Decades ago, the Proving Ring was conceived as a device to be used for the calibration of force measuring dial gauges. It consisted of a steel ring with a micrometer mounted so as to measure the vertical deflection when loads were applied through the threaded blocks at the top and bottom.
Figure 4. Proving Ring
For many years proving rings were considered the standard of excellence for force calibration. However, they suffer from the following adverse characteristics:
When a force is applied to a solid material within its elastic limit, the resulting deflection will increase very subtly with time if the force is held constantly. This is true whether the force is in tension or compression. This phenomenon is called creep, and by definition, is not permanent but is recoverable. The signal from a load cell exhibits this creep, and therefore it should be understood in all load cell applications. Both loaded creep and creep recovery are exponential with time, as illustrated in Figure 5.
Figure 5. Creep versus time
Referring to Figure 5, a force is applied in time 0 to (A), and the deflection goes from point 0 to point (K). Then in a stable loading condition, the deflection increases up to point (M) during time (A) to (B). This is positive creep. It is also possible that it could have been negative creep, in which case the curve from (K) to (M) would have gone negative rather than positive. When the force is released at time (B), the deflection quickly goes to point (J), then creep recovery occurs and the deflection goes back toward 0 with about the same curve shape as the loaded creep inverted.
Creep continues as long as a force is applied, but the rate of creep decreases significantly with time. It is typically measured and specified for a 20 minute interval. To illustrate the concept, the creep in the first 20 minutes is about equal to the creep in the succeeding 24 hours.
When forces are applied to the proving ring, it departs from its circular shape and becomes slightly egg-shaped. The determination of the deflection of a proving ring depends on the subtraction of two large numbers, namely, the inside diameter of the proving ring and the length of the micrometer measurement assembly. Since the difference is so small, any slight error in measuring either dimension leads to a large percentage error in the number of interest, the deflection.
Any mechanical deflection measurement system introduces errors that are difficult to control and/or overcome. The most obvious problem is resolution, which is limited by the fineness of the micrometer threads and the spacing of the indicator marks. Nonrepeatability of duplicate measurements taken in the same direction depends mainly on how much force is applied to the micrometer’s screw threads, while hysteresis of measurements taken at the same point from opposite directions is dependent on the preload, friction, and thread looseness.
Variation in the temperature of either the steel ring or the micrometer assembly will cause expansion or contraction, which will result in a change in the deflection reading. A first-order correction would be to make all of the parts out of the same material so that their relative temperature effects are equal, effectively causing them to cancel one another. Unfortunately, this presumes that all of the parts track one another in temperature, which is not true in practice. A light shining on one side of the ring or a warm breeze from a furnace vent will cause differential warming, and a proving ring is very susceptible to temperature gradients in the proving ring mechanism. Also, the spring constant changes with temperature, thus changing the calibration
Response to Extraneous Forces
The construction of a proving ring does not lend itself to the cancellation of extraneous forces, such as side loads, torque loads, and moment loads. Any load other than a pure force through the sensitive axis of the ring can result in an extraneous output.
The proving ring requires trained personnel for proper operation because of the possibility of errors introduced by creep, as well as the potential for errors due to temperature and extraneous loads.