Building a numeric potentiometer for AC current

I'm building a bench to calibrate a device that measures AC currents. As I won't build something accurate, the current will be measured with a current clamp and a multimeter (voltage mode).

For the current load, I first tried a diode bridge with big capacitors and after it, a DC current load (voltage / current converter).

The result is a non sinusoïdal voltage at the clamp output due to the diode voltage which leads to 0 current when the input voltage is less than twice the diode voltage. Like in this screenshot (orange = input voltage, yellow = clamp output (volt)).

The problem is that the multimeter does not integrate correctly the non sinusoïdal voltage of the clamp and outputs a wrong rms value. If I plug the multimeter in-line as an ampere meter, the measure is correct but I need the clamp because I will measure currents that are beyond what the multimeter can do

That's why I want to build a device that operates likes a rheostat. Its resistance should be defined by a microcontroler and the current vaweform should be proportionnal to the voltage waveform. Maximum dissipated power is around 60W (for the purpose of my specific design, I can add a serie resistor to drop the power in the device down to 20W) Input voltage: 6V 50Hz AC sinusoïdal.

How can I do?

• how about resistors? They behave like resistors! – Neil_UK Mar 6 '18 at 9:29
• I would need an infinite number of them. User can choose the value. – Julien Mar 6 '18 at 9:48
• Do you mean a potentiometer or a rheostat i.e. a programmable load? – Andy aka Mar 6 '18 at 9:51
• I suppose what you want to do is actually an AC active load, basically a power resistor with user-settable value, correct? – peufeu Mar 6 '18 at 10:04
• Julien. What is the real problem you are trying to solve? Why can you not use a triac controlled dimmer to vary the output voltage? Does it have to be a sinewave output? Please add the information into your question. – Transistor Mar 6 '18 at 10:11

I will interpret this as an AC electronic/active load, basically a "resistor emulator" for AC. I'm thinking about several solutions:

• Oldskool

Several relays, switching resistors of values in powers of 2 in parallel or in series. Basically a 4-bit number, and each bit controls a resistor. Say you have 4 relays, and resistors of 1,2,4,8 ohms.

If they're in series you can make any value between 1 and 15 ohms.

If they're in parallel, then the binary number controls the inverse of the resistor value, which is basically the current.

• Simple PWm

In this case we would generate a PWM signal, switching a low value resistor in circuit with an AC switch (like a pair of back to back FETs).

Current is proportional to V/R multiplied by the duty cycle, which makes this an adjustable resistor.