Resistance measure - what's the optimal Ref resistor value?

[bt36]

Hi Guys,

I'm trying to measure resistance using a simple voltage divider and ADC.

The resistance to be measured is actually a variable resistor that changes its resistance linearly due to turn of a knob.

Actually I don't care about the absolute value of the resistance , only know when it's at Max (100%) and min (short), and anything in between.

I'm wondering, what's the optimal value of the Rref (in relation to R, the Resistance to be measured), that will yield best resolution?

Since I don't care only about relative values, I'm thinking of making Rref a rheostat.

What is its recommended value?

Many thanks in advance

 

[frankliner]

The best resolution would be given for zero resistance, but this would cause problems when you pot is shorted.

Since you don't care about the absolute value you could just pick R = Rpot@100%. You would loose 1MSB accuracy.

 

[FvM]

Re: Resistance measure - what's the optimal Ref resistor val

The best resolution would be given for zero resistance

No. The best resolution is exactly achieved for Rref = Rpot. Compared to a potentiometer circuit, the (minimum) resolution
is worse by a factor of four. The non-linear characteristic is another disadvantage.

 

[hakeen]

Re: Resistance measure - what's the optimal Ref resistor val

Hi FvM,

Why Rpot is the best?

 

[FvM]

Re: Resistance measure - what's the optimal Ref resistor val

You have a resolution varying with R, it's minimal for maximum R. With R = Rref, du/U = 0.25 dR/R, with higher or lower Rref values,
the relative resolution is < 0.25.

 

[bt36]

Re: Resistance measure - what's the optimal Ref resistor val

Guys thanks very much

I actually did some of the math that lead to this aswer, it's not very difficult but not trivial as it may initially seem.

Thanks again