0.01 Ohm resistor

--
Buy one and save yourself a lot of grief.

How about using a length of copper wire? #20 can handle 11 amps, and one foot of it is 0.011 ohms according to one wire table. If you run it back and forth, not too tightly bundled it will dissipate the 1 watt with little temperature rise.

Geez, does it really need to be that accurate? If this is for your current limiter, I wonder whether it really matters if you limit at 9A or at 11A. After all, the downstream stuff won't be precise enough to know the difference.

But anyway, you can buy .01 ohm resistors. You can also buy 0.1 ohm resistors and parallel 10 of them, which will help spread the heat load. I've seen that done in switching power supplies.

A bit of climb-with-heating could be a good thing, for a current sense resistor; it would tend to decrease the current limit when overheated.

You can buy two and for terminal current sense resistors. Search digikey for those terms. Based on the fact that you say you can't tollerate a variance of much more than a couple of percent, this is probably the way to go. It sounds like you are interested in using this to make current measurements, which is what I am doing in my project and I used a four terminal resistor for this purpose. I am able to read the current to an accuracy of better than 1% with it and have pushed over 10 amps through it. I believe I may have used a .005 ohm because it was readilly available and the .01 ohm was out of stock.

How much of a problem is heat dissipation at 100 mWatts !!! IxR= 10 Amp x 0.010 Ohms = 0.1 Watts !!! The real trick is how do you measure a 0.010 Ohm resistor, now that is the real Problem !

Yukio YANO

You don't want to use any pure metal for shunt resistors--their resistances rise very rapidly with temperature (something like 6000 ppm/K for copper, IIRC). Nichrome works, or constantan thermocouple wire (which is easily solderable).

Cheers,

Phil Hobbs

No, it's I^2*R which is 1 watt.

Ian

I misspoke. 4000.

Cheers,

Phil Hobbs

1 watt, as Ian pointed out. Measuring .01 or even .001 ohms isn't that hard. All you need is an accurate current source and a volt meter. I made the current source with a LM317 and a length of resistance wire. Calibrating such a current source to 1% accuracy or better is fairly easy to do just by locating the right point to tap the nichrome wire. I wound the length of nichrome around a straw and secured it with heat shrink tubing.

Anyway I thought I could fabricate a low value sense resistor, but it looks like buying one is the better solution. Thanks all for the advice.

Go to the hobby shop (model trains, RC airplanes, etc) and get some brass shim stock. Cut a strip of the appropriate size - maybe a little oversized, so that you can trim it with a Dremel or so.

Figuring out the dimensions of the strip is, of course, left as an exercise for the reader. ;-)

Or, bite the bullet and get a proper current shunt. ;-)

Good Luck! Rich

What problem? All you need is a known current source and a millivoltmeter.

Done! ;-)

Cheers! Rich

--
"The appropriate size"???

Ever considered using a hall sensor?

"Ancient_Hacker"

** It is often important that current shunt have low *self inductance* so its value remains the same to some high frequency.

A 1 foot length of #20 looped wire has enough inductance to make it no longer an accurate shunt above about 1 kHz.

........ Phil

Based on a diameter of .032 inches and the formula on this page:

I calculate an inductance per foot of about 360nH.

That produces a 45 degree frequency of about 4.5kHz. I wonder how that would change if you zig zagged the wire into a 1 inch square.

Much lower inductance, I am quite sure.

** I used the rule of 25 nH per inch - close enough.

• What a strange criterion.

By the time the phase angle is 45 degrees, the impedance value has risen by a whopping 41 %.

An increase of about 1% happens at a mere 450 Hz.

** Much easier to use a short length of a high resistance, low tempco material.

As used to make WW resistors.

Get L down to a maybe 20nH then.

Moves the 1% error frequency up to 9kHz.

........ Phil

Hmm 1 Watt...

Nichrome wire keeps a failry constant resistance, it's not the easiest stuff to solder though and you'll need to parallel several pieces of nichrome wire to get the current capacity you want unless you can get some thicker than usual nichrome somewhere.

Bye. Jasen

According to Grover and Rosa of the NBS, such a non-inductive (so-called) zig-zag arrangement would give about 103 nH. If instead of making it into a 1" square, you made it 1" by .384", and kept the strands right next to each other, the inductance would be 39 nH.

Just folding the 1 foot of wire into a hairpin, with the two strands as close together as possible would give about 57 nH.

The real problem with measuring low value resistances is contact resistance - soldered contacts do have very low and stable resistances, but a lump of solder on the resistor side of the joint can affect the resistance of your resistor!

The solution is to take four terminal (Kelvin) measurements.

The voltages that you measure tend to be fairly low, so an AC excited bridge can get you better accuracy than DC - reversing DC can be almost as good. If you are really pusshing for high accuracy (as national standards laboratories do) you use an AC excited transformer-based (Blumlein) bridge, where the transformers are precision ratio transformers, with ratio's stable to about one part in ten million. See

Lots of good advice. My copy has seen very little use, but when I've needed it, it has been invaluable.

--
Bill Sloman, Nijmegen