Current Sensor Module

In order to measure the current we need some sort of current sensor, which can simply be just a short section of wire inserted into the current path. The wire resistance needs to be small compared to the circuit resistance so that the measurements don't alter the behaviour of the system. Since many coilgun circuits have resistances of the order of 1 and run currents of 100A or more, a suitable sensor should have a resistance of say 0.01 or 1% of the circuit resistance. In this case the sensitivity will be 10mV/A. Something to note is that although the power dissipation would be around 100W at 100A you don't need a high rated resistor since the event time is of the order of a few ms. You're not going to easily find a 0.01 resistor but you could wire up several low value resistors in parallel to drop the resistance. Alternatively you could make one using a piece of enamelled wire. What you need to do is decide on the resistance of the sensor then pick a sensible wire diameter, between 1 and 2mm is a safe bet. Now you can work out the required wire length using the resistance formula -

Here L_W works out as 722 mm using a 1.25 mm diameter wire (resistivity of copper is 1.7x10e-5mm). It's best to wrap up the wire so as to minimise its inductance. You can do this by wrapping it back and fourth as shown below. Note that there is an inherent inductance in any conductor - even a straight wire. This arises due to the flux which exists inside the wire (this gives rise to the skin effect) and is independent of the diameter of the wire. The value of the internal inductance is 5e-8H/m which is insignificant compared to typical coil inductances of several uH.

Fig 1. Flat winding

Alternatively, you could wrap the wire so that the turns are wound in alternate directions. This does have quite a lot of inductance compared to other winding configurations so it is best used for low frequencies, if at all.

Fig 2. Opposing winding

The current sensor module is wound with 1.25mm wire, with the current terminals made from brass. The sensor voltage is brought out to a screw-less terminal block.

Fig 3. Current Sensor Parts

Fig 4. Completed Current Sensor

After several attempts at getting the winding length correct the accuracy of the sensitivity is better than +/- 2%. The inductance is less than 1uH.

Version 2:

I made this sensor using constantan resistance wire so that the wire length and inductance would be minimal.

Fig 5. Constantan element sensor.

The accuracy of the sensitivity is better than +/- 2%. The inductance is primarily internal inductance and so it is irrelevant unless the frequency of the current approaches several hundred MHz. The limiting factor with this sensor is the skin effect which causes more charge to flow close to the outer surface of the wire than in the bulk of the conductor - the resistance is frequency dependant. This sensor consists of several strands of wire to reduce the skin effect but the useful frequency range is still limited to about 400kHz.

Version 3:

This sensor design is based on a commercially available 0.01 resistive element. The element is reshaped to reduce inductance and a coaxial lead is soldered directly onto the ends of the element. Fig 6 shows the completed sensor with BNC terminated coaxial lead. Fig 7 is a close up of the sensor element. Note the thin red plastic strip separating the opposing sides of the element.

Fig 6.

Fig 7. Sensor element with plastic seperator strip.