Field testing and maintenance of 3G basestations
By Stephen Hire and Ian Rose of Aeroflex Wireless
Comprehensive on-site testing of base stations using real-world scenarios, both at initial installation and then during ongoing maintenance, plays a vital role in preventing and solving performance problems before they impact subscribers. Thorough independent testing gives operators greater confidence in the quality of the deployed network. Poorly performing base stations have a significant impact on the quality of service experienced by users of the higher data rate services available on EDGE and 3G networks. Whether this is caused by incorrect installation, a gradual degradation in performance or complete failure of a particular module, the end result is that subscribers will suffer poor performance and will be less satisfied with their service provider.
Traditionally, test equipment has been selected on the basis of test coverage versus price. However, today, operators additionally need to consider field technician productivity and skill level – favourable changes in either of these can have a significant impact on the cost of maintaining the network and more than offset increased spending on test equipment.
For 3G test equipment, this means that test coverage, test speed and ease of use are absolutely critical and compromising on any of these is likely to be a false economy. Inadequate test coverage can also mean that routine maintenance misses potential problems resulting in repeated visits and additional costs from return visits to the same site. The rest of this article looks at what testing needs to be done on a Node B.
What effects do problems have?
Figure 1 shows a cell site, the outer circle of which represents the planned coverage of the cell. However, if the Node B is not operating correctly, the actual coverage could well be smaller as represented by the inner circle. This would mean that subscribers in the shaded area could suffer dropped calls or even a loss of coverage.
Figure 1: Cell site showing planned coverage In reality, a subscriber in the shaded zone will probably be covered by adjacent cells. However, this makes the network less efficient and at peak times cells may reach capacity faster resulting in increasing numbers of subscribers getting a network busy signal and being unable to make or receive calls.
Add to this the cell-breathing effect common to all CDMA systems, and it makes it all the more important that there are no differences between the actual and planned cell site coverage. Unlike FDMA or TDMA systems, the size of a CDMA cell is not fixed. With cell breathing, as more and more calls are set up, it becomes harder to separate weaker signals from the combined spread spectrum signal. As a result, a weak (i.e. more distant) signal can be demodulated more easily when traffic levels are low and there are fewer interfering signals than when traffic levels are high. This means that the cell size shrinks as traffic levels increase and then grows again as they fall off.
What tests need to be performed?
Looking again at Figure 1, it is obvious that, if the Node B transmitter is not working correctly, signals from the base station will not reach the outer limit of the cell. But what is often overlooked is the impact that a poorly performing receiver in the Node B will have. If the receiver sensitivity is below specification, it will become harder to demodulate signals from more remote mobiles. And as traffic levels increase, the effect of cell-breathing will become more pronounced if the sensitivity is poor. It is, therefore, important to test both the transmitter and receiver paths in the base station. It is also advisable to make some functional checks of the base station as a whole.
The performance of the Node B will have been fully checked as part of the production process. However, during shipment there is the risk of damage. The configuration could also be incorrect for the network on which it is to be installed. If such instances it is important to test these two aspects prior to connection to the network. If testing is not done prior to connection to the network, there is a risk that the Node B could either interfere with other operators or provide poor quality of service for users.
Unfortunately some manufacturers do not test the Node B in its final configuration. Individual modules are tested and the complete Node B is not assembled until the unit is installed on site. If damage occurs during shipment, for example a component or module fails or a connection breaks, this will almost certainly show up as a failure on several tests. Therefore, only a subset of the production tests need be carried out to confirm that the Node B has not been damaged in shipment. Another important consideration is that, with many Node Bs being located at the tops of buildings or on hilltops, installation and maintenance engineers prefer to use just one or two pieces of easy-to-carry test equipment on site.
Because of this, it is very rare that adjacent channel selectivity, blocking or receiver intermodulation are carried out during installation or maintenance. Any problems affecting the receiver will almost certainly show up as a failure of the reference sensitivity level. Similarly, spurious emissions are not usually tested as most will show up as failures of the spectrum emission mask.
For installation and maintenance testing, there are a number of other requirements that are not covered by the 3GPP tests. For example, each operator will want the Node B to be configured in a specific manner. To ensure that an incorrect configuration (perhaps because two modules are swapped over) does not cause network problems, it is useful to conduct functional tests before the Node B is connected to the network. Also, VSWR or distance-to-fault tests should be carried out on antenna feeders to make sure that they are also working correctly.
The following table shows the key tests that are usually carried out at each stage. Conformance testing requires many tests, whereas installation and commissioning requires fewer but does add some additional ones that aren't in 3GPP 25.141. The table has been produced based on Aeroflex's extensive experience of dealing with GSM manufacturers and operators and also from discussions with 3G Node B manufacturers and operators.
The above lists have been produced based on Aeroflex's extensive experience of dealing with GSM manufacturers and operators and also from discussions with 3G Node B manufacturers and operators.
The importance of receiver testing
With 3G, the operation of the transmitter and the receiver are closely linked. Any problem with the receiver is likely to have an impact on the operational performance of the transmitter. If the Node B's receiver is insensitive, for example, the power control loop will instruct each mobile to increase its output power to maintain the target bit error rate for the uplink. As each mobile is a source of interference to all the other mobiles, and the multiple access system works by sharing power across the code domain, the result is a reduction in uplink capacity. Network performance is degraded for everyone.
The reference sensitivity test provides a single uplink channel at minimum power to the Node B's receiver, simulating a mobile at the edge of the cell under ideal propagation conditions. If the Node B can decode the signal and meet the BER requirement, its small signal performance is good.
The dynamic range test simulates the scenario where many mobiles are active, by setting up an uplink channel for a single mobile (the wanted signal), along with superimposed Additive White Gaussian Noise (AWGN) to represent the interference from all the other mobiles. The unwanted noise is much greater in overall power then the wanted signal so the test measures the Node B receiver's ability to extract the wanted signal from the interference with a higher overall power at the Node B's receiver. For network operators, testing the performance of the receiver is critical to ensure that the service that they provide to their subscribers is good enough. Performing the reference sensitivity and dynamic range tests is usually adequate to verify the performance of the receiver.
How receiver tests are made
On mobiles, receiver tests are usually made by activating an internal loopback. However, most basestations do not include an internal loopback. Even when there is an internal loopback available, the approach has three important drawbacks which can mean that potential faults are missed or are harder to diagnose. First, a loopback tests both the transmitter and receiver together making it difficult to determine which is at fault if there is a problem. Second, there is no defined pass/fail limit for the combination of the two as the standards only state performance for transmitter and receiver separately. An incorrectly chosen limit could lead to failing base stations to be declared as working correctly. Finally, where internal loopbacks do exist they do not usually cover the complete receive and transmit paths with the result that parts of the chain are likely to go completely untested.
The best solution is to include code in the test set that enables the Node B to be directly controlled and Iub data to be decoded. This allows the test set to put the Node B into the required state, and then to send the demodulated data back to the test set via the Iub interface. From this, it is relatively straightforward to measure the BER and to obtain accurate measurements of the receiver. Figure 2 shows the test setup.
Figure 2. Equipment configuration for receiver testing
Summary
Aeroflex's RIWS 6413A base station test set has been developed to deal with the scenarios outlined above. Its concept is the same as the company's 6113 GSM base station tester, which is used by GSM operators throughout the world. However, the 6413A is aimed at operators' and infrastructure vendors' field technicians installing and maintaining 3G Node Bs. It is designed to allow technicians to perform a complete set of transmitter and receiver measurements that will give confidence that the Node B under test is working correctly, and, if not, to give sufficient information to enable faulty modules to be replaced or repaired.
Contact information:
Aeroflex Wireless Burnham Berks SL1 6BE Web: www.aeroflex.com Info: wireless@aeroflex.com
Photo 1: Aeroflex's 6413A base station test set
Photo 2: A typical receiver BER measurement result
|
