immediately above the zero crossing. Use similar triangles to calculate the zero. Any single answer is likely to be off. However if you average the zero crossing calculations, you should be able to get close to truth.

All of this assumes you are using an interrupt driven process to control your sampling times. Jitter is not your friend.

Best Wishes,

APDahlen

P.S. If you really want a challenge, capture several thousand samples and then perform an FFT to see the waveform spectral components.

4 Likes