PIM: Components, Material, Handling & Testing
By Wolfgang Damm, AWT Global
New connector surfaces look and feel very smooth, but the picture changes quickly when viewed under a microscope. The atomic lattice size of metals is often no more than 25 Angstroms (0.0000025mm) wide. Machined metal surfaces will never have such a degree of smoothness. Metallic surfaces look very rough under high magnification. That causes surfaces of mated connectors to touch at only a few spots, called asperities. Tightening connectors applies localized pressure to these asperities, which causes them to deform. Deformations increase the contact area, but it is still limited to some “load-bearing” areas, called a-spots. They add up, but their overall area is still smaller by several orders of magnitude than the apparent contact surface of connectors.
Simplified, a-spots can be regarded as electrical RC models and a mated connector can be seen as a network of thousands of unequal RC circuits. Such a network does not behave in a linear way. Passing currents of different frequencies will respond differently. This causes passive intermodulation.
Structural material discontinuity also causes discontinuity in current flow. Regardless of the contact material, discontinuity in electron flow is characterized by:
Constriction Resistance - due to bending of current lines of flow in the vicinity of an a-spot
Tunnel Resistance - due to conduction through thin insulating contaminant layers via tunnel effect
Contact Capacitance - between the two essentially parallel equipotential surfaces.
Lengths of the constrictions are very short, so inductive effects are small compared to capacitive and resistive effects.
Skin effect is the property of alternating current to show higher current density closer to a conductor surface. Current flows mainly at the “skin” of the conductor, between the outer surface and a level called the skin depth. Skin effect is caused by eddy currents that are induced by the changing magnetic field of alternating current. The effect is more pronounced with higher frequencies. At 1 GHz, on a silver plated surface, around 98% of the current density occurs within approximately 0.01 mm of material depth. For comparison: an average human hair has a diameter of 0.08 mm, 8 times larger in diameter than the skin depth at 1 GHz. This fact underlines the importance of connector plating. It serves not only as protection for the connector but also carries almost all of the current in RF systems.
The remarkably diminutive skin depth of conductors at high frequencies is very susceptible to scratches in the material. Even if only on a microscopic scale, the tiniest groove, dent or jag interferes with homogeneous flow of current, and with that causes unwanted passive intermodulation.
Working with Low PIM Components
Whether connectors, cables or components, low PIM components are precision building blocks of RF networks. Low PIM products require manufacturing processes that meet highest standards, 100% quality sampling, careful handling and shipping with sufficient protection.
These components must also be treated carefully in the field to avoid degradation or damage. Since components like duplexers or loads are often hermetically sealed, their internal elements are relatively protected, but their connectors are exposed. This is also the case for cables. This chapter is about treating connectors of cables and component ports.
When Working with Low PIM
• Prevent mechanical damage
• No touching of RF conductors with bare fingers
• Avoid alien bodies of all kinds
• Avoid humidity
• Avoid electrical damage
Mechanical damage can be inflicted by a variety of events. Dropping a component is the most obvious mishap, but it can occur in the factory and during shipment if loose components are allowed to bump into each other. A less obvious cause is improper connector tightening. RF connectors are designed in a way that they can be screwed on almost completely by hand.
Wrenches are to be used only for the last half turn. It clearly indicates non-parallel mating of connectors if it is too tough to screw them on by hand. This can be caused by cables that are too short or too long, which apply sideways pulling or pushing forces to connectors. If RF connectors do not turn easily during mating, their threads and connecting parts are forced against their counterpart surfaces, causing extreme friction and even deforming. High forces can chip off parts of the plating. These—conducting—chips are alien bodies that interfere with the current flow in the RF path, causing passive intermodulation distortions. The cables’ geometry is paramount for proper functioning. It can be damaged by external force (denting) or too tight bending radii. A good practice is to install release cable loops to avoid both forces at the connectors and too tight cable bents. Such loops cost a bit more material, but the investment goes a long way as the installation tends to be more reliable over time. After connectors are pre-tightened by hand, they have to be mated with a torque wrench to apply exactly the right tightening force. Connections with too little torque result in insufficient contact force; too much torque causes contact areas to deform. Both are consecutively resulting in passive intermodulation distortion.
Alien bodies like dust, dirt, and metal chips can very easily find their way into connectors. Base station sites or in-building installations are never dust free, and dust and dirt kernels are difficult to avoid. Keeping protection caps on connectors helps. It is suggested to always wipe connectors with alcohol wipes and dry them with moist–free canned air before mating. While connector dust might not be visible to the bare eye, dark areas of used wipes will clearly show that it has been there. Connectors of test equipment, measurement cables and low PIM terminations have to be cleaned frequently with alcohol wipes and dried with moist–free air.
Connector wear is an issue that concerns test equipment including PIM analyzers, test cables and low PIM terminations. It is not so much an issue for field installations because connectors of low PIM components are mated only a few times, for initial system measurements and final mating. Test equipment, on the other hand, is in permanent use and has to endure many mating cycles. Manufacturers typically guarantee 500 mating cycles with sustainable PIM ratings before connectors start to degrade. The reason is clearly not low quality but the fact that asperities can undergo only a limited number of deformation cycles. Furthermore, attrition of conductors’ plating due to mechanical friction steadily reduces the thickness of the plating.
Humidity and moisture are creeping enemies of low PIM networks. Over time they cause oxidation. While initial measurements may look good, connectors with accidentally enclosed humidity and moisture will degrade. An often overseen but common source of humidity is human breath. It is tempting to blow into a connector to remove a little dust fluff. Never do it. Exhaled air has a relative humidity of 100%!
No Touching of RF
Sweat cools the body and skin lubricates itself with oily matter. What is helpful to maintain our health is adverse to proper function of RF connectors. Even minuscule amounts can alter PIM performance of connector contact areas. Low PIM RF conductors are very susceptible to such external influences. There is a good reason why manufacturers of low PIM components require their workers to wear gloves.
Electrical damage is easily overseen but often the cause of serious PIM problems. It can happen by applying power levels to a device that exceed its actual power rating. Without question, that has to be avoided. Another occurrence that happens sometimes unintentionally is mating and disconnecting connectors under RF power. If this happens, spark discharges are unavoidable. They cause craters in the material, altering current flow significantly, which again is a source of PIM.
Testing Low PIM Components & Networks
Low PIM components are a key factor when building telecommunication networks with the lowest passive intermodulation interferences. However, 60% of all PIM issues are not caused by faulty components but are manmade and happen during installation. This is particularly true and happens during installation, mainly because RF cables are usually assembled in the field.
Unintended scratches in the plating, chipping, and entrapped dust are just a few of many issues that can occur. The only way to ensure that base station installations operate at the expected low PIM levels is to conduct thorough PIM tests of both individual RF branches and the complete installation. Three simple PIM tests have gained general acceptance and serve as excellent reference for both installation and component testing. The tests are described below. They will detect virtually all sources of PIM in cables, connectors and components.
The static test analyzes both components and cables. PIM analyzers that deliver continuous 2 x 20W measurement signals are connected to the Device Under Test (DUT). PIM measurements are performed for at least 30 seconds to fully energize the system and also apply thermal stress to the tested components, similar to a live telecommunications signal. This test detects bad materials, scratched surfaces and alien bodies like dust or metal chips in the RF path.
Measurement values of static tests need to be below the required limit, but they should also be stable. The signal should not alter too much during the measurement. Changes of 2-4 dB are acceptable; higher swings even if they are within the required limits can be an indicator of (future) PIM problems.
The first dynamic test, also called the “wiggle test,” checks the quality of assemblies between cables and connectors. Tested cables are moved in a circular way (turn diameter about 10 cm). The test is to be conducted with at least 10 turns in each direction. Wiggle tests detect loose contacts and poor workmanship of cable assemblies. They find also bad soldering and shielding cracks. PIM measurements must be stable throughout the test. Correct cable-connector assemblies will endure this mechanical stress test easily. The second dynamic test is known as the “tap test.” It requires PIM analyzers to be set to “PIM versus time mode.” PIM readings are continuously shown over a time axis in this mode. A harder device, made of wood or plastic material but not metal, is used to tap 10 times at all connectors. Field technicians often use the handle of a screw–driver for this test. It is hard, but does not dent or scratch the connectors. PIM readings should stay stable during this test. Possible contaminations like dust, metal chips or other alien bodies in the connectors will cause spikes in the reading whenever the connectors are tapped. The remedy is to open the connectors again, clean them with alcohol wipes and dry them with moisture–free pressured air. Afterwards, the tap test has to be repeated.
this article to a friend!