Diagnosing Passive Intermodulation

Executive Summary

A PIM test does not tell you what is broken. It tells you that something, somewhere in the tested path, is generating intermodulation — and on a modern multi-band network, that path can run through a dozen connectors, several metres of cable, and whatever metalwork happens to be nearby.

The instinct, faced with a high reading, is to blame the jumper or cable under test. That instinct is usually premature. A high reading confirms that intermodulation exists somewhere in the measured path. On its own, it says nothing about which component is responsible.

This paper sets out a systematic method for closing that gap: verifying the test system, inspecting connections, substituting components one at a time, weighing installation and mechanical effects, and accounting for environmental PIM — in that order, before anything gets replaced.

Done properly, this cuts unnecessary returns, avoids repeat site visits caused by fixing a symptom instead of a cause, and gets the fault resolved the first time.

1. Introduction

Passive Intermodulation occurs when two or more RF signals meet a non-linear point in a passive system and generate additional frequencies. Where those products land inside a receive band, they raise the apparent noise floor and degrade system performance — invisibly, until someone runs a test.

Connectors, cable assemblies, antennas and filters can all generate PIM. So can poor electrical contact, contamination, corrosion, mechanical damage, and dissimilar metals working against each other at a junction.

Establishing that a high reading exists is rarely the hard part. Establishing where it is coming from is. A PIM test measures the complete path between the analyser and the termination, not a single component in isolation, so the result reflects everything sitting in that path ^[1].

That makes replacing the most obvious part an unreliable fix. If the replacement doesn't address the actual source, the fault comes back. And even where a reading does improve, changing several things in the same visit makes it impossible to say which change did the work — which means the next fault on the next site gets diagnosed with no more certainty than this one.

The more reliable approach: establish a known baseline, then change the system in a controlled, recorded sequence, one variable at a time.

2. Understanding the PIM Measurement

A PIM analyser injects two RF carriers into the system under test — per IEC 62037, typically +43 dBm (20 W) each — and measures the intermodulation products that result, expressed in dBc relative to the carrier power ^[1].

Figure 1 — The test path. Every stage between the analyser and the termination is inside the measurement.

That measured value belongs to the whole arrangement, not to any single part of it. The analyser, test leads, adapters, load, connectors and device under test are all inside the loop. On an installed system, cable routing, mechanical condition and nearby metalwork join them too.

This is the detail that gets missed when a result is read too quickly. A high reading proves intermodulation is present under the conditions tested. It does not locate the source within them. A jumper measured against one test arrangement can produce a materially different result against another, not because the jumper changed, but because the configuration, the connections, or the conditions around it did.

PIM should therefore always be measured against a known, repeatable arrangement. Without one, two genuinely identical components can appear to perform completely differently, and neither reading can be trusted.

3. Establishing a Test-System Baseline

Check the test system before the assembly under test. If the analyser, leads, adapters or termination are generating PIM of their own, nothing measured downstream of them means anything.

A verified low-PIM load establishes this baseline: measure the test system's residual PIM before the device under test goes anywhere near it, and hold that baseline to a tighter limit than the one you're actually trying to meet. Hughes' own precision test loads, for reference, are independently rated to an absolute minimum of −165 dBc while handling 50 W continuous ^[7] — that is the kind of margin a baseline load needs below whatever limit the job specifies, so the load itself is never in question.

If residual PIM drifts between repeats, or sits higher than expected, that is the test system to investigate — not the assembly under test. Chasing a fault through the device under test while the analyser's own leads are the actual source wastes a site visit before it has properly started.

4. Inspecting Connections and Interfaces

Once the test system is verified, the connections within the test path come next.

Connector interfaces get investigated early because their performance depends on two separate things: the condition of the mating surfaces, and how well the connector was assembled in the first place. Either one alone can cause a fault. Most real faults on site are some combination of both.

None of these findings, on their own, prove a source. A worn thread or a dissimilar-metal junction is a condition to weigh against the rest of the investigation, not a verdict to close it on.

5. Controlled Substitution

If the test system and the connections both check out and the reading is still high, the next step is substitution — one component at a time.

Repeat the cycle for each remaining component in the path.

This is precisely what controlled substitution is for. Swap a jumper, reseat an adapter, alter the cable route and adjust connector torque all in the same pass, and an improved reading afterwards tells you nothing about which of the four changes actually did the work. The more variables move at once, the less the result can tell you — and the more likely a working component gets condemned in place of the real fault.

6. Installation and Mechanical Effects

PIM performance can also be affected by the way a cable assembly is installed. An assembly that performs correctly during a controlled test may produce a different result when subjected to mechanical stress in an installation.

Controlled Mechanical Testing

Where appropriate, controlled mechanical movement can also be used as part of the investigation. A defined section of cable or a connection can be lightly flexed or tapped while monitoring the PIM result.

If the PIM reading changes when a particular section is moved, the test can be repeated to determine whether the change is consistent.

A repeatable change may indicate a problem associated with the cable, connector, termination or mechanical installation. The result should then be investigated further rather than treating the movement itself as proof of a particular fault.

7. Environmental and Alien PIM

The source of PIM does not always form part of the electrical circuit being tested.

Metallic objects in the surrounding environment can interact with RF energy and may contribute to the measured result. This is sometimes referred to as alien PIM.

When investigating an unexpected or intermittent result, the surrounding installation should therefore be considered as well as the RF components themselves.

This can be particularly relevant when a result changes with the position or movement of the installation, or when a result obtained on site cannot be reproduced under laboratory conditions. The surrounding environment should not be assumed to be the source of the PIM, but it should be included in the investigation where the measured behaviour indicates that it may be relevant.

8. Test Records and Reproducibility

PIM troubleshooting depends on being able to compare measurements. Test conditions should therefore be recorded in sufficient detail to allow the result to be reproduced.

The sequence of changes is particularly important. Recording the initial result, the change made and the subsequent result provides a useful record of the investigation.

Where a component is identified as a potential source of PIM, it should be retained for further testing wherever practical. Cleaning, re-terminating, repairing or disposing of the component before the investigation is complete can remove evidence that may be needed to establish the cause.

9. A Practical PIM Troubleshooting Procedure

When a PIM result is higher than expected, the following sequence provides a practical starting point for the investigation.

Figure 2 — The troubleshooting sequence. Unresolved readings return to substitution for the next component rather than moving on early.

If step 4 doesn't resolve the reading, the cycle returns to step 3 for the next component rather than moving on to installation and environment prematurely — substitution is exhausted before the investigation broadens.

10. Conclusion

A high PIM result does not, by itself, indict a cable or a jumper. The measurement spans the complete test path, and the equipment, the connections, the installation and the surrounding environment all sit inside it.

Replacing the most visible component is rarely the fastest way to resolve that — it is usually the slowest, once the return shipping, the re-test, and the second site visit are counted.

A reliable investigation starts with a known test-system baseline, moves through the connections, and substitutes components one at a time while recording the effect of each change. Where the evidence points there, it extends to mechanical condition and environmental PIM. What holds the whole process together is reproducibility: without a result that repeats, there's no way to know whether a change fixed the fault or just coincided with it.

For network operators, that discipline is the difference between a fault fixed once and the same fault revisited on a second callout. It is also, over enough sites, a meaningful reduction in components returned that were never actually at fault — see Hughes' companion papers on connector-related PIM and VSWR in cellular networks for the mechanisms behind two of the most common culprits ^{[6, 8]}.

11. References