"Dynamic Error Band Basics"

Let us cover the basics and understand just what the diagnostic traces should look like on a good performing valve.

This feature will be dedicated to the "Dynamic Error Band" (DEB) curve (ValveLink) and/or the "Total Valve" curve (FlowScanner).  Though labeled differently, these are identical curves that present the valve’s performance by plotting control signal (input) verses travel (output).  For this writing, we will refer to them both as a DEB plot.

Although encompassing the entire valve assembly (i.e. instruments/valve/actuator), the DEB plot is more reflective of the instrument’s performance then that of the valve.  We will concentrate our efforts on valves equipped with a positioner or digital valve controller (DVC).  Also for this discussion, we will assume that the instrument is meant to perform linearly either via a linear cam within the traditional instrument or linear characteristic configuration within the DVC.

The first thing one looks for is the linear relationship between the input and output.  The change in output travel should track the change in signal in a linear relationship as shown here.

Zero to 100% change in input signal span should result in zero to 100% valve travel movement.

The bulges one sees at both ends are referred to as boots.

These boots are the result of the time delay that occurs from the time the valve is commanded to move until the time the movement is detected and satisfies the signal.  Keep in mind that the test is run by changing the input signal from zero-to-100% and then back to zero at some controlled ramp rate.  At the initial start of the test the actuator is void of pressure.  As the valve is commanded to move, the positioner/DVC is supplying as much air, within its capability, to load the actuator and start the valve moving.  As the actuator is being loaded, the input signal is continuing to change (ramp).  As the actuator volume is filled and the pressure becomes sufficient enough to overcome the spring and frictional forces, the valve moves.  It now moves very rapidly to the position in travel that correlates with the input signal at that particular time in test.  At this point the actuator void had been filled and the travel now tracks the change in signal.  The same is true as the signal is reversed.  However, in this case the actuator is saturated with pressure and thus the exhaust of this volume causes the delay.   These boots will always be present but should appear minimal.  The auto analysis routine of the FlowScanner is performed between 5% and 95% of the total travel while ValveLink’s analysis occurs between 10% and 90%.  This is done to eliminate the effects of the boots on the DEB and Linearity analysis values.

Now look at the spread in data points between those collected during the increase in input signal and those collected during the decrease in signal.  The lines should appear evenly spread throughout the full test with the exception of the boot areas.  This spreading is what is defined as Dynamic Error Band (DEB).  This DEB is the cumulative result of Hysteresis, Dead Band, and time lag.  In an ideal world, we would like to see the increase/decrease signal lines to fall right on top of each other.

This would say that at any precise input signal (setpoint), the valve would be at the same travel position regardless from which direction (increase/decrease) the setpoint is approached.  But, we operate in the real world with these mechanical devices known as valves.

All mechanical devices have, to some degree, inherent Hysteresis and Dead Band.  This roughly translates to slop (loose connections) and friction within the device.  Hysteresis and Dead Band is a static measurement determined by changing the input and then pausing to allow the output to come to some equilibrium or steady state rest.  The difference in output due to Hysteresis and Dead band on a well-designed instrument is typically less then 0.5%.   This percentage is the amount of spread (output difference at any given input) divided by the total output.  Because of the way we conduct the test (continuous ramping of the input signal instead of stopping to allow for steady state equilibrium) we introduce a sluing error due to the test speed.  This means that the output is always lagging the input in varying degrees depending on the test’s ramp rate and the response of the instrument being tested.  Knowing this, we can put it to our advantage.  By designating a specific ramp rate based on actuator size or volume, we can control the sluing error.  This allows us to compare apples to apples.  You can take any given positioner and mount it on any size actuator and come up with the same DEB results by just adjusting the speed of the test.  Or, you can take competing or different instruments mounted on the same size actuator and determine response variance between the lot.  DEB is a very useful measurement for determining the integrity of the positioner and/or I/P.

<< Back to Dabbling in Diagnostics Index