I NTRODUCTION

In today’s market place, group delay specifications of Radio Frequency (RF) receivers are generally not tested in production due to the prohibitive nature that these tests exhibit. In addition, as these specifications are usually guaranteed by design, the tests of such parameters are generally not needed. In general, catastrophic defects in a LPF can be detected by simply testing the amplitude response; however, latent defects that cause a small change in the amplitude response can cause a larger effect in the group delay and hence need to be caught. Also, this goes a long way to improving the quality of devices in the field as over time a latent defect will become a catastrophic defect causing the entire system to fail. Therefore, in the future, as automobiles incorporate more RF receivers and as many of these devices will be going into autonomous vehicles, and as described in 2.8, all datasheet parameters of an automotive device have to be tested in production; therefore the need to test these group delay specifications in RF devices will become mandatory.

Comparing techniques for measuring group delay using conventional Vector Network Analysers (VNA’s), such as the Aeroflex 6480 Microwave System Analyser provide phase or group delay data relative to a golden device [25,26]. Other companies such as

ANALOGUE MIXED SIGNAL TEST DEVELOPMENT

88 Peter Sarson - December 2018

Keysight technologies have developed their own calibration standards for downconverters using three broadband standards, a power meter as a magnitude standard, a comb generator as a phase standard, and an S-parameter calibration kit [27, 28]. Other techniques such as using a transmitter with the same characteristics as the receiver under test have been used so that the same frequency range can be used with a VNA to measure the group delay directly [25,29]. There have been efforts made to measure the group delay of a system without having the system phase locked by using a comb generator [28], however this is not feasible at ATE as no such calibrated standard is available on an Automated Test Equipment (ATE) test system. There has also been a drive to use a two-tone signal where measuring the phase change between the two tones at both baseband and RF will indicate the group delay through the device [29]. The problem with using [29] is the problem of how to measure the phase of the RF input signal. At the end, the user is restricted to down converting the input by means of the testers RF measure, which results in the inability to calibrate a part of the signal path due to the weaknesses of the standard RF ATE tester infrastructure (Figure 70). In addition, using the two-tone option of an ATE would be very slow to sweep across the band. This is because only two tones are applied at the same time and thus to build a complete picture of a device then the two tones would need to be swept across the band that take much time due to the settling time of the generators. However, if the idea in [29] is expanded upon it, then a scheme for measuring and characterizing group delay in production could be developed. The fundamental idea in [29] is that the phase of the first frequency is not important to measure but the phase delta between two frequencies.

By modulating an RF source with a two-tone signal would give controllability over the phase would only require one generator and an Arbitrary Waveform Generator (AWG), where the Digital to Analog Converter (DAC) resides, to do the same as described in [29] but with less cost and more flexibility. However, a two-tone signal between a large bandwidth could have a large error. The technique in [29] can be further improved if a high-resolution chirp is used to modulate an RF generator. To use such standards and techniques [27-29] is dependent on the architecture of the test equipment available to an ATE test engineer and for the majority of the ATE testers available today; one possible solution would be to use a golden sample correlation method to work around the weaknesses of the standard RF ATE tester infrastructure.

One major benefit of an ATE test system is that all instruments are phase locked to the tester 10 MHz reference clock, Figure 71. The main items to note in Figure 71 are

the SMATE, which are RF Generators, SMA100A, which is used as the Local Oscillator (LO) for down-converting RF signals. The Digitizer (DIG), which is used to capture analogue Baseband (BB) signals or digitize a down-converted RF signal and the Arbitrary Waveform Generator (AWG), which is used to supply analogue signals to a Device Under Test (DUT) or to modulate an RF Source. There are two RF bricks has 4 port modules each with 2 ports, labelled A + B, per port module. As there are 4 RF generators, this allows the possibility to source 4 RF signals simultaneously.

Another feature of the RF Source is that it is possible to apply a modulation waveform to the IQ modulation inputs of the generator, Figure 92. For a limited band sweep, such as 25 MHz,

High Speed AWG

RF Source RF Measure

High Speed Digitizer RF TX

BB Loadboard Loopback

Path RF Loadboard

Loopback Path

Figure 69 – RF tester setup showing the weakness with calibrating the full RF path whilst testing RF transmitters.

High Speed AWG

RF Source RF Measure

High Speed Digitizer RF RX

BB Loadboard Loopback

Path RF Loadboard

Loopback Path

Figure 70 – RF tester setup showing the weakness with calibrating the full RF path whilst testing RF receivers.

ANALOGUE MIXED SIGNAL TEST DEVELOPMENT

90 Peter Sarson - December 2018

Figure 71 –Standard RF Tester Overview

RF BPF RF LNA IF LPF

LO

BB (OUT) RF+BB

(IN)

Figure 72 –RF RX Block Diagram

it would be possible to chirp [17,20,29,30] the RF source frequency hence sweeping the frequency of the generator in a fast and efficient manner compared to using a two-tone setup [28].

By utilizing the benefits of the test system and using a proven technique from the laboratory, it would be possible to produce a reliable and equivalent test on ATE. This would result in high confidence of the end quality that is the starting requirement of the shipped devices to end customers in the automotive industry for use in autonomous vehicles where a slight error in a communication could result in a fatality.

One of the features of a chirp is its distortion properties. The distortion seen at the carrier frequency, 1, is too far out of band to interfere with the signal integrity, therefore, distortion properties of a RF amplifier can be ignored.

In this chapter, we will discuss how to implement a Group Delay test on a Very High Frequency (VHF) or Ultra High Frequency (UHF) RF Receiver, Figure 72, using todays RF ATE tester installed base. Figure 72 shows a simplified block diagram of a RF Receiver. The device consists of a RF bandpass filter (BPF), RF Low Noise Amplifier

(LNA), Mixer and a BB, Low Pass Filter (LPF). The main two direct contributors that could cause a group delay failure of the system described would be due to the LPF and BPF. As these components primary function is to alter the phase at specific frequencies, if there is a defect that changes this characteristic, it could be possible that a group delay failure could go undetected by testing only the magnitude response. However, we will also show, through simulation, that if the non-linearity components of the mixer are not tested that this could have a tremendous impact on the measured group delay due to distortion of the signal.

This chapter is organized in the following way, Chapter 4.2 will describe the basics of mixer theory, and how a chirp as described in [29, 30] can be used to modulate an RF generator to be able to sweep a RF downconverters input. Chapter 3.3.3 will discuss how to generate a discrete linear chirp; Chapter 4.3 will describe the effects of a chirp on a non-linear system, Chapter 4.4 will describe the effect of a defect on a filters performance and how the chirp technique can detect the fault. Chapter 4.5 will describe the architecture of a standard RF tester, and the issues pertaining to making an RF group delay test of an RF downconverter. Chapter 4.6 will detail some measurements using the RF tester in loopback to verify the technique and then conclusions will be drawn.