English is currently available.
Control & Instrumentation Engineering DISCUSSION
I have to calibrate a 0 to 10 bar, 4-20 mA pressure transmitter against a reference gauge and a milliamp meter and fill in an as-found / as-left sheet. The stated tolerance is ±0.5 percent of span.
Which test points should I use, how do I convert a reading such as 12.10 mA at 5 bar into a percent error, and how can I tell from the results whether the problem is zero, span or something else?
Use five points across the range, 0, 25, 50, 75 and 100 percent, going up and then coming back down, so nine or ten readings. For 0 to 10 bar the ideal outputs are 4, 8, 12, 16 and 20 mA at 0, 2.5, 5, 7.5 and 10 bar.
Error in percent of span = (measured - ideal) / 16 mA × 100. Your 12.10 mA at 5 bar is 0.10 / 16 = 0.625 percent of span, which is outside ±0.5 percent. In current terms the tolerance is ±0.08 mA at every point.
The pattern of the errors tells you what is wrong. The same offset at every point, for example +0.10 mA from bottom to top, is a zero shift. An error that is nil at 0 percent and grows steadily towards 100 percent is a span error. Correct at both ends but off in the middle is non-linearity, which a simple zero and span adjustment cannot remove.
If the upscale and downscale readings at the same pressure differ, that is hysteresis, often mechanical in origin. Approach each point from one direction without overshooting, or the test itself hides it.
Record the as-found readings before touching anything. They show how far the instrument drifted since the last calibration, which is what decides whether the calibration interval is right and whether past measurements were affected. Only then adjust, and repeat the full run for the as-left record.
On analog transmitters zero and span interact, so set zero, set span, and go round again until both hold. Adjust zero at the 0 percent point and span at the 100 percent point.
With a smart transmitter, separate three different operations. A sensor trim corrects how the device reads the applied pressure against your reference. An output trim corrects the 4 mA and 20 mA currents against your milliamp meter. Re-ranging only changes which pressures correspond to 4 and 20 mA and is not a calibration at all.
If the digital pressure value is right but the current is off, do an output trim; if the digital value itself is off, do a sensor trim.
The reference has to be clearly better than the instrument under test. A common rule is a ratio of 4 to 1: to verify ±0.5 percent of span, the combined uncertainty of the pressure reference and the current meter should be about ±0.125 percent of span or better, and both need valid calibration certificates of their own.
Also mount the transmitter in the same orientation as in service where possible. On low ranges a liquid-filled impulse line or a tilted sensor shifts the zero.
