How to measure cable loss

Overview of cable loss measurement

All coaxial cables attenuate radio frequency signals that pass through them, and this attenuation is commonly referred to as either “cable loss” or “insertion loss.”

Cable loss is a function of the cable length and the frequency of the signal passing through the cable. Cable loss generally increases linearly with increasing length - doubling the length doubles the loss. Cable loss also increases with increasing frequency, but this is not a purely linear relationship.

Loss is an important specification provided by cable manufacturers, and it is often expressed in decibels (dB) per meter or foot. The frequency-dependent nature of this loss is usually represented through tables or graphs, and understanding the amount of cable loss is important for various RF applications.

Despite manufacturer-provided specifications, however, you may still need to measure actual cable loss, especially when the cable type is unknown or when factors such as connectorization or wear affect performance. The most common tool for measuring cable loss is a vector network analyzer (VNA).

Cable loss measurement with VNA

There are two ways to measure cable loss with a VNA:

• One-port cable measurement (S11 or reflection measurement): You connect one end the cable to the VNA, and the other end is either left open or shorted. The VNA injects a signal, and the reflected power is compared to the transmitted power to calculate the cable loss.
• Two-port measurement (S21 or transmission measurement): You connect both ends of the cable to the VNA. One port sends a swept signal through the cable, while the other measures the signal’s magnitude at the far end. This method is preferred for cables with high loss or when both ends are accessible.

For one-port cable loss measurements, a source or tracking generator is used to inject a signal into a cable. The frequency of this signal is swept over a user-defined range. The far end of the cable is either left open or is terminated with a short. In both cases, a signal that reaches the end of the cable will be reflected back to the source port.

At the source port, the amount of reflected power is compared to the known transmitted power. The cable loss in dB is the total or “round trip” attenuation divided by two. As mentioned above, the total loss of the cable is a function of both the signal frequency and cable length.

Before beginning the measurement, you should configure the VNA. This involves three main groups of settings:

• Sweep frequency range: This is the frequency range over which the tracking generator or stimulus signal is swept. It should cover the frequencies for which the cable will be used.
• Number of measurement points over the span: Increasing this number will provide greater detail, but more frequency points will also increase the amount of time you need for a single sweep.
• Averaging multiple sweeps: This can be used to reduce noise and obtain a more accurate result and is especially useful for cables with very high loss. However, increasing the number of sweeps will also increase the overall measurement time.

After configuration, you can connect the cable under test to the VNA in two different ways:

• Directly to the analyzer port
• Using using a short, high-quality phase-stable DUT cable

Why would you want to use a DUT cable? Well, a DUT cable is useful when the cable under test has a connector that is difficult to access, such as when the cable terminates in an enclosure or is attached to a tower or mast. Another reason is that a DUT cable can reduce wear and mechanical stress on the analyzer port. You can remove the effect of the DUT cable on the measurement results during calibration.

Calibration is necessary for accurate cable loss measurements. To do this, you sequentially attach an open standard, a short standard and a match (or load) standard to the cable under test. These standards can be in the form of discrete standards or may be combined into a “calibration tee.” In addition to these manually attached standards, electronic calibration units (autocal) can also be used; these units switch their internal standards automatically and are controlled by the attached VNA.

Calibration is usually a “follow the prompts” process in which the VNA will indicate which standards are to be connected in which order and at which times. It is a quick process (usually only a few minutes), and automatic calibration units tend to be faster than using manual standards.

How you connect the calibration standard to the VNA depends on how you will connect the cable under test to the VNA. That is, if you connect the cable under test to the VNA directly, the calibration standards should also be directly connected to the port. If you use a DUT cable, the calibration standards should also be connected to the end of the DUT cable.

Let’s take a look at an example one-port cable loss measurement result. In the image below, you can see cable loss as a function of frequency between 1 GHz and 5 GHz with the y-axis showing loss or attenuation in dB. This trace is typical in two ways:

• The attentuation increases with increasing frequency.
• The trace has a wavy pattern or “ripples” caused by reflections.

You can quantify the cable loss by averaging the minimum and maximum values. In this example, the minimum value is -1.2 dB and the maximum value is -3.5 dB, so the loss would be -2.35 dB.

Now, let’s talk about two-port measurements. Two-port measurements are preferred over one-port measurements in two cases:

• There is easy access to both ends of the cable.
• The cable has a very high loss (above 20 dB).

For most two-port cable measurements, you can just directly connect the cable under test to both analyzer ports. If, however, DUT cables are used to connect the cable under test to the analyzer, then a normalization should be performed to remove the DUT cables’ influence from the measurement.

Cable loss in two-port measurements is still a function of frequency, but the trace has fewer ripples than in a one-port measurements because both ends of the cable are terminated in their characteristic impedance. Although there are many cases when attaching both ends of a cable to a VNA is impratical or infeasible, two-port cable loss measurements are generally preferred over one-port cable loss measurements.