Hi John, Beware that copper wire doesn't make a great shunt because the resistance is very temperature dependent. You can find very nice shunts on eBay for not much money - they are made from copper alloys which have almost zero temperature coefficient of resistance around room temperature. It is entirely possible to run >30 Amps through PCB traces. I have a design which runs 125 Amps through a 1 inch wide trace BUT that is 2 oz copper and the trace is duplicated on several layers and they are paralleled with stitching vias. However, you should note that Allegro Microsystems makes a wide range of hall-effect current sensors and some of them have wide tab leads that can be soldered directly to wire instead of a PCB, like these: http://www.allegromicro.com/en/Products/Current-Sensor-ICs/Fifty-To-Two-Hun= dred-Amp-Integrated-Conductor-Sensor-ICs/ACS758.aspx How accurately do you need to measure current? Are you trying to keep track of amp-hour throughput? Do you have a significant amount of time at very low current? If you have current in the range of 30 Amps and you only need about 1 Amp accuracy, then I would definitely say that Hall effect is the way to go. Shunts will allow you do get better accuracy at low current. To measure current flow in both directions, I would float the ground of the entire datalogger and reference it to one end of the shunt. If the shunt resistance is small enough, then the peak drop across the shunt could be kept less than half a volt (+ and -) and there are many instrumentation amps which can handle small common mode voltages below their negative supply rail, or you could even make your own diff amp (since the common mode voltage would be in the same range as the differential mode so not very difficult to reject) with 0.1% resistors or similar and get very good accuracy. The product I mentioned above that has the 1 inch trace carrying 125A is actually a battery management system which does pretty much exactly what you are trying to do - it has two gain paths - one which measures +/- 10 A with 5% of reading plus10mA accuracy, and the other that measures +/- 150A with 5% of reading plus 150mA accuracy. For each sample in time it chooses between the low range and the high range path depending on what the actual current is. It uses a PCB-mounted shunt which is only 150 micro-ohms along with an auto-zeroing op-amp in a diff amp configuration. Communication with the rest of the system is via isolated CAN so the measurement circuit can be locally referenced to one side of the shunt. I still use a diff amp because there would be additional error due to power supply current returning through the four-terminal Kelvin connection sense traces otherwise. Unfortunately, this isn't a product we sell separately, it is part of a larger robot, so I don't think I can get you one, but I am willing to help if you go down that path. Still, as I said, unless you need sub-1A accuracy, I would stick with Hall sensors. Sean On Tue, Jul 7, 2015 at 8:47 PM, John Gardner wrote: > Very kind of you, James, but my "skills in the analog domain" are > > quite modest; don't mind improving them, though. > > Pololu is right down the road from me; running 30+ amps through > > PCB traces does worry me a bit, I confess - My 1 milliohm shunt > > is ~19" copper #8 AWG, which seems unlikely to contrive a failure > > mode that'll keep me from getting home... > -- > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist > --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .