LT1078 Opamp as precision AC current measurement (full wave voltage rectifier)

Hi All,

i am struggling to get this working before i order the parts to do actual test. Forgive me with the design of the circuit but how accurate would this circuit be ?

In the attached ‘Weld cable 1’ is section of the weld cable that would be used for spot welding. U3 acts as differential amp to measure the voltage drop at ‘Weld cable 1’ U1 and U2 are full wave voltage rectifiers as mentioned here by Linear Technology.

The idea here is to measure the amps put into a nickel strip during spot welding 18650 cells. Since the current is very HIGH 2000 - 3000A other alternatives like CT and products like Allegro current sensors would not help.

In the attached change the .txt of LTSpice to .asc

Precision Rect xfmr.txt (2.79 KB)

I don’t understand the U3 operation, with its output shorted to Gnd.

The rectifier should work, measuring the voltage drop across R2 and the weld cable. What’s the purpose of R2, which consumes 8 times the power of the weld cable?

The remaining part of the cable that goes to the electrode, there is obviously some resistance to it as voltage flows through it to the electrode. here its shown in milli volts. just update the schematic a bit more.

Here the voltage drop across the inverting and non inverting pins is about 335mv where the secting of the cable where the pins are tapped on to becomes a shunt the. The output of U1 fed into the non inverting pin of U2 so the signal is amplified and as well keeps it steady should the capacitor start loose charge which in turn gives a steady dc voltage that can be fed into the arduino.

Precesion Full bridge.jpg

1/ - that's a very peculiar precision rectifier circuit - see the old NatSemi databooks for the conventional approach

2/ Using probes directly across the unknown wire resistance is highly unreliable - use a current transformer .

In particular the inevitable sparks on making/breaking the current will cause large voltage spikes due to the leakage inductance ot the transformer driving the current,

  • so

  • pass your main current wire through a toroid with a number of turns of thin wire also wound on the toroid as a secondary - ie make a current transformer. Load this with a known resistor, and feed that into your precision rectifier etc

Total galvanic isolation => your circuitry survives!
..

Allan

Thank you Allan.

1/ - that's a very peculiar precision rectifier circuit - see the old NatSemi databooks for the conventional approach

Had been around for a while and is the simplest form for a peak detector.

Regardless of your choice of the input and rectifier circuit, I'd check the amplification and dynamic behaviour of the circuit. Start with zero current, then turn on the current for some time, then turn it off again, and watch the resulting output. And that for various currents, to get the offset and sensitivity of the circuit. The same can be done with a real circuit, a transformer of low output voltage (5-12V), a pot for current adjustment and a resistor for the weld cable (shunt). Plug in an AC meter for the shunt voltage or current, and a DC meter or the Arduino for the output signal. Still better were a 2 channel scope, but that's not a requirement. Only if that works as expected, connect it to the welding station.

How do you intend to calibrate your device, so that it will show real kA values?

LTSpice seems to give fair results. here is the schematic.
Precesion Full bridge.jpg
Precesion Schematic.jpg

How do you intend to calibrate your device, so that it will show real kA values?

Short the weld electrodes for about 1 second or less and adjust the gain via the pot to the max volt, here 5v

Not sure why you do all of this.
The Arduino that you use for your spotwelding timer should be able to measure (AC) welding current as well.
A galvanic connection to the welding wire seems dangerous.
Allan gave you the solution with a toroid.
Leo..