Do you cut your coax?

A benchmark standard approach to installing coaxial cable for mobiles will improve RF power measurements and improve reception quality in fleet vehicles.

Do you cut your coax cable for your mobile two-way radio installations? We at Washington State Patrol (WSP) do. The reason? Repeatable test results.

The Washington State Patrol installs equipment in as many as 500 new vehicles and retrofits more than 100 vehicles each year. During the check-out phase, RF power measurements must fall within WSP standards (e.g., VHF: 100W forward power, 5W reflected power. UHF: 30W forward power and 1W reflected power). If these measurements do not meet specifications, the vehicle is set aside for further investigation to determine the fault. This is the point where frustration begins. The coaxial cables used to connect the radios to the antennas may be random in length. Coax that is terminated resistively with its own characteristic impedance will read “zero” reflected power. If SWR exists, the wattmeter readings will be different for each length of cable.

An antenna cut to exactly one quarterwavelength (1/4 lambda) will have a nominal impedance of 37 + j221/2ohm, (assuming sufficient ground plane). Reducing the length slightly will eliminate the reactive component, resulting in 37ohm resistance at the base, which yields a VSWR of 1.35:1. Therefore, if the wattmeter indicates 100W forward power and 2.2W or less reflected power, everything is normal.

As shown in Figure 1 above, a 1/4 lambda cable can show an impedance from ground to infinity, depending on its length on either side of the 1/4 lambda. Note also that the impedance repeats itself every 1/2lambda, except that the phase angle is different. This feature can confuse the situation.

Consider a cable, less than 1/4 lambda long, that is inductive-reactive. When attached to an exact 1/2 lambda cable, the end result is capacitive-reactive because of the 1808 phase shift of the 1/2 lambda cable. It takes a full wavelength (1 lambda) cable to repeat the exact phase. If the cable length is somewhere between 1/2 lambda and 1 lambda, and not terminated in its characteristic impedance, its impedance can be changed as seen from the sending end because of the 1/4 lambda influence. A 1/4 lambda cable can also be attached as a stub to a cable of the same impedance, adding reactive components, or an open or a short, as determined by the length.

This method was popular many years ago in the CB business when someone used the wattmeter in the reflected mode and began cutting the coax cable from the antenna until the meter read zero. This fooled the transmitter, but the cable mismatch was still there.

“Ya got yer coax cut” was the catch phrase of the time.

WSP cuts cables to establish a common point whereby all measurements would use the same reference. Case in point: WSP had both 150MHz and 450MHz radios connected to a dual-band antenna via a duplexer. During the checkout phase, not all cars met the reflected power standard. Changing duplexers sometimes solved the problem, and at other times changing antennas solved the problem. Soon there were boxes of duplexers and antennas that would be labeled “good?”, “bad?” and so on. Investigation found the problem to be intricate.

Each individual component-antenna, duplexer and radio-performed according to specifications when tested with dummy loads. Specifications were met when connecting the antenna only to the radio, but when they were connected through the duplexer, a high VSWR was often indicated. When the same test was tried on another vehicle, it successfully passed the test. The discrepancy was caused by different cable lengths. Although the antenna by itself was OK, the duplexer added a small amount of inductance on VHF that either canceled or worsened, depending on the length of the cable. Similar effects occurred at 450MHz.

The solution was to have all cables cut to an “electrical” multiple of Pi lambda to provide a constant reference point whereby each faulty component could be identified. Physical length is unreliable because of manufacturing differences among cable suppliers and varying velocity factors.

Instrumentation To correctly establish a reference point, a test cable is assembled to the electrical 1/2 lambda, including the NMO antenna mount and the UHF-to-NMO adapter.

This cable is attached to the wattmeter for subsequent measurements.

Connect a cable to port 1 of a coaxial “tee.” Connect a 50ohm resistive (dummy) load to port 2 of the tee. Connect the male side, port 3, to a wattmeter, (one capable of forward- and reverse-power readings), as shown in Figure 2 above. Temporarily connect the other end of this cable to the UHF-to-NMO adapter, and screw it onto a NMO antenna mount. Do not attach cable to the NMO mount at this time. Use a 5W portable radio at the desired frequency as the wattmeter power source for the measurements. Cut the cable until no reflected power is indicated on the wattmeter. The result is an electrical Pi lambda, including the NMO mount. You have just set up the lab conditions whereby all further tests will be done.

Cutting the coax Once this benchmark cable is available, the process can begin. The vehicle installation will obviously include a cable from the NMO antenna mount to the transmitter location. To cut this coax to the right length, install the NMO-to-UHF adapter with the instrumentation cable, wattmeter and dummy load as previously described and shown in Figure 2. Apply power to the wattmeter and note the reflected power. Cut the coax cable in small segments (1/4 inch) until the wattmeter reads “zero” reflected power. At that length, the cable is a multiple of an electrical 1/2 lambda. The 1/2 lambda cable is open where the cut was done, which reflects to the wattmeter an open circuit that does not affect the 50 ohm fixed load.

Now install the proper connector on the transmitter end of the cable and connect this end to the appliance of your choice. If you are now curious about the velocity factor of the cable, measure the physical length and compare it to the 1/2 lambda formula of 5,905/MHz) in inches. The quotient divide d by the physical cable length is the velocity factor-usually 0.66 for most coax and 0.8 for foam.

Because the cable is some multiple of 1/2 lambda, all measurements will be reasonably accurate. If some component other than the cable is not within specifications, it is easy to replace one or the other items connected on either end of the coax cable.

Note that 450MHz = 3 X 150MHz. A 1/2 lambda for 150 MHz is 3/2 lambda for 450 MHz. Makes it simple, doesn’t it? When using a duplexer for dual-band operation, it is necessary to make the cable connecting each radio to the duplexer also a multiple of 1/2 lambda. Now that everything is compared to a 1/2lambda, any measurement out of order can be found by replacing the defective component. Should one, by chance, replace a good antenna with a good antenna, test results will be the same. However, replacing one bad component with a good one will yield immediate acceptable results.

Samples of cable lengths of RG8X used in the WSP shop are as follows: * wattmeter to UHF/NMO adapter, 29.5″ (tip-to-tip). * 450MHz radio to duplexer, 23.0″ (tip-to-tip). *150MHz radio to duplexer, 21.25″ (tip-to-tip).

Note: The Patrol has found that it is most efficient to use RG8X instead of the normal run of RG-58/U, especially for the 450MHz band. Some vehicle installations have coax runs in excess of 25 feet. Instead of having two sets of cable and connectors, it is easier, and most cost-effective, to have one type each of cable, connectors and crimp tools on the job.

This attention to detail may seem trivial for the shops having but a few mobile installations per month. However, for “mass production,” this concept may prove well worth the effort, especially if one cares about workmanship. Additionally, improving the RF power standards will yield better reception for your fleet vehicles operating in fringe coverage areas.

Buller is an electronics design engineer for the Washington State Patrol, Bellevue, WA. He is also a member of IEEE, NARTE, RCA, APCO and ARRL.