Radio frequency interference occurs in everything from commercial wireless networks and devices to military communications, radar and electronic warfare (EW) systems. Addressing this problem can be especially difficult since measuring interference is unpredictable. Additionally, the intermittent failure modes in typical signal analyzers make data capture particularly challenging. Consequently, when the root cause of a problem is not yet known, it can be difficult for engineers to set up a measurement that captures the failure.
Despite the challenge, the task of finding, identifying and analyzing interfering signals in a crowded spectrum—whether intentional or not—has become increasingly important in a wide array of applications. One RF recording technique that may prove particularly useful in addressing this problem is gapless capture. Using this technique, system engineers can now measure data continuously over long durations and ensure the capture of all RF events when they occur.
Understanding the measurement challenge
When characterizing system interference, system engineers have traditionally relied on a signal analyzer performing continuous long-duration recordings, as shown in Figure 1. The main limitation to long-duration recording is that test equipment typically has limited on-board memory. Signals-of-interest enter the analyzer’s RF input and are processed by subsequent stages, resulting in the displayed waveform on the right of Figure 1. Up to the blue vertical line, all signals-of-interest within the instrument’s capture bandwidth are processed in real-time, assuming a fixed local oscillator. However, once the samples fill the memory buffer or RAM, the instrument no longer looks at incoming digital samples. Instead, it must process previously recorded samples.
Figure 1: Shown here is a block diagram of a typical signal analyzer Click on image to enlarge
The signal analyzer does not capture any samples while it post-processes previously captured data, effectively creating a gap in its continuous acquisition of data. If events occur while the previous event is being processed or if the new event lasts longer than the available memory, it falls into this gap and may be missed. Moreover, the analyzer’s trigger setup only captures signals for one set of limited conditions. Once the analyzer fails to capture the event, it is gone forever.
Not covered in this paper but also very useful is the fact that these recording systems have significant amounts of pre-trigger capability that allow us to record many seconds of data prior to the actual trigger event.
This opens up the possibilty of applying some signal processing to the data in the triggering channel to make a sophiticated criteria based trigger.
It doesn't matter that this processing could take seconds; the pre-trigger capability of the recording channel provides the time.
Thank you for your comment that it is actually the knowledgable Engineer that is critical to solving the interference problem and your keen observation [borne out of experience] that there is often a whole lot of detective work done before the measurement system even becomes necessary. I certainly agree with your assertion.
This was a short paper in which I attempted to show that a recording system can be very useful for helping to root-cause a specific interference problem caused by signals in the external environment that are both intermittent and probably non-compliant.
Having a permanent record of the signal environment, which is known to cause an interference problem with some RF system or device, affords us the luxury of processing this data "offline" back in the lab.
Even with a recording system, it can still be tricky to capture the actual culprit signal, as it occurs, because the analyzers triggering system may false trigger on other signals too.
Note that I'm suggesting it is generally not too efficient to record hours and hours of data.
Even a few seconds of recording at wide RF bandwidths can produce Giga Bytes of data.
After all, who has time to process that much data.
It is usually much better to selectively record.
In that case I posited that an additional independently tunable receiver could be tuned specifically to the frequency and bandwidth of the suspected intererence signal to provide more selective triggering capability, since the recording in the afflicted channel could be initiated by this trigger channel.
This is funcionally equivalent to a frequency selective trigger with the added benefit that we can trigger off of out-of-band signals.
in standing down, the users, will tell you the noise interference problems, and the locations, and when they have occurred , "you just have to ask" don't worry about ambiguity, and when you have solved, " good luck " ask again.
Dave, i so enjoy your company's expertise. i value the technical thought as first class. active voice and problem solvers, " i thank you "
from Ralph T. Gerwing yvr.ca Metro Vancouver Cheers.
pretty penny this analyzer your promoting, but, in real world problem solving " especially building new band services " it is very valuable indeed.
"" Tracking down interference in complex RF environments "" well yea, but:
Interference usually is in the receivers, hand-helds, base stations or auxiliary receivers.
If the hand-helds are moving about, then interference zones may be noticed by the users. , in a particular building, a geographic square block, or fraction of a larger area.
The engineer is in the know, who is a close frequency user. as adjacent channel user, or third intermodulation contributor. all of this comes into play "the math" and analysis before, the recording data comes into play.
The recorder itself, just shows, the place and time values. ' it is not really the solver of the fact " however an item of merit you may of wanted to touch on was.
Propagation measurements, of band by band comparison, to fulfill an existing radio footprint plan system. that is of merit, similar or better.
In my studies of this, equal merit or better, the recording data is "Missing" or has not been done at all.
A lot of new systems are poorly designed in coverage. getting in building coverage has been a failure, the recording analyzer will tell you what you have now, and when the new system goes in, what you have got to replace it with .
Now i am simplifying my thought here, as this is a very important topic. \
i will say, in my work experience as a principal, i ordered ten percent of project value to be in Test equipment, and bought many pieces of gear, by your wonderful company ""AGILENT""
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.