June 27, 2017
Passive intermodulation (PIM) is one of the most important measurements that needs to be made on distributed antenna systems (DAS), as well as one of the more challenging. Its importance stems from the fact that PIM can significantly inhibit the ability of a DAS to operate at KPI levels established by mobile carriers.
Complicating matters further is locating the source of PIM when tests indicate it is affecting performance. One reason is that there may be hundreds – if not thousands – of potential PIM causes in a DAS environment. Lighting fixtures, fire suppression pipes, HVAC ducts, and ceiling tile frames are just a sampling of the many elements that can lead to PIM.
Figure 1 shows a typical DAS configuration that has 63 RF connections, 31 cable assemblies, 15 antennas, 14 power dividers, one hybrid combiner, and one RF termination. It only takes a single PIM issue on one of those branches in the sector shown to create enough interference to make the DAS “blind” to a mobile device.
When conducting PIM tests, selecting the proper analyzer is important. One specification to keep in mind is the frequency coverage of the instrument. Frequency is important when measuring PIM because a neutral-host DAS is typically designed to support bands from 700 MHz - 2600 MHz to satisfy a variety of mobile operators.
Just how essential is frequency when locating and determining PIM? To answer this question, we conducted a series of tests that included measuring components, connectors, and antennas.
Testing DAS Components
Our first test was to measure components and adapters used in a DAS system. Separate PIM analyzers with frequency support of 700 MHz, 850 MHz, 1900 MHz, and 2600 MHz were part of the test. For each measurement, a QN connector was used on the analyzers to minimize the possibility of metal flakes being created by the constant connecting and disconnecting associated with the tests. By doing so, we had greater confidence in the data acquired.
The tests revealed that PIM response varies greatly depending upon the frequency. At the lower bands (700 MHz and 850 MHz), the frequency penetrated the nickel skin of the DAS components while at the higher bands (1900 MHz and 2600 MHz) the nickel was not excited by the frequency. As a result, there was a higher magnitude of PIM at the low frequencies.
The second test was to determine the impact connectors can have on PIM levels. We took output from four analyzers and merged them using a hybrid combiner, then ran them to a single-paired connector mounted on a vibration plate.
Four connectors – Type N, 7/16 DIN, 4.1/9.5, and 4.3/10 – were measured. The configuration used allowed for simultaneous PIM measurements for highly effectively measurements of the magnitude of PIM impact at four frequencies. A baseline test was conducted on each when they were tightly connected. Once the data was acquired, we purposely loosened each connector slightly.
Among the findings was that Type N connectors used at 700 MHz had excellent PIM levels when the connector was tight. Once they became loose, however, the PIM levels escalated to failing levels. Similar results were achieved when testing the 7/16 DIN connectors. The 4.1/9.5 and 4.3/10 connectors passed at 700 MHz and 850 MHz but failed at the higher frequencies. A complete breakdown of each connector group’s performance is shown in figure 2.
Not surprisingly, the test showed loose-connector PIM is worse at a higher frequency. Overall, there was a 9 dB to 28 dB increase in PIM levels from 700 MHz to 2600 MHz, depending on the connector type.
Our final test was to measure the effect of PIM at the antenna. Once again, four PIM analyzers were used. To create a real-world environment, an omni-directional antenna was placed on a mast and an engineer carried it while walking around an approximately 30-square-foot office.
One factor that was considered was the large loss between the antenna, combiner network and PIM analyzer. Loss was 20 dB at 2600 MHz and dropped all the way to 16 dB at 700 MHz.
Multiple tests were conducted, with the power level being adjusted each time. The first test was done with 30 dBm power at the antenna. Results showed that PIM levels were so high throughout the room at 700 MHz and 850 MHz that neither frequency could be used in this setting. At 1900 MHz, 67% of the room passed, and 83% of the room was acceptable at 2600 MHz.
When the power level at the antenna was 24 dBm, 700 MHz remained unusable while an 850 MHz DAS system would provide coverage over 20% of the room. Again, the higher frequencies performed well, as 95% of the room had coverage at 1900 MHz and PIM levels were passing in the entire room at 2600 MHz. A 700 MHz DAS finally provide some coverage – 25% of the room - when the antenna power was 8 dBm. Coverage was 57% at 850 MHz, 99% at 1900 MHz and 100% at 2600 MHz. Figure 3 provides a complete summary.
Overall, the test revealed that external PIM is much worse at the lower frequencies, which was to be expected. One thing to remember is that if the power levels are lowered to achieve coverage at 700 MHz and 850 MHz, more antennas must be installed.
Another finding is that if PIM levels are acceptable at the lowest frequency, they will also be passable at the higher bands. Where not restricted by DAS system components, a 700 MHz PIM analyzer can be used to conduct the tests on all DAS networks. If the DAS has filters in it, however, it may be necessary to test with higher frequency analyzers, as well.
To learn more about this, watch this informative PIM webinar on the importance of frequency when measuring PIM.