** Introduction: **
Wattmeter is an important measuring instruments. It allows to measure true electric power (wattage).
Determining the true power in AC circuits can not be made simply by multiplying the RMS voltage and current, because
power factor is usually not equal to one. You must use the meter that continuously measure
instantaneous current and voltage, multiplies and produces a mean value. Analog electromechanical devices do that
using through the coil (solid) and voltage coils (moving with the pointer). The magnetic force acting between the coils is equal to the product
of magnetic fields. Analog wattmeters are not too accessible, they never measure with great precision.

** Electronic wattmeter theory: **
I decided to make it solid state way and build an electronic power meter with analog sensing and digital reading.
Display is provided by a digital multimeter, which can now be bought under CZK 100, and therefore does not make sense to build your own
digital voltmeter. It is also possible to use panel digital voltmeter or a common analog uA meter.
Immediate voltage is sensed using a voltage divider. The current is sensed by shunt. The voltage and current are then multiplied by an analog multiplier AD633.
The output provides a voltage proportional to the instantaneous power. To obtain the average power is necessary to filter it using RC filter.

The biggest pitfall in circuit design is generating the product of two analog voltage that is not so simple as
might seem. There's possibility of multiplication using operational amplifiers and discrete transitions diodes or transistors, which have an exponential characteristic.
Their principle is logarithm both signals, add them and finally de-logarithm. Accuracy is not too good, there are problems with calibration, a huge
temperature dependence and differences between individual pieces of transistors or diodes. That's why I rejected this option. Another option is
using pulse-width multipliers, but this solution is also quite awkward. Even more complications occurs when it is necessary to work
with both polarities of current and voltage (4 quadrants). So I decided to use a specialized integrated circuit AD633 (AD633JN in classical THT case DIP8),
a four-quadrant analog multiplier with differential inputs and precision of 2%. For more info see AD633 datasheet.
Note, that SMD version has a different pin layout!
The output voltage is given by a function:

** w = (x2-x1) * (y1-y2): 10V + z **

I wanted to try the integrated circuit MPY634 with an accuracy of 0.5%, but I couldn't find one.
Maximum voltage range at which the circuit AD633 operates precisely is +/- 10V. This must match both the input voltage.
The circuit must be designed to the amplitude of current and voltage, not only an effective value. The mains voltage therefore expect
with 325V, not only 230V. As best shown with a divider ratio of 1:40, which allows you to work with peak voltage up to 400V.
Shunt voltage is lower than the voltage of voltage divider, so is connected to the input Y, which has greater accuracy.

** Simple Wattmeter circuit: **
Fig. 1 is the simplest power meter (wattmeter) design with the AD633 and a single range.
The current is sensed by shunt. If we require the output signal conversion 1mV/1W, the value of the shunt would be calculated
0R4. Maximum effective current through meter is determined by the maximum allowable shunt loss.
For 40W shunt max continuous current is approximately 10A. Max. measured power is 2300W for ideal resistive load, for different loads must be lower.
Another limitation is the maximum input voltage 10V of multiplier,
so that the maximum momentary peak amplitude must be below 25A.
Calibration is done by setting P1 according to known loads. The sum of the values of P1 and R1 will be about 390k and the dividing ratio of 1:40.
If you can not set the correct value, change R1. Multiplier inputs are protected against overvoltage by 12V zener diodes.
Supply voltage +/- 15V is capacitively knocked down from the network and stabilized by 15V zeners.
In combination with a conventional multimeter with a resolution of 0.1 mV you will get a wattmeter with a resolution of 0.1 W. We will use
200mV, 2V and maybe 20V ranges where the power
is displayed directly in watts (1mV = 1W) or kilowatts (1V = 1kW).

Warning! The whole circuit including the output for a multimeter (voltmeter) is electrically connected to the mains voltage, which is deadly dangerous.
It should be handled accordingly.
A fuse or circuit breaker should be used to reduce the risk of fire.
You do everything at your own risk. I do not take any responsibility for any of your harm.

Fig. 1 - The schematic of simple wattmeter

AD633JN in DIP8 case.

First experiments with wattmeter

Wattmeter prototype in breadboard.

Wattmeter on the PCB board.

Video - testing Wattmeter

Added: 7. 3. 2011

home