230 V AC measurement with Microcontroller

[design.s]

Hi,
I've been trying to measure 230 volt AC with PIC18F4550 to trip a AC generator on overvoltage or undervoltage and display the reading on an LCD.
Currently, I'm using a 230~6V (1VA) transformer along with resistor divider and a diode. But since there's a 0.7V drop across the diode, im going with the precision rectifier as in this figure:

I would like to eliminate the transformer since i will be isolating the board completely with a cover, so there isnt any physical contact of the user with the board directly, and in addition i'll be using MOV and PTC fuses . I prefer not using galvanic isolation to reduce the bulk and cost. Would it be appropriate to connect the AC neutral line to the digital ground (Vss) of the microcontroller directly for ground reference for ADC measurements or is there a way to neutral reference by any other means?


thanks

 

[andre_luis]

I would suggest you perform a rectification at 230v side, once this way the drop voltage take on diode will be divided by the same ratio yielded by resistors net.



+++

 

[design.s]

thanks for the reply andre,

using a bridge, i could get the dc full rectified wave, but I will be developing it further for power factor measurement once this part is done through, by detecting the zero crossing of the voltage and current, so i will have to go with half wave rectification to keep it simple. in that case, again, the microcontroller ground would have to be referred to neutral. Also i want the current consumption to be kept to the minimum.
Would it be possible to use the LM324 itself to interface both line and neutral to the PIC? The answer is somewhere in differential or instrumentation amplifier, but I haven't been able to figure it out.

 

[vilela]

design.s, connecting the AC neutral line to the digital ground (Vss) of the microcontroller is not a big problem if the neutral is efficiently grouded. If not, the neutral voltage can assume pretty high values during surges in the distributions system (if grounded, these values probably will be just high, and the PIC may survive).

If during the PIC coding you need to know if the half cycle is either positive or negative, additional circuitry will be necessary (in addition to the rectifier and divider). In this case, the strategy of "attenuate and shift" may be a good option. Please see the following image.



Note the resistive voltage divider, than the OpAmps. After dividing the voltage (circa of - 2.3 to + 2.3), the first OpAmp acts as a buffer, and the second one shifts up the signal in 2.5 V. The diodes are for protection. In the end, you will have the complete AC signal in the PICs ADC. However, this is not the best option if zero crossing detection is the point (and positive-negative semi cycle reference is not).

If you don’t need positive-negative semi cycle reference, just stick to what andre_teprom suggested: rectification at 230v and then a voltage divider.

 

[design.s]

thanks andre and vilela,
really appreciate the help.