multi turn pot 1

[gurujeff]

I'm having trouble finding the definition of the linearity tolerance and a resistance tolerance for a multi turn pot. Can anyone help me out.

We will be using this as a position feedback device and I would like to know if for example we had a 10k multi turn pot that has a linearity of .5% and a resistance tolerance of 5%. If we calibrate the device such that we have a given "X" distance at lets say 5000 Ohms what would be the tolerance on the lower end. Is it 5% of the 10K or +/- 500 ohms for any given resistance or is it 5% of the resistance value which would in turn make a difference where I calibrated the pot and the distance relationship.

Any help explaining this would be appreciated. Thanks

 

[IRstuff]

It's neither. Both the linearity and tolerance are calibrated out, assuming you're calibrating each and every device. The only remaining uncertainties are your rotational position repeatability, the potentiometer slider repeatability, and the resistance measurement uncertainty, none of which you've specified. Oh, and there's a temperature coefficient as well.



TTFN

FAQ731-376

 

[gurujeff]

I'm not sure if you understand. This isn't a sliding pot and I don't think the calibration would eliminate the tolerance on both ends of the pot. It certainly wouldn't eliminate the linearity. That is inherent with the pot design. The calibration method is not an electronic mapping. I set the pot at 1K within a few ohms at "X" position. Lets say I turn the pot exactly 4 rotations and it should be at 5K. I've been told that best accuracy that I can get given any number of turns is the resistance tolerance which is 5% of the total value of the pot in this case that is 10K total or a +/- 500ohm accuracy. I know there are 3% and 1% pots out there but I want to make sure I understand the tolerance error properly.

 

[IRstuff]

a multi-turn pot is a slider, only helical.

In your specific example, the answer is no. The value at any point is (FR±NL±PE)*(10K+5%-6%)

Where
FR = fractional value = turns/10
NL = nonlinearity
PE = rotational position error
-6% accounts for the end resistance error
so at turn 5, the value should be 5K +5%/-6%, nominally.

In order to correctly calibrate, you need to get a minimum of 2 points, ideally bracketing your operational range. If you do that, then you'll only have the nonlinearity error coming from the pot.

TTFN

FAQ731-376

 

[gurujeff]

ok well straight or helix or multi-turn etc that doesn't factor into the spec.

1) How can the non-linearity always be positive? Shouldn't this spec be based on taking any 2 points in the range of the pot. If you plotted resistance vs. rotary position it should come out a linear plot and the linearity error should be the deviation from a straight line of those 2 points. This value could be positive or negative. I know this the value is usually pretty small something like .25% seems standard. I believe I understand this part pretty well it is the resistive tolerance I'm questioning.
2) Let’s assume there is no rotational positional error because that is not what I’m concerned about. I’m concerned about the error from the pot not the error of the device I have turning the pot. So Lets say the value of the pot is R. I will set R to 1K through a manual calibration (i.e. multimeter and adjustment). I will then turn the pot EXACTLY 4 turns. Let’s say the new value of R should nominally be 5K. The error from the pot will include the non-linearity and some sort of resistance tolerance error. The non-linearity is .25% of 10K or 25ohms. The resistance tolerance if I’ve been told correctly is 5% of 10K or 500 ohms (I’m not sure if it is based on the highest value of the pot or lowest or what but I’ve been told it is the highest). Worst case would be if these errors added together. My resistance value could be 525ohms greater than the nominal or 5525 ohms. Is this correct?

 

[Skogsgurra]

Hello Guru!

One thing that you haven't even touched upon is how you are using the potentiometer.

The normal way of using it is as a voltage divider (three wire circuit). Then, your resistance tolerance is not important at all. Just the linearity spec. If you, as I am starting to think, actually do use it as a series resistor in a circuit (two wire circuit), then rethink. At least if you are concerned about the error budget.

BTW non-linearity is given as a +/- value, i.e. deviation from a (best fit) straight line. The deviation is hence given as an ABS value, which - by definition - always is positive.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...