Measuring the resolution

A "meaningful distinction" between two measured values ​​is given, for example, when a person sees a clearly perceptible difference in the graphic display of the measured values. This means that the difference is visible over a sufficiently perceptible period of time and over a sufficiently perceptible amplitude.

Resolution = 1x noise amplitude (according to peak value definition)

ME measuring systems define the resolution as the difference between the maximum value and the minimum value of the last 10 seconds or alternatively the last 100 measured values.

This is a strict definition of resolution. Alternatively, you could use the RMS value of the last 3 seconds, or the last 30 measured values.

Transferred to a graphics screen, one could say: For the resolution, 1x line width must be a visible, clearly perceptible difference.

Parts = Measuring range / Resolution

Since the absolute resolution is usually very small compared to the measuring range (e.g. 1/10,000 to 1/100,000 of the measuring range), and since it depends on the measuring range of the sensor, we at ME-Meßsysteme create a relative numerical value for the resolution: We relate the resolution to the measuring range and create the reciprocal of this (for better "readability").

Since the resolution is essentially determined by the measuring amplifier used, the resolution is related to the standard measuring range 2.0 mV/V or 3.5 mV/V of the measuring amplifier.

The resolution is therefore described by a numerical value. This numerical value describes how often the measuring range can be divided into line widths. The higher the number of parts, the better ("higher") the resolution.

The GSVmulti software allows the resolution to be displayed using different definitions: resolution as a numerical value in "parts", resolution as a dimensioned, absolute value "peak values ​​max minus min", resolution as noise amplitude in decibels, etc.

Factors influencing the resolution

The relative resolution (related to the measuring range) is essentially a quality feature of the measuring amplifier: The inherent noise of the first amplification stage is decisive for the resolution. The resolution of the analog/digital converter (the digitization noise) is usually better (finer) than the noise amplitude with 16 or 24 bit technology.

This fact is also responsible for the fact that an exact adjustment of the gain to the measuring range of the analog/digital converter does not lead to a significant improvement in the resolution: Doubling the gain also results in a doubling of the noise amplitude.

Bandwidth

The bandwidth of the measurements essentially determines the noise amplitude. If the bandwidth is limited by filters to e.g. 0...10Hz, the noise amplitude is significantly lower than with a bandwidth of 0...100Hz or even 0...1kHz.

Bei einem über alle Frequenzanzteile gleichverteiltem Rauschen (weißes Rauschen) ergibt sich bei 10-facher Bandbreite eine √10 ≈ 3 -fache Rauschamplitude.

Shielding, grounding, ground concept

In addition to the "inherent noise" of the measuring amplifier, external influences determine the achievable resolution. In particular, the shielding of the sensor cables is an essential prerequisite for high resolution.

Other influences include, for example,

• ground cables that are not laid in a star shape (ground loops),
• line-based interference via USB ports,
• power supplies,
• capacitive interference (particularly with 1/4 bridges) due to a lack of grounding.

Environmental influences

Vibration, drafts or heat input reduce the resolution.

Influence of the bridge supply voltage

A higher supply voltage is often considered as a measure to reduce noise, since an increase in the supply - and thus the bridge output signal - is associated with a decrease in the gain and thus a reduction in the inherent noise.

However, the temporally and spatially uneven self-heating of the strain gauges within a sensor leads to thermal noise, which reduces the resolution. 2.5V to 5V have proven to be the optimal supply voltages. High resistances of the strain gauges lead to higher resistance noise and to more interference due to the higher input impedance of the measuring chain.

Reducing the bridge supply voltage can lead to better stability of the bridge circuit.

Influence of strain gauges

A higher k-factor of the strain gauge does not necessarily lead to a higher resolution. A high resolution requires good "self-compensation" and, above all, a temporal behavior that is as uniform as possible with regard to drift and creep phenomena.

The following figures illustrate the requirements:

• typical maximum expansion of a sensor: 1000 µm/m
• achievable resolution "10000 parts": 0.1 µm/m
• underlying resistance change at 0.1 µm/m: 0.05 10E-6 * 350 Ohm = 17.5E-6 Ohm
• thermal expansion of aluminum: 23 µm/m /K
• required thermal symmetry for 0.1 µm/m: 1/230 K = 0.004 K
• absolute length change at 3mm measuring grid length: 0.1 µm/1000mm * 3mm = 0.0003 µm = 0.3 nm
• distance between two atoms: approx. 0.1 nm

In fact, the maximum strain of most sensors is only 500µm/m, the resolution of the GSV-8 measuring amplifier is about 100,000 parts and the measuring grid length is often only 1.6mm!