How to measure a resistors value as accurate as possible?

I got a few cement 5W and 10W resistors in the range 0.1 to 20 ohm. I want to measure their resistance as accurate as possible but my multimeter has bad resolution in that range.

Any idea how I can be done?

Google "Arduino measure low resistance".

groundFungus:
Google "Arduino measure low resistance".

Of cause... found this.

I'll try it.

The standard approach requires two meters, one measuring the current through the resistor,
the other the voltage across it.

Its vital that 4 wires are used, two from each end of the resistor, so that any voltage drop across
the wires carrying the current aren't included in the voltage measurement. Thus one set of wires
goes only to the voltmeter direct from the resistor terminals.

Since the volt meter draws negligible current there is no error from the test lead resistances
(being the big problem using a standard multimeter in ohms range).

You need some way of supplying a steady current, around 100mA - 1A would be good for a 0.1 ohm resistor,
10 -- 100mA for a 20 ohm resistor (you want the power dissipation in the resistor not to be so high it
gets hot, but more current gives more accurate readings).

Good google terms are "kelvin connection" and "four-terminal measurement"

The problem is I only got one multimeter. The four-terminal measurement require both a A and V meter, which I do not have.

Then you are not in a position to make accurate measurements. Simple as that!

If you are serious about this, you will need to build a precision constant-current source.

I got a few cement 5W and 10W resistors in the range 0.1 to 20 ohm

You could buy a few precision resistors to calibrate/correct your readings. Is 1% close enough? But 0.1 Ohm measurement is still likely to be unreliable because of the resistance in the wires & connections.

If you are serious about this, you will need to build a precision constant-current source.

And, you still need to make accurate measurements...

The problem is I only got one multimeter. The four-terminal measurement require both a A and V meter, which I do not have.

At work, I have an old 4-wire HP bench-multimeter. It probably cost a few thousand dollars but Google found [u]this[/u] for only $100 !

Still time to write to Santa and ask for a proper meter. You could ask for two.

It’s always nice to have a few precision resistors of different values for such situations.

thehardwareman:
Of cause... found this.

I'll try it.

Just more proof that any idiot can post absolute crap on the Internet. Yes, please show everyone how to measure resistors of less than one ohm all built on a solderless breadboard. Yep, that will work. Kind of, sort of. But the results are meaningless.

Knowing the precise value of said resistors is almost pointless, they’re typically good for 5 or 10% accuracy which good enough for their typical applications.

Just curious, what is your proposed usage of the power resistors that the absolute resistance matters so much?

I got a few cement 5W and 10W resistors

I bet they are actually ceramic.

$100 for a used 4-wire meter is a bit overkill just to measure one or two resistors. Then it's better to buy/borrow a second multimeter (I can probably borrow a good second multimeter at work).

A constant current source has been on my mind.

The resistors are 10% but if I can determine the resistance down to 1% I would me happy.

Thanks for the feedback.

WattsThat:
Just more proof that any idiot can post absolute crap on the Internet. Yes, please show everyone how to measure resistors of less than one ohm all built on a solderless breadboard. Yep, that will work. Kind of, sort of. But the results are meaningless.

What? Why do you think the results are meaningless? The method should work well. Since he is measuring on the leads of the reference resistor and DUT he has Kelvin connection and the results will be quite precise. The only problem I see is that connecting the test leads may move the measured part and change contact resistance (and so current). But two DMMs or repeated measuring should fix this so even solderless breadboard is no real problem.
Or you simply don’t like the idea everyone can easily measure <1 Ohm resistors? Are you selling $1000+ precision meters or what?

Lacking the required equipment to really get good and accurate measurements I would just seek some help from a source which has what is needed. They can note the temperature and give you some numbers on a tag. Since you only have a few resistors it obviously is not worth it to invest in expensive equipment just to do a one time measurement.

All you need to do is actually state an accuracy (uncertainty) you want the numbers taken to.

Ron

Smajdalf:
What? Why do you think the results are meaningless? The method should work well. Since he is measuring on the leads of the reference resistor and DUT he has Kelvin connection and the results will be quite precise. The only problem I see is that connecting the test leads may move the measured part and change contact resistance (and so current). But two DMMs or repeated measuring should fix this so even solderless breadboard is no real problem.
Or you simply don’t like the idea everyone can easily measure <1 Ohm resistors? Are you selling $1000+ precision meters or what?

Contact resistance is a problem even if you use a full 4-wire Kelvin connection in the breadboard, as
there is a contact carrying current between the volt sense wire and the resistor terminal. You need to
use croc-clips or terminal block of some other means of making a solid connection (breadboard springs
tend to be rather weak and may not scrap away surface oxide like a good crop-clip or screwterm.

At first I considered (and rejected) the idea of measuring an RC time constant.
So we are back to potentiometric methods.

You need to measure the voltage across the resistor under test, (RUT) and know the current through it.
We will make kelvin connections to the RUT, (with croc clips) and use screw terminals to ensure good connections for the current flow.

pot2.png

The diagram shows how this is done. You need to measure V1, V2 and V3.

V2 should be very near zero, and V3 about 0.9 volts.

So you can use the 1.1V reference to make these measurements.

V1 - V2 gives you the voltage across the RUT.

V3 gives you a measurement of the supply voltage and allows you to have a correction if its only say 4.7V.

Knowing V1, V2, V3 you can calculate the voltage across the 10 ohm resistor, and hence the current through the RUT and hence the resitance.

You COULD use high precision resistors throughout - but all you really need is 1 resistor of known value to do a calibration on each range.

pot2.png

MarkT:
Contact resistance is a problem even if you use a full 4-wire Kelvin connection in the breadboard, as
there is a contact carrying current between the volt sense wire and the resistor terminal.

Did you look at the mentioned link? He is connecting the DMM probes directly to the resistor leads. There is no current carrying contact in the voltage measuring path. I think it is directly by definition - if there is such contact it is not a Kelvin connection.

Cf. Wikipedia Wheatstone bridge.

Smajdalf:
Did you look at the mentioned link? He is connecting the DMM probes directly to the resistor leads. There is no current carrying contact in the voltage measuring path. I think it is directly by definition - if there is such contact it is not a Kelvin connection.

Of course!

+1

That Wheatstone bridge was a "fancy" solution but I suspect that the variable R2 must be difficult to obtain. A simple trimming resistor would probably not do.

I assume that of the different methods except for a simple voltage divider, give better result than a simple voltage divider. But I do not need lower than 1% accuracy (which is 10 times the accuracy of the resistor).

See my post above - its a "modified" wheatstone bridge, uses components you will likely already have, and you only need 1 "known value" resistor to calibrate it. NAything you dont understyand just ask?