Thanks, Russell, that is very helpful information. Most of the time when I have had to measure current in a circuit I have used a resistor with four terminals, designed for current sensing. For example, the Ohmite TGHG series (chassis-mount with screw terminals) is awesome (although they are expensive). I have made some DC energy-measuring dataloggers using these which, right out of the box, no calibration, showed no discernible error relative to a 4 digit multimeter in the 1 Amp range, for a 10 milliohm resistor. They are also non-inductive and, if heatsunk, can handle up to 100W. For PCB-mount applications, I have used Vishay's WSL series, both in the four-terminal package that looks like a large-ish traditional SMD resistor, and in the bent metal package, which still has four terminals and surface-mounts. In one of my designs I had to pay significant attention to the PCB layout to maintain thermal symmetry, to reduce parasitic thermocouple effects. However, I was able to get down to 50mA offset, +/- 1% scale factor error, across the initial production lot, out of 10 Amps, with no calibration, and down to a few milliamps with calibration. I did the calibration automatically on power-up - it is a simple zero-offset calibration assuming zero current flow at start-up. This is a battery management system application. Sean On Tue, Feb 12, 2013 at 8:10 PM, RussellMc wrote: > Optimize High-Current Sensing Accuracy by Improving Pad Layout of > Low-Value Shunt Resistors > > When current sense resistors in the order of milli-Ohms are used, > significant errors can be introduced by bad choice of sense trace > locations on the resistor pad and by poor pad design. > > This article is a useful contribution to "how to do it right". > > http://www.analog.com/library/analogDialogue/archives/46-06/shunt_resisto= rs.pdf > > My very recent experience is that even resistprs in the 0.1 Ohm range > can cause problems if poorly managed. > I installed two x 0.1 Ohm SR's (sense resistors) in a product to > measure current. Currents were <150 mA and < 500 mA in each case > giving 0.015V and 0.05V drop in 0.1R. Acceptable in the context. > > I was using 1% through hole parts. To use these in 4 wire sense mode I > added a measurement wire to each lead, wrapped it around the lead > several times, slid it along the lead to make body contact and > soldered it in place so the solder joint was against the resistor > body. Should be perfect, right ? :-). > While this method is not 'approved' and cannot easily be replicated > using a PCB it should work very well indeed. I thought. The resistors > are about 10% low. By making connection at the lead copper where it > would enter a PCB at minimum approved lead length the results would > have been almost correct. Calibrating a resistor so the minimum lead > length is a necessary part of the calibration seems "strange". BUT the > 0.9 Ohm reading suggests this MAY have been done, as it is well > outside the 1% tolerance. > Readings shall be multiplied by 1.1 in this case. > > > Russell > -- > 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 .