Sensitivity Training

Of all of the measurements made on today's sophisticated mobile radio handsets, perhaps no other measurement can have as dramatic of an effect on overall performance and customer satisfaction as receiver performance tests. A phone receiver's sensitivity that is only a few decibels out of tolerance can play havoc with even the best net-work's performance. New phone designs need to be tested for proper sensitivity performance before they are allowed in customers' hands.

Sensitivity is commonly defined as the smallest RF signal that can produce a usable baseband output from a receiver -- the key word being usable. When testing the analog radio, you determine what is a usable level via a sinad ratio. For instance, the 12dB sinad test refers to the RF level required to reduce the open squelch distortion by 12dB when a 1kHz tone is injected (typically a level of 0.35mV). Any level below this would not produce a usable baseband output.

In the digital world, such ratios are not practical. Instead, you would measure the bit error rate (BER). In some formats, with various forward error correction (FEC) systems, you would use the frame error rate (FER), and sometimes both BER and FER (as in the GSM system) to ensure proper receiver performance. In this way, you can determine a usable error rate on transmitted bits or frames. In many TDMA systems, a BER of more than 3% is not usable.

With analog communications, radio sensitivity generally depends on the gain and noise figure of the RF amplifier as well as the IF amplifier performance. In digital systems, the FEC systems and various decoding functions also will play an important role in overall radio performance. Thus, in a digital system, sensitivity not only is influenced by RF performance, but also by the performance of the digital decoding block.

Network modeling for today's high-technology digital air interfaces is complex because of factors such as receiver performance under operating conditions, complex handoffs, system acquisition performance and complex power control operation. Accurate, repeatable and documented testing of handset receivers can allow technicians to rule out the handset when problems are reported with the forward channel. The technician then can check for other possible causes, such as the terrain, installation, path-loss model, or the network hardware and/or software that controls handoff and power control systems.

Several receiver tests can be performed on CDMA/TDMA handsets, including adjacent channel desensitization, intermodulated spurious-response attenuation and protection against spurious-response interference. All of these are measures of a receiver's ability to perform correctly in the presence of various interfering signals. There also are tests of the mobile-assisted handoff system in TDMA and a number of pilot detection and handoff tests in CDMA. Although these tests are important, the primary concerns in the field are the absolute RF sensitivity and dynamic range as well as the ability to demodulate properly in the presence of operating conditions. This can be simulated by fading and additive white gaussian noise (AWGN).

TDMA Handsets In an IS-54/136 handset, the BER of a received forward-channel signal can be determined two ways. First, the handset can estimate the BER based on cyclic redundancy check and report a range back to a test set via a channel-quality message. However, a more accurate measurement of BER can be obtained by putting the handset in a mode in which the received forward channel signal is decoded, looped back, encoded and re-transmitted to the test set on the reverse channel. The test set then can perform a comparative analysis between the original data sent and the received data.

IS-137, the minimum performance standard for IS-136 handsets, calls for two specific sensitivity tests. The first is the lowest RF level that can be received in which the BER is less than 3%. This is the absolute sensitivity in a static (no fading or signal distortion) operating environment. A test set generates a signal with p/4 DQPSK modulation using a pseudo-random pattern for data. The BER then is measured using one of the above two methods, and the signal's RF level is reduced. The sensitivity then is the RF level at which the BER exceeds 3%.

A second sensitivity measurement involves the same procedure, only this time the signal will be subject to simulated flat Rayleigh fading. IS-137 calls for the sensitivity measurement at a simulated speed of 100km/h (62mph) and 8km/h (5mph).

CDMA Handsets CDMA technology is more complex, particularly when testing CDMA handset receivers.

There are many types of receiver tests for CDMA handsets, most of which are complex and often require expensive test systems. These tests generally need to simulate two base stations and then monitor the handset's ability to receive and react to different signals. Although these tests can be useful in high-level service and manufacturing applications, carriers generally cannot afford such elaborate and costly test equipment in the field. Also, the basic RF sensitivity/dynamic range and demodulation in AWGN tests will give the receiver a good workout, generally assuring proper operation of the receiver.

The CDMA receiver tests use FEC as the basis for determining what constitutes a usable level. The test setup is similar to the IS-54/136 system. A test set will generate a forward-channel signal using a pseudo-random data pattern. The handset is put into a loopback service mode, which will allow the test set to count the number of frames transmitted and the number of good frames at the base station. The test set must measure until it has 95% confidence in the measurement (in which the FER cannot exceed 0.005), using defined setup parameters for the signal at both the lower channel power and the high-er channel power.

Demodulation within AWGN is similar, except that a noise source (usually internal to a test set) is used to create a simulated operating environment. The test involves measuring all of the data rates and varying levels of AWGN Eb/Nt. Minimum standards vary for both rate set 1 (9,600kb/s vocoder) and 2 (14,400kb/s vocoder) and depend on which data rate is being tested and how much noise is added. FERs range from 0.003 at full rate and an AWGN Eb/Nt of 4.7 to 0.05 at full rate and an AWGN Eb/Nt of 3.6.

Although many tests can indicate proper operation of the receiver block in a digital handset, the most basic and necessary tests are RF sensitivity and dynamic range as well as tests that involve reception under simulated operating conditions such as fading and AWGN. These basic field tests give carriers the confidence that their customers' handsets can receive properly even in the most demanding environments.