What sets the resolution of an analog resistive sensor?

Question

When I read about film or silicon strain gauges their data-sheets mentions about their resolution.

But in analog world what sets the resolution? For example if you vary a potentiometer the output voltage changes accordingly so one can say there is relation between the rotation and the output voltage which would be a continous function not quantized.

Can you give an example which would explain the resolution concept for strain-gauge?

Link to a data sheet that talks about strain gauge resolution please. — Andy aka, Feb 27 '19 at 12:32

score 9 · Answer 1 · answered Feb 27 '19 at 14:50

If we define resolution of an analog measurement as the smallest change which can be detected, there are some limiting factors. In this case, we're talking about change in strain, not the resulting change in resistance, though obviously they are related.

For the sensor itself, there are factors such as temperature coefficent, hysteresis, 1/f noise (drift) and Johnson-Nyquist (white) noise. The white noise effect can be mitigated by reducing the bandwidth, but then 1/f noise becomes more important, so even if you have no constraint on the time to take a measurement you can't reduce the noise effect without limit.

To the extent you can know the temperature (and to the extent that it's consistent over the element) and you know the strain history you may be able to compensate partially for some of those factors.

score 2 · Answer 2 · answered Feb 27 '19 at 12:34

2

The resolution is determined by the instrument used to measure the effective resistance.

A given resistive sensor would be characterized by its precision (the repeatibility of measurements under identical conditions) and its accuracy (how well the change in resistance truly reflects the change in strain). Over the range of measurements you could also talk about the linearity of the sensor.

answered Feb 27 '19 at 12:34

Elliot Alderson

31,192
5
29
67

Yes but there is a point that the change in resistance stops reflecting the change in strain. What is the cause of it? Is it the noise or? – pnatk Feb 27 '19 at 12:40
It depends on the exact nature of the sensor. For your potentiometer example it happens when you near the end of the travel of the wiper. This effect would cause a severe loss of **linearity** and **accuracy**. You have not told us enough about your particular sensor to give you a better answer. – Elliot Alderson Feb 27 '19 at 12:59

Marko Buršič · Answer 3 · 2019-02-27T13:00:57.870

2

If you are talking about strain gauges, then the problem has to be expanded also to the application. Strain gauges are used in weighing scales, pressure transducers,...they are glued on piece of metal.

Things may differ if you use some alloy vs other alloy due to different temperature coeficients, moreover the metal body doesn't always return to it's initial position when unloaded, it has some hysteresis.

Many constraints putted together makes a weighing scale to have a finite number of counts/resolution. It's not just a property of strain gauge.

Most legal to trade scales are C3000 standard, meaning 3000 count over full scale.

edited Feb 27 '19 at 13:00

answered Feb 27 '19 at 12:55

Marko Buršič

23,562
2
20
33

What is meant by "count" in this context? Is it like a count of a counter or a weigh unit? – pnatk Feb 27 '19 at 13:27
Measuring intervals, like quants. Or similar with ADC counts - 12 bit = 4096 counts. – Marko Buršič Feb 27 '19 at 13:43

analogsystemsrf · Answer 4 · 2019-02-28T04:07:28.127

Here is a system-analysis for 100 microvolt sensor, into 90dB gain low-noise (3-stage) amplifier, into 10 Hertz RC low pass filter, into 32-bit Analog Digital Converter. The amplifier produces 3 volts PP into the 5vPP ADC.

The right-hand numbers tell the resolution story:

---- total noise 4.94 milliVolts RMS (all uncalibratable errors, causing Code Spread in the ADC output binary code)

---- thermal noise 140 microVolts RMS (dominated by first opamp), with lower left plot showing how that opamp and the Rg (resistor to ground) dominate

---- ADC quantization noise 336 picoVolts RMS

---- Power Supply noise: 4.94 millivolt RMS (60 and 120 Hz; see the table in lower-right corner)

What is the limit to resolution? The power-supply-rejection of the first operational-amplifier (which the tool set to 80dB at low frequencies).

What sets the resolution of an analog resistive sensor?

4 Answers4