Copper wire as a current shunt?

[BiOzZ]

I have made my fair share of current shunts with a set of 12,14,15,16 AWG spools labeled suspiciously as "low resistance shunt wire" ... with a label maker ... but i need a low resistence high current shunt for a personal project and i want to use materials i have instead of materials from work/school

i have a spool of 15AWG solid copper wire used for house grounding ... i worked out using loose math and with allot of expectations of perfection that 6 inches of the wire will have a resistence of 0.00159 ohms giving a rounded output voltage of 0.04v at 25A and with an instrumentation amp with a gain of 130 i will get 0-5v out i should get about 5v ... ish out at the 25A giving me a resolution of ... with a 10 bit ADC of 25MA or so if we lived in a perfect world

im not looking or remotely expecting to get 25ma ... 250 will do me really ... and im not expecting the math to work at all because my guy feeling is it will be less than the calculated ohms so i plan on sticking in a 10 turn pot over the gain resistor

anyone have ANY if this will work with any reliability? im going to only use it in lab for high current measurements so the temperature drift should only be 1 or 2 *f

if anyone is kinda sure or thinks it should work ill build it up and make a post about it!
if it simply wont work or will be ungodly unreliable i wont even bother!

[Psi]

I used 1.5 meters of thick copper wire as a current limit in some crazy spark wands i made (shorting out 40cm electrodes connected to 9x 2.5V 2600F boostcaps in series)

The caps were rated to 600A max discharge so i selected the wire length and gauge to limit the current to those specs.

It worked perfectly, i just had to be careful not to burn myself on the wire 

[Harvs]

Only thing you need to be aware of is the high tempco of copper.

[amspire]

From memory, copper wire has a 0.4%/C coefficient. That means that if the wire has a 100 degC temp rise, it will increase by 40%.

25A in 0.0016 ohms is 1W. If you put 1W into a bare copper wire, it will get a lot hotter then 100 degC temp rise.

So forget about a 10 bit ADC. Most of the bits will be meaningless. 

Perhaps a 2 bit ADC will be about all you can manage with the wire.

If you do try it, you have to try a way of conducting the heat away from the wire.  Where copper wire can be useful is where you actually want the resistance to increase with temperature. A bit like a PTC thermistor. Even with no added heat in the wire, if the ambient can go from, say, 0 degC to 40degC, that means a change in the resistance of 16%. If you can keep the temperature rise to 10 degC, that means you end up with an accuracy of 20%.

You are much better off buying a current shunt resistor.

[BiOzZ]

From memory, copper wire has a 0.4%/C coefficient. That means that if the wire has a 100 degC temp rise, it will increase by 40%.

25A in 0.0016 ohms is 1W. If you put 1W into a bare copper wire, it will get a lot hotter then 100 degC temp rise.

So forget about a 10 bit ADC. Most of the bits will be meaningless. 

Perhaps a 2 bit ADC will be about all you can manage with the wire.

If you do try it, you have to try a way of conducting the heat away from the wire.  Where copper wire can be useful is where you actually want the resistance to increase with temperature. A bit like a PTC thermistor. Even with no added heat in the wire, if the ambient can go from, say, 0 degC to 40degC, that means a change in the resistance of 16%. If you can keep the temperature rise to 10 degC, that means you end up with an accuracy of 20%.

You are much better off buying a current shunt resistor.

i thought the tempco of copper was 0.0039%/*c

and i can guarantee you its not going to reach anything around 100*c even with 1 watt ... thats going to disperse very very quickly when coiled up

im going to have to build this up and do some tests ...

[kfitch42]

http://en.wikipedia.org/wiki/Electrical_resistivity_and_conductivity

Looks like it could be anywhere from about 0.4%/C to 0.7%/C

[muvideo]

i thought the tempco of copper was 0.0039%/*c

ampspire's memory is pretty good, probably you are referring to the value of "alpha"
from the known relation R(T)=R₀(1+alpha(T-T₀))

and i can guarantee you its not going to reach anything around 100*c even with 1 watt ... thats going to disperse very very quickly when coiled up

You are citing 15AWG x 6 inches, if i'm not mistaken that's equivalent to circa 1.5mm diameter x 150mm lenght,
that's sums up to about 270mm², or 2.7cm². Maybe it will not reach 100°C, but will go
far from ambient temperature. A fan could help. You could find manganin or NiCr wire from some vendors on ebay.

[JackOfVA]

A variant on the original poster's question ...

Suppose for some reason you have to make a shunt from copper. What's the easiest way to compensate for the temperature shift. To make it more in keeping with a challenge, restrict the answer to analog techniques.

I can think of a few - lots of problems with these, but perhaps it will inspire a better solution.

1) Use a second identical piece of wire in thermal contact with the shunt, but electrically insulated. With a constant current source, drive a small current through the non-shunt wire to provide a voltage proportional to the shunt temperature. Scale the voltage to some useful value with a DC op amp. Use the scaled voltage to drive an analog multiplier to ratio up or down the shunt voltage.

2) Drive a low value AC current through the shunt, coupling via capacitance (I assume the shunt is to measure DC current) and derive the resistance from the AC voltage and then apply the approach of (1).

(4) Put the shunt in a temperature controlled oven at say 100 deg C. Measure the oven temperature by the AC resistance of the shunt or some other means. When DC is applied to the shunt, the oven heater will be throttled back so that the shunt is always at a constant temperature and hence constant resistance.

4) Maintain the shunt in a temperature controlled oven at say 100 deg C. Calibrate at 100 deg C. As current is applied, the oven temperature sensor will reduce heating power to maintain the 100 deg value. (The temperature of the shunt could be measured with the AC approach of (2).


3) Scale up the old microwave trick. Put the shunt and an identical wire in a bridge and drive a known current through the second wire until the bridge is balanced.

[Harvs]

You could find manganin or NiCr wire from some vendors on ebay.

Actually that's a lot harder than it sounds.  I've been down this path before trying to make a shunt equivalent to what they use in commercial electronic loads (low tempco, very large for temp stability.) It's easy to find fine resistance wire, but trying to put you're hands on decent gauge low tempco resistance wire is far more difficult in small quantities (easy if you're able to buy 1000m reel.)

[metalphreak]

brass has a temp co half that of copper but twice the resistance. This means you'll have to use thicker gauge brass wire to achieve the same resistance. The benefit of that is greater thermal capacity reducing temperature rises further reducing the change in resistance

[codeboy2k]

toasters.

[richard.cs]

Consider using a longer length of thicker wire to get the same resistance. The heat is then dissipated over a larger surface area and the temperature rise will be less.

I have also seen thin stainless steel threaded bar used as current shunts. Many stainless alloys have a very low temperature coeficient of resistance, and they are nice and adjustable - cut a little longer than you think you want, feed the current to the two ends, and move the voltage take off points along the bar by using a solder tag sandwiched between two nuts.

[mikes]

You can get a decent manganin current shunt, with Kelvin connections, off ebay for <$20.

[eevblogfan]

hey

I uses 1mOhm Shunt I wounded when I was 13 . simply 10AWG can't remember the length , but untill 55A It doesn't heat up so even though it is copper the spec is pretty good , can't remember if 0.1% or 0.5% ( compare to my fluke 287 on the 10A range )

I use my 3478A and measure that low voltage with resolution of 0.0001A on the "30A range"  so that's pretty powerful thing to have

If I was you I probebly would have bought this : http://www.ebay.com/itm/400443484310?var=670076741352&ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

hack it to measure 0-30mV DC and uses some 20CM of 10 AWG wier to create 1mohm resistor , so 0-30.001A device , and if you know how to hack that to 30mV range you probably better do 2 range selectable . 0-30A and 0-300A ( for > 30A )

[nukie]

What's wrong with commercially available current shunts? They are too cheap for you to make them.

[Ghydda]

What's wrong with commercially available current shunts? They are too cheap for you to make them.

I suspect it's more a 'can it be done with usable result'-exercise than a 'looking for the least hazzle'-way.
I often probe into such waters myself. Doing things the hard way merely for the exercise and curiosity.

[nukie]

I like to take the 'homemade' approach too but usually end up many times the cost of an off shelf product, due to obsessive pursuit for accuracy and precision. I guess its a nice project to do if you have time and with the right type of metal it would be very reliable.