The broadband noise generator: an alternative to sweep testing
September 1, 1997
In many routine uses, the broadband noise generator can be a viable alternative to hauling out the heavy old sweep generator and going through the hassle of interconnecting all the cables and setting up the system for a simple sweep test or alignment. The BNG-1000A broadband noise generator from Avcom, Richmond, VA, as shown in Photo 1 below, can, in many instances, greatly simplify sweep testing and alignment chores.
The unit is lightweight and portable_weighing in at just 2 pounds (0.9kg). The beauty of the instrument is its utter simplicity of operation and use. Aside from the on/off switch, the only operator control is the bypass switch that enables the operator to send the noise directly to the spectrum analyzer for a reference-level reading. The unit is powered by external power (112Vdc to 124Vdc @ 200mA) through a coaxial dc power jack on the rear panel. A 115Vac to 12Vdc adapter is provided with the unit. The unit is designed to operate over a bandwidth from 3MHz to 1,000MHz. The output level is about 230dBm 13dB over the range.
Using the noise generator As shown in Figure 1 below, there are only three cables to connect: from the BNG-1000A to the DUT (device under test), from the DUT back to the BNG-1000A and from the BNG-1000A to the spectrum analyzer input. All connectors on the BNG-1000A are BNC, allowing quick and easy cable connections. Figure 2 at the right shows a spectrum analyzer display with no video filtering. This trace is obtained by switching the switch on the noise generator to the “bypass” or “direct path” mode. Notice how thick the noise appears on the spectrum analyzer display. To get a sharper image on the display, video filters on the spectrum analyzer can be activated. Figure 3 at the right shows a sharper image after the wideband video filter of the spectrum analyzer is switched in. Figure 4 at the right shows the best image using the narrowband video filter. Be aware that as filter bandwidth becomes more narrow, the sweep rate of the spectrum analyzer’s timebase must be reduced. If the sweep rate is too high, the trace will become distorted. To get an accurate trace, the sweep rate must be correct for the frequency span, resolution bandwidth and type of video filter chosen. Many spectrum analyzers provide a warning indicator to signal the operator that the chosen sweep rate is too high for the frequency span, resolution bandwidth and video filter settings. Other analyzers prevent improper sweep rates by interlocking controls.
Typical displays To check the frequency response of a quarterwave bandpass cavity, the setup shown in Figure 5 at the lower right is used. The typical response curve is shown in Figure 6 on page 59. The setup for a notch filter is the same as shown in Figure 5, along with the response curve in Figure 7 on page 59. Figure 8 shows the setup for checking the return loss on an antenna. A return loss bridge such as one shown in Photo 2 on page 60 would be appropriate for this test setup. The chart shown in Table 1 on page 60 converts return loss to VSWR. The antenna is connected to the DUT port of the return loss bridge. With the spectrum analyzer set up to sweep the appropriate frequency span, the resonant frequency of the antenna will appear as a null in the display, as shown in Figure 9 on page 60. The minimum amplitude represents the resonant frequency. Note the amplitude at the null, then remove the antenna and leave the DUT port open. Note the new amplitude at the resonant frequency. The difference between this amplitude and the amplitude of the display with the antenna connected is equal to the return loss of the antenna at the resonant frequency. For example, if the difference is 20dB, then the return loss is 20dB and the VSWR is 1.22, according to Table 1. The operating bandwidth of the antenna can be easily determined from the display. For example, if the maximum allowable VSWR is to be 1.5:1, then check the frequencies (minimum and maximum) where the return loss is 14dB.
Another use of the noise bridge is to check the operation or tuning of an isolator. The setup shown in Figure 10 on page 61 is used for checking the response of an isolator. Note that the output of the noise generator is fed to port 2, or output port, of the isolator. With the proper frequency span on the spectrum analyzer, the frequency response of the isolator will be displayed. The null should be located at the desired operating frequency. If not, the isolator can be tuned to the desired operating frequency as long as the desired frequency is not too far away from the frequency for which the isolator was designed. To get a reference mark, the noise generator switch can be switched to bypass the DUT. The difference between the reference mark and the null amplitude is equal to the isolation of the isolator in decibels.
The frequency response of the narrow bandpass filters in a receiver can be checked as shown in Figure 11 on page 61. Note that a probe is used to sample the low IF output beyond the bandpass filter(s). The probe is connected to the “return from dut” port on the noise generator. Since the RF front end and high IF sections of the receiver are wide compared to the low IF section, the overall response curve represents the low IF filters.
Summing up The BNG-1000A noise generator can be used to check the frequency response on almost any frequency-sensitive device. The results you get will depend somewhat on the quality of spectrum analyzer you use with the device. Some of the low-end spectrum displays on service monitors are not satisfactory for this purpose. Others are well-suited to the task.Remember to keep the sweep rate down to prevent distorting the waveform. Slow sweep rates are best for getting the most accurate representation of the frequency vs. amplitude response on the spectrum analyzer display. This may cause a bit of a problem when trying to tune to a peak or a null. If the sweep rate is too slow, the display won’t follow the tuning fast enough for real-time response. In this case you may have to speed up the sweep rate and set the notch or peak to the correct frequency, and then reduce the sweep rate to check for the accuracy of tuning. It may be necessary to go back and forth several times to get the peak or null at the correct frequency.
If your spectrum analyzer has a manual sweep feature, use it to set the spectrum analyzer to the desired peak or null frequency and tune the device for maximum or minimum amplitude of the dot on the screen. Then recheck the response curve using a slow sweep rate again.
Although the BNG-1000A broadband noise generator can’t beat the tracking generator and spectrum analyzer combination, it definitely has its place in the two-way shop. The simplicity of operation and portability would certainly enhance its use in the field as well.
For further information on the BNG-1000A broadband noise generator contact: Avcom of Virginia 500 Southlake Blvd. Richmond, VA 23236 Phone: 804-794-2500 Fax: 804-794-8284
Until next time_stay tuned!
Kinley, a certifiied electronics technician, is regional communications manager, South Carolina Forestry Commission, Spartanburg, SC. He is a member of the Radio Club of America. He is the author of Standard Radio Communications Manual: With Instrumentation and Testing Techniques, which is available for direct purchase. Write to 204 Tanglewylde Drive, Spartanburg, SC 29301..