Experts on testing: Reducing measurement errors through correct cabling

As products become more complex, the requirements for measurement accuracy and reliability are also rising. Due to space limitations, I will use this article in three parts to introduce the correct wiring and grounding methods for instruments and test devices to reduce measurement errors. Of course, the principles involved in the article can be applied to basic measurement settings, data acquisition systems, and automatic test systems.

Cable specifications

A variety of general and special cables can be used. The following factors will affect the type of cable you choose.

· Signal requirements—such as voltage, frequency, accuracy, and measurement speed.
· Interconnection requirements—such as cable diameter, cable length, and cable routing.
· Maintenance requirements—transition joints, cable terminations, strain, cable length restrictions, cable routing, etc.

Various methods are commonly used internationally to calibrate cables. Be sure to check the type of cable you plan to use and be sure to pay attention to the following indicators.


· Nominal impedance (insulation resistance)—can be found on the cable, from DC to a certain frequency. It changes with the frequency of the input signal. Check the shielding between high and low, and between channels. High-frequency and RF applications have special requirements for cable impedance.

· Insulation voltage—Must be high enough for your application. In particular, consider the maximum insulation of all channels in the system to prevent electrical shock or equipment damage. It is recommended that you use wire with 600 V rated insulation.

· Cable resistance—The resistance of cables varies with wire diameter and length. Use the largest wire diameter possible and the shortest cable length possible to minimize cable resistance. The following table lists the cable resistance of copper wire for several typical wire diameters (the temperature coefficient of copper wire is 0.35%° C). Special instrument sense leads are used in some instruments, such as DMM four-wire resistance measurements and high-performance power supplies for remote sensing, to compensate for voltage losses caused by cable resistance.

· Cable capacitance—Capacitance varies with different insulation types, cable length, and cable shielding. Keep cables as short as possible to minimize cable capacitance. In some cases, low-capacitance cables can be used.

Grounding Techniques

One of the goals of grounding is to avoid ground loops and minimize common-mode noise. Most systems should have at least three separate ground loops:

1. The first is the signal ground. You also need to provide separate signal grounds between high-level signals, low-level signals, and digital signals.
2. The second is the ground for high-noise hardware, such as relays, motors, and high-power devices.
3. The third ground is for chassis, racks, and cabinets. The AC power ground should generally be connected to this third ground.

In general, for frequencies below 1 MHz or low-level signals, use a single-point ground. Parallel grounding is better, but it is more expensive and more difficult to wire. Parallel grounding followed by a single-point grounding is necessary. The most important point, especially for small signals or the most accurate measurement requirements, should be the nearest ground. For frequencies above 10 MHz, use a separate grounding system. For signals between 1 MHz and 10 MHz, you can use a single-point grounding system if the longest ground wire loop does not exceed 1/20 of the wavelength. In all cases, loop resistance and inductance should be minimized.

Shielding Techniques

Noise shielding must take into account capacitive and inductive coupling. There is a large capacitive coupling between the conductor and the surrounding ground shield. In a switched network, this shielding is related to the shape of the coaxial cable and connector. For signals above 100 MHz, it is recommended to use double-shielded coaxial cable to reduce the shielding effect. Reducing the loop area is the most effective way to deal with electromagnetic coupling. For signals below a few hundred kilohertz, twisted pair can resist electromagnetic coupling. Use shielded twisted pair to avoid electromagnetic and capacitive pickup. Maximum protection can be provided for signals below 1MHz, but make sure that the shielding does not conduct any signal to the user.

 Separating Small and Large Signals

If the ratio of signal strength between them exceeds 20:1, they should be physically separated as much as possible. The entire signal path should be reviewed, including the wiring and adjacent connections. All unused wires should be grounded or tied low and placed in the sense line channel. In a data acquisition system or ATE system, be sure not to affect the connection and function of adjacent channels when using screws to secure wires to the connector.

Wireless radiation interferes with

most voltage measuring instruments and can generate false readings if there are high-intensity, high-frequency signals around. Possible sources of high-frequency signals include nearby radio and television transmitters, older computer CRT monitors, and cell phones. High-frequency energy can couple into the wiring of a DMM system. To reduce interference, try to minimize exposure of the cabling connections to high-frequency RF sources. If the test you need to complete is extremely sensitive to RF radiation from the instrument, a choke coil, used in the system cabling, as shown below, will be required to attenuate the instrument's radiation. Note that you are most likely to see this coil on your computer monitor's video input cable. It looks like a cylinder, and has a small choke coil in the center.

Thermal EMF Errors

Thermal EMF errors are the most common source of small DC voltage measurements. Thermoelectric voltages are generated when you use dissimilar metal connections at different temperatures. Each metal-to-metal connection forms a thermocouple, which generates a voltage proportional to the temperature difference at the connection. You should take the necessary precautions to minimize the thermocouple voltage and temperature changes in low voltage measurements. The best connection is a copper-to-copper crimp connection. The following table shows the thermoelectric potentials generated by common dissimilar metal connections.

Noise Caused by Magnetic Fields

If you are measuring near magnetic fields, you should take precautions to avoid induced voltages in the measurement connections. Voltages can be induced if the input connection wiring is shaken or the magnetic field changes in the field. Unshielded, bare input wires moving in the earth's magnetic field can generate several millivolts. The varying magnetic fields around AC power lines can also induce voltages up to several hundred millivolts. You should be especially careful when working near conductors carrying large currents. If possible, you should keep wiring away from magnetic fields. Magnetic fields are prevalent around electric motors, generators, televisions, and computer CRT monitors. In addition, if there are magnetic fields near the work area, make sure your input wiring is straightened and securely fastened when working. Use twisted-pair wiring to connect to the instrument to reduce the loop area for noise pickup, or keep the wires as close together as possible. When measuring milliohm meters and millivolt meters during calibration, measurement fluctuations caused by magnetic fields can cause significant problems. To reduce these fluctuations, we can build a metal shielding box to isolate the instrument from the surrounding magnetic field. The case can have a small opening, just large enough to read the measurement and change settings.

Low-Level AC Measurement Errors

When measuring AC voltages less than 100 mV, be aware that these measurements are particularly susceptible to errors introduced by extraneous noise sources. The exposed test leads can act as antennas, and the internal DMM will measure the received signal. The entire measurement path, including the power leads, can act as a loop antenna. The current in the loop will produce voltage errors after passing through a series of series impedances including the instrument input. For this reason, you should use shielded wire to input low-level AC voltages into the instrument. You should also connect the shield to the low end. If this cannot be avoided, be sure to minimize the area of ​​the ground loop. A high-impedance source tends to pick up more noise than a low-impedance source. You can reduce the high-frequency impedance of the source by connecting a capacitor in parallel with the instrument's input terminals. You may need to experiment to determine the correct capacitor value for your application. Most external noise is uncorrelated with the input signal. You can use this formula to determine the error, as shown below.

Noise associated with the device under test is extremely rare, but particularly harmful. This noise will always be added directly to the input signal. When measuring low-level signals, it is easy to cause errors if there are other signals with the same frequency, such as the same frequency as the local AC power supply. If you switch large and small signals on the same switch card or module, you must be very careful. Because in this case, it is possible that the charging power of the large signal will be released on the channel of the small signal. In this case, it is recommended that you can use two different modules, or separate the small signal channel from the large signal channel, and add an unused channel between them, connected to the ground.