I've been pondering current....

CannibalRobotics

I want to measure battery current on three large battery systems, dual gang 12 and 24 volt systems. I don't want to use a load cell because the batteries are in the bilge of a salt water boat and any new connections or components increase the likelihood of failure plus the whole dissimilar metals in salt environment chemistry.
So, a passive approach is desired. I'm looking at either making my own inductive wrap about the battery wire or using some of the hall effect sensors I've seen. Sensors are pretty pricey and seem to be hard to get in small quantities. Anybody have any direct experience in something similar?

erco

You can measure the voltage drop across an existing length of wire and deduce current from that. The longer the wire, the more voltage you'll get from any given current.

Mark_T

I have tried cutting a slot in a ferrite-ring core and glueing a Hall-sensor in the gap. It did sort of work but the ferrite seemed to have rather more hysterysis than I wanted (several percent I think).

The inductive wrap idea only works for AC - I presume you are interested in DC current. If the currents are large enough you might get away with gluing a Hall sensor onto one of the wires (magnetic field lines go around the wire). Remember to get an analog Hall sensor (and its easy to calculate the field strength from the current and distance from centre of the wire so you can get a good estimate of sensitivity for a given sensor and geometry).

BTW a load-cell is a force sensor not a current sensor.

CannibalRobotics

Well, the wire is a big fat 2ga and I've only got about 8-12" to work with from battery to buss. So resistance measurement would have to be pretty accurate. I think I'd also have to look at Temp compensation at those voltage levels.
On the inductor idea as DC I was thinking about looking at change then using some sort of integrator math. However, your hysteresis comment is well taken, that would probably not work.
Your right on the 'load cell'. I guess my mechanical lobe was typing at that second; "shunt resistor".

erco

I imagine you've got some serious current flowing through 2-gauge wire. Throw a multimeter measuring millivolts across your longest wire and see whatcha get! You can tweak an ADC to get meaningful readings. Not sure that temp compensation is required. Without rewiring, the best you can do is a good estimate anyway.

CannibalRobotics

Starter motors for big block marine engines - lots of current.
I'm thinking your right.
Without a real shunt I'm down to turning the cable into a shunt.