Shunt or Hall effect for DC side current measurement

[frackers]

Greetings

Up to now I've used a 1000 Siemen shunt with a DS2438 1-wire smart battery monitor chip to measure from 100mA up to 200A (the chip will measure to 250A with this value shunt) it dissipates upwards of 40w at full wack.

On the other hand, a Hall effect device, such as a ACS759ECB-200B-PFF-T with a tenth of the resistance will result in lower losses (only 4w) and is fast enough to capture fast pulses on the supply (120KHz bandwidth). It also doesn't have any current loop (common ground) issues that can be a pain in the design of a shunt based system. It can also do high side instead of low side measurement (not that it matters!!)

Does anyone have any preferences? I already have the code for the 1-wire devices (and hooking in extra devices such as temperature sensors would be easy once the basic driver in installed) and the Hall effect devices look tricky to get a good high current connection to.

I reckon Hall effect for the AC side measurements as its self calibrated, I can measure power factor of loads if I can find a 'safe' method (i.e. isolated) of monitoring the mains output voltage as well as displaying and logging power.

Be interested in feedback.

[rossw]

Hall devices sound great, but I've found them far noisier than I was happy with.

The better ones are all chopper-stabilized which helps a lot with stability and drift, but adds to the cost somewhat.
The last several systems I've built all used either an analog meter (for instant visual assessment) or simply an existing cable, as the shunt. I've always thought it backwards, but low-side metering makes things ever so much easier. A simple opamp at low-side gives you a low-noise, low-impedance, ground-referenced signal that goes cleanly into whatever ADC you choose to use, and using cable or meter means no "additional" losses being introduced into the system.