Introduction

In a manufacturing environment QC (quality control) is an important process that ensures customers always get a product of consistent value. QC criteria identifies parameters that, when checked, certify the finished product is what it is meant to be.

QC is well established where work practices and skills are critical to the performance of the end product. This generally involves a process of objective testing and analysis specific to the product. It is not always practical to carry out the same tests that would normally be used in the laboratory or on an assembly line in a field environment, generally because of the portability of the test rig. Instead QC for the field installation of factory certified components is more subjective than objective relying on the knowledge, skill and reputation of the work force rather than a functional performance test. This is a high risk option in some cases and acceptance of the RF interconnection at a cellular BTS is a typical example. Here objective testing is limited to measuring RL (Return Loss) which, while an important characteristic, can often mask a poor quality installation.

The most effective way to check the physical installation of the RF interconnection is to test for PIM (Passive Intermodulation) however portability of available test equipment has previously made this difficult to do in the field. Triasx has recently removed this limitation by introducing the IMT1000 portable PIM test set.

PIM is a critical QC parameter when certifying the RF interconnection assembly quality.

PIM has long been understood by RF design engineers as a phenomenon that introduces additional, often unwanted interfering signals in any device or system carrying two or more wanted signals. These unwanted signals occur as a result of using incorrect materials or the poor physical design and assembly of the device, perhaps both. Once the design and selection of materials has been finalized any significant residual PIM will be present only as a result of poor physical assembly.

RF component suppliers are required to meet high levels of PIM during manufacture (generally close to or > 110dBm) to remove the possibility of significant internally generated interference in a live system. Achieving these levels requires fastidious attention to cleanliness and assembly processes. It is not uncommon for experienced assemblers to need more than one attempt to get it right. Interestingly, these finished components are assembled on-site without the benefit of a suitable objective QC test. The same fastidious attention to detail common to factory assembly is required here as recent testing of BTS sites in Australia has proved. Many installations fail the PIM test. In the finished RF interconnection, PIM figures are not an average of those for each interconnected component instead they are the direct result of the worst PIM source in the interconnection. This is usually a physical connection made during the onsite construction of this infrastructure. IEC 62037 defines methods for taking PIM measurements. These methods define the parameters of the test rig and describe the physical treatment needed to confidently assess the quality of the assembled RF system.

Poor quality assembly will make it very difficult if not impossible to co-locate technologies or frequency bands where PIM can be generated in the receiver band of any of the BTS receivers. Simply put, the receivers will become deafened by the PIM interference and this will result in low quality subscriber connection or dropped calls. Triasx iQA products and services will help installation teams ensure their work does not contribute to these problems as networks are constructed or expanded.

Limitations of RL (Return Loss) measurements as a QC parameter

It is common practice to use swept RL measurements as the only method of quality certification of the assembly of the BTS RF interconnection.

RL is an important measurement as it is an indication of how well the RF interconnection is able to carry, radiate and collect RF signals to and from the air without undue loss. It is a measurement of how well the RF interconnection is electronically matched to the antenna. Provided that the physical assembly of this infrastructure is of good quality RL will confirm the health of the system. It cannot measure for poor assembly. Bad terminations or loose or damaged connectors will not generally be identified by an RL sweep. In fact, feedback from the field provides details on how connections close to the antenna are sometimes “tweaked”, (over tightened or loosened) to achieve the required RL “sign off” on an installation. This practice, regardless of how common it may be, shows that RL sweeps can provide a false impression of the quality of the RF interconnection assembly. Only an RL sweep carried out after the installation quality has been certified by testing for PIM can be confidently accepted as a true representation of the expected performance of the antenna system.

Purpose of the RF interconnection

All wireless based communication systems consist of radio transmitters and receivers. The transmitters of a base station generally have enough power to radiate signals to the subscriber’s mobile phone at the edge of the coverage foot print. If this is lacking then it is possible to build a more powerful amplifier. The sensitivity of receivers, on the other hand, is now about as good as it can be making this component a major limiting factor for cell coverage. Radio signals need to be radiated and collected from the air and conveyed to and from the radio equipment, ideally without the introduction of attenuation or receiver deafening interference. It is the RF interconnecting infrastructure that is designed and installed for this purpose. A high quality RF interconnection will ensure that the BTS radio equipment can communicate reliably with the weakest practical subscriber signals.

Handicapping the receiver

In cellular networks the subscriber is switched from one cell to the next depending on the strength of the signal that is being received from their mobile phone or the bit error rate detected by the receiver for that connection. It is the receiver’s ability to reliably detect the weakest subscriber signals that determine the geographic topology of the network hence the number of BTS or repeater sites needed to provide the desired coverage. This factor is a major influence on the cost of the network and therefore the cost of providing the cellular service.

Any performance limiting characteristic of a BTS installation will mean more sites, significantly increasing the cost of service.

It is the physical RF interconnection and the electromagnetic environment that will have the most influence on the ability of the receiver to work at its best. The two primary limiting factors are losses in the RF interconnection and RF interference.

Losses are generally a function of the electrical characteristics of the components used to build the RF interconnection. Where these are significant Tower Mounted Amplifiers in the receiver path (uplink) are used to compensate for loss. RF interference is more complicated to determine and resolve.

RF interference is unwanted signals that fall within the receiver band with the detrimental effect of significantly reducing or blocking the receiver’s ability to detect subscriber signals. In some instances the interference comes from a source external to the antenna while in others it is internally generated.

Active components such as TMAs can be a source of IM (Intermodulation), however internally generated interference is more often caused by PIM as the result of a poor quality RF interconnection assembly. This is an unnecessary handicap on the receiver, the network and the cost of providing a cellular service.

Reducing Cost of service

In the future cellular subscribers will require better services (less dropped calls, more bandwidth and more functionality). Environmental concern will mean less BTS installation sites will be available and competition will mean Operators (Carriers) will need to reduce the cost of providing a service.

The demand for more cellular bandwidth, more services and greater capacity, combined with the increasing difficulty and cost of introducing new sites, is driving a trend to co-locate cellular technologies on existing infrastructure. This can be a very cost effective way of meeting subscriber demands while, at the same time, lowering the average cost of service. However, a poor quality RF interconnection will make this option impossible to implement.

In the past, PIM interference has been managed by selecting a combination of frequency bands that ensure any likely PIM interference will fall outside the receiver bands. Where this has not been possible, channels that were affected by this internal interference were not used. This is no longer an option for most network Operators, particularly in a spread spectrum CDMA and WCDMA environment where the presence of PIM can mean much more than the loss of one or two channels.

A quality certified RF interconnection will protect the investment in this infrastructure and potentially reduce the cost of providing the cellular services as subscriber demand forces an increase in capacity.

The consequence of accepting a poor quality RF interconnection installation is likely to be a higher than necessary cost of service in the future, if not from the start.

Results of poor QC in the RF interconnection

Figure 1 is showing a significant increase in dropped calls when in July 2004 a CDMA network, overlaid on GSM at 900 MHz, was switched on. This was caused by a poor quality RF interconnection.

Figure 2

Figure 2 shows that the interference seen here is on one channel only. This can only be internal interference resulting from a poor quality RF interconnection for one of the two receiver paths. External interference would equally affect both. This confirms the value of quality certification of the RF interconnection.

Figure 3

Figure 3 is the display from a spectrum analyzer on “Maximum Hold”, capturing wide band noise resulting from the presence of PIM in the RF interconnection over time. In this case GSM 900, CDMA at 800 and 1800MHz were co-located with resulting PIM products reducing the raising the noise floor to as high as -95dBm/30kHz. This loss in receiver sensitivity will mean the loss of significant calls as confirmed by the data collected to produce figure 5 below.

Figure 4

Figure 4 is showing a drop in mean hold time for subscriber calls during a period where the RF interconnection quality had deteriorated. Here the reduced time of each subscriber call directly relates to loss of revenue and the chance that an unsatisfied subscriber may be thinking of taking their business to a competitor. This was caused by overlaying multiple technologies on a poor quality common RF interconnection.

Outcome of Field trials
 

Figure 5