WO2000005591A1

WO2000005591A1 - Method of and apparatus for generating a multitone test signal

Info

Publication number: WO2000005591A1
Application number: PCT/GB1999/002351
Authority: WO
Inventors: Harold Thomas Brown; Neil Edwin Thomas
Original assignee: Ifr Limited
Priority date: 1998-07-21
Filing date: 1999-07-20
Publication date: 2000-02-03
Also published as: GB9815889D0; GB2339917A

Abstract

The invention consists in an apparatus for and a method of generating a test signal (T'), the method comprising the steps of generating at least first and second main signals (f1, f2), the first main signal having a frequency component at a first predetermined frequency, and the second main signal having a frequency component at a second predetermined frequency, combining said main signals to form a composite test signal (T') containing one or more intermodulation components, characterised in that the method further comprises the step of introducing at least one cancelling signal (2f1-f2, 2f2-f1) at a frequency substantially equal to that of one of the intermodulation components so as to suppress that intermodulation component in the composite signal. The cancelling signal may be combined with a main signal or the test signal, or may be introduced by modulating one of the main signals. The modulating signal (f2-f1) may be a composite modulating signal so as to generate two or more cancelling signals. Also, each of two or more main signals may be modulated by a respective modulating signal, or two main signals may be modulated by a modulating signal equal to the difference in frequency between them.

Description

METHOD OF AND APPARATUS FOR GENERATING A

MULTITONE TEST SIGNAL

This invention relates to a method of and apparatus for generating a test signal. The test signal may be used for measuring the linearity of a system.

When testing the linearity of a system, a known method involves sending a test signal into a

device under test (DUT) and observing the output spectrum on a spectrum analyser. A

commonly used test signal consists of two closely spaced tones at frequencies ft and f₂, although test signals containing more tones are also common. Non-linearities in the DUT mix these input tones to create extra tones which can be detected by the spectrum analyser.

The levels or powers of these extra tones may be used to characterise the degree of

non-linearity in the DUT. Typically, for a two tone test signal the largest of these extra tones are the third order products at frequencies of 2fι-f₂ and 2f₂-fι. Higher order products are also generated but the magnitude of these tones tends to be much smaller than the magnitude of

the third order products.

Accurate measurements of non-linearities in the DUT can be made by using a clean input test signal together with an accurate spectrum analyser. If the input test signal entering the DUT

contains intermodulation distortion products then the dynamic range and/or accuracy of non-linearity measurements will be affected. For a two tone input signal, the individual tones

are generated respectively in independent signal sources and are subsequently combined to

form a composite signal. However, the process of combining the individual tones generally

produces unwanted intermodulation products in the composite signal in addition to the two required tones. Figures 1 to 3 illustrate three known methods for combining tone signals generated in

independent signal sources Si and S₂. Intermodulation products are formed in the composite

test signal T when the two tone signals, fi and f₂, are present simultaneously in the output of either signal source. Each method attempts to prevent the formation of intermodulation

products, by minimising both the amount of f₂ signal that appears in the output of the fi signal

source and the amount of fi signal that appears in the output of the f₂ signal source.

The first method shown in Figure 1 is applicable only to narrow band systems and uses filters to isolate the two sources. This method is commonly used where several basestation

frequencies share a common antenna. The rationale behind this is twofold - as well as

reducing output intermodulation, it also reduces the power loss that occurs with other types of combiners. This is the combination method of choice for PIM (passive intermodulation)

measurement systems, which operate at fixed frequency.

The second method shown in Figure 2 is for broad or medium band systems. An isolating

combiner is made using either a Wilkinson or a 3dB hybrid combiner for very high frequencies, or a transformer based combiner going down to or near DC. The bandwidth of

the latter depends on the ferrites used for the transformer core, but 10MHz to 2GHz are

achievable.

The third method shown in Figure 3 uses only resistive combiners and is broadband but has

the lowest performance of all the methods. The best resistive combiner arrangement is

shown, giving the maximum source to source isolation for any given signal isolation. The aim of all of these known methods is to prevent creation of intermodulation products in the desired composite test signal by effectively buffering each signal source.

According to a first aspect of the present invention there is provided a method of generating a

test signal, the method comprising the steps of generating at least first and second main

signals, the first main signal having a frequency component at a first predetermined frequency

and the second main signal having a frequency component at a second predetermined

frequency, combining said main signals to form a composite test signal containing one or

more intermodulation components, wherein the method comprises the step of introducing at least one cancelling signal at a frequency substantially equal to that of one of the

intermodulation components so as to suppress that intermodulation component in the composite signal.

According to a second aspect of the present invention there is provided an apparatus for generating a test signal, the apparatus comprising signal generator means for generating at

least first and second main signals, the first main signal having a frequency component at a

first predetermined frequency and the second main signal having a frequency component at a

second predetermined frequency, and means for combining said main signals to form a composite test signal containing one or more intermodulation components, wherein the

apparatus further comprises means for introducing at least one cancelling signal at a

frequency substantially equal to that of one of the intermodulation components so as to

suppress that intermodulation component in the composite signal.

A method of or an apparatus for generating a test signal in accordance with the first or second

aspect of the invention has an advantage that it can generate a test signal with reduced intermodulation distortion, which in turn makes the test signal cleaner. The method in

accordance with the invention is thus able to generate a test signal which may be used in test systems with improved performance.

Ideally, the test signal is used for measuring the linearity of a system. Suitably, the first main

signal is generated in a first signal path, the second main signal is generated in a second

signal path, and the composite signal contains frequency components at the first and second predetermined frequencies. Preferably, the cancelling signal is introduced at a suitable

amplitude and phase so as to reduce or in some cases cancel one of the intermodulation

components in the composite signal.

In one embodiment of the invention, the cancelling signal is introduced into the main signals before they are combined to form the composite test. In an alternative embodiment, the cancelling signal is introduced into the composite signal. In a further embodiment, the

cancelling is introduced into the main signals in the combining step to produce the composite signal.

In a preferred embodiment of the invention, the step of introducing the cancelling signal

comprises modulating a main signal with a modulating signal. The step of modulating the main signal may also generate a sideband signal at substantially the frequency of another

main signal.

Preferably, the modulating signal has a frequency substantially equal to the difference

between the first and second predetermined frequencies. Suitably, the intermodulation components are third-order intermodulation components.

The modulation of the main signal may be achieved using amplitude modulation (AM). In an

alternative embodiment, modulation of the main signal is achieved using angle modulation

which may involve frequency or phase modulation techniques. Typically, the angle

modulation is low-index angle modulation.

Preferably, the method comprises a calibration step in which the composite test signal is

analysed to determine the residual signal levels at the frequency of the intermodulation

components, and the amplitude and phase of the additional frequency component is adjusted in dependence on the determined residual signal levels. Preferably, the calibration step optimises suppression of the intermodulation components in the composite test signal.

Further features and advantages of the invention will be apparent from the description below.

Embodiments of the invention will now be described, by way of example, with reference to

the accompanying drawings, in which:

Figure 1 is a block diagram of a known system for producing a test signal T from two signal

sources Si and S₂;

Figure 2 is a block diagram of alternative known system for producing a test signal T from

two signal sources Si and S₂;

Figure 3 is a block diagram of another known system for producing a test signal T from two

signal sources Si and S₂; Figure 4 is a block diagram of a system for producing a test signal T' in accordance with the

invention incorporating two modulators;

Figure 5 is a block diagram of a system for producing a test signal in accordance with the

invention incorporating two amplitude modulators and two phase angle modulators;

Figure 6 is a block diagram of a system for producing a test signal T' in accordance with the

invention incorporating two IQ modulators;

Figure 7 is a diagram illustrating the process by which a test signal T is produced in the

known systems shown in Figures 1 to 3, and the process by which a test signal T' is produced

in the modulating systems shown in Figures 4 to 6;

Figure 8 is a block diagram of an alternative system for producing a test signal T' in

accordance with the invention incorporating two frequency synthesisers;

Figure 9 is a block diagram iof a system for producing a test signal T' in accordance with the

invention incorporating a composite modulation of one of the tones; and

Figure 10 is a diagram illustrating the process by which a test signal T' is produced in the

system of Figure 9.

Referring to Figure 4 there is shown a block diagram of a system 40 for producing a relatively

clean two-tone test signal T. The system shares some of the basic elements of the prior art systems shown in Figure 1 to 3 such as the two independent signal generators Si, S₂ and a combiner 41.

The signal generator Si generates a tone signal at a frequency fi which is modulated by a

modulator 42. The resultant modulated signal is fed to a first input of the combiner 41. In a

similar way, the signal generator S₂ generates a tone signal of similar amplitude and at a

frequency f₂ slightly higher than frequency fi. This tone signal is modulated by a second modulator 43 and the resultant modulated signal is fed to a second input of the combiner 41.

The combiner 41 adds the two modulated signals together to produce a test signal T'

containing frequency components at the frequencies fi and f₂. This test signal T' may be used to test the linearity of a system or a device under test.

The combiner 41 is a standard component used in the art and may, for example, be equivalent

to one of the combiners used in the prior art systems shown in Figures 1 to 3.

A modulating signal which is supplied to the modulators 42, 43 is generated in an oscillator

44. The oscillator 44 is adjusted to generate a tone signal at a frequency equal to the difference in frequency of the tone signals generated by the signal generators Si and S₂ ie. f₂ -

fi. The modulating signal is supplied along a first signal path via a variable phase adjuster 45 and a variable attenuator 46 to an input of the modulator 42. The modulating signal is also

supplied along a second signal path via a variable phase adjuster 47 and a variable attenuator

48 to an input of the modulator 43.

Referring to Figure 7, there is shown a frequency domain plot i) representing the tone signals at fi and f₂ generated by the signal generators Si and S₂ in the prior art systems of Figures 1 to 3 and in the system 40 shown in Figure 4. The tone signals are shown as vertical peaks

whose height is proportional to the magnitude of the tone signal. The prior art method for

combining the tone signals generated in the independent signal sources Si and S₂ is symbolised by the arrow C. The resultant test signal T is shown in the frequency domain plot

ii) and contains in addition to the frequency components fi and f₂, third order intermodulation

products at the frequencies 2f f₂ and 2f₂ - fi. Contamination of the two-tone test signal T

with these intermodulation products is undesirable and can result in a degradation of

performance when used in a linearity test system.

The process shown in Figure 4 for producing a two-tone test signal is represented by the

arrows M and C in Figure 7. The arrow M symbolises the modulation stage and the arrow C symbolises the combination stage. Referring to Figures 4 and 7, modulation of the tone signal fi by the modulating signal f₂-fι in the modulator 42 generates a modulated signal

containing additional frequency components at the frequencies fi ± (frfi) ie at 2fι-f₂ and f₂. Modulation of the tone signal f₂ by the modulating signal f₂-fι in the modulator 43 generates a modulated signal containing additional frequency components at the frequencies f₂ ± (f₂ - fi)

ie at fi and 2f₂-fι. The modulated signals are represented in the frequency domain plot iii).

Combination C of these modulated signals in the combiner 41 produces a test signal T

represented by the frequency domain plot iv). Third order intermodulation products which would normally be generated in the test signal T' during the combination step are cancelled by

the additional frequency components 2f f₂ and 2f₂-fι in the modulated signals.

Cancellation of the intermodulation products requires the additional frequency components in

the modulated signals to be set to an appropriate phase and amplitude. This is achieved by

independent adjustment of phase and amplitude of the modulating signals using respectively the variable phase adjusters 45, 47 and the variable attenuators 46, 48. Optimal adjustment of the modulating signals produces a test signal T' with two tones at fi and f₂, and suppressed

frequency components at the intermodulation frequencies.

Optimal adjustment or calibration of the test signal device of Figure 4 is achieved in practice by sending the two-tone test signal T' directly to a signal analyser. The analyser measures the

levels of the frequency components at the third order intermodulation frequencies and either displays the results for manual adjustment of the test signal device or automatically

implements the test signal device adjustments on the basis of a feedback algorithm. As the third order intermodulation frequencies comprise an upper and a lower sideband, each with an associated amplitude and phase, 4 parameters are required to be adjusted in the test signal device.

A basic technique for calibration relies on independent amplitude measurements of the upper

and lower side bands and performs iterative adjustments of the two modulating signals in

phase and amplitude by means of the variable phase adjusters 45, 47 and the variable attenuators 46, 48. Calibration is achieved when amplitude nulls are detected for the upper

and lower sidebands corresponding to the intermodulation frequencies. According to one

algorithm, each parameter (amplitude or phase) of the modulating signals is systematically

varied until a minimum is observed for signal components in the corresponding upper or lower sideband. This systematic variation is then repeated to determine a minimum for the

signal components in the upper or lower sideband with respect to both adjustment parameters

(amplitude and phase). In another embodiment, the signal analyser measures both the amplitude and phase of the

upper and lower side bands and calculates a single adjustment for each parameter in the test

signal device to achieve optimum cancellation of the intermodulation products.

The modulators 42, 43 in the system 40 may be amplitude modulators or angle modulators.

Amplitude modulation will form modulated signals containing the two sideband signals

shown in plot iii) in Figure 7 together with the main signal at fi or f₂. For angle modulation, however, low index modulation techniques have to be used to form a modulated signal containing only the two sideband signals shown in plot iii) in Figure 7 together with the main

signal at f i or f₂. Low index modulation, requires the magnitude of the angle variation of the

main signal fi or f₂ to be small. This occurs when the modulating signal is at a low level or when the modulation factor of the modulator is small. Low index modulation generates a modulated signal in which the sideband signals are at a relatively low level compared to the

main signals. This restriction is acceptable in the test signal device of Figure 4 as the intermodulation products which require cancellation are also at a relatively low level.

In an alternative embodiment of the invention, the signal source S I and modulator 42 of Figure 4 can be replaced by a first frequency modulated source that generates a modulated

signal with the frequency components fi, f₂ and 2fι-f₂, and the signal source S2 and modulator

43 can be replaced by a second frequency modulated source that generates a modulated signal

with the frequency component fi, f and 2f₂-fι, thus producing the same result as the

embodiment of Figure 4.

Referring to Figure 5, there is shown a block diagram of a system 50 in accordance with the

invention for producing a relatively clean two-tone test signal T. In this embodiment, the modulating signal previously generated by the oscillator 44 in Figure 4 is now generated by

mixing the signals fi and f₂ from the signal sources Si and S₂ in a mixer 54. The phase of the

modulating signal generated by the mixer 54 is not directly controllable. Therefore, the

arrangement of the two modulators 42, 43 in Figure 4 has been replaced by an arrangement of

two amplitude modulators 52, 53 and two angle modulators 92, 93. Each modulator 52, 53,

92, 93 is fed with the modulating signal from the mixer 54 via a respective variable attenuator

55, 56, 57, 58. The two variable attenuators 55, 57 together control the amplitude and phase

of the lower sideband frequency component 2f f₂ whilst the two variable attenuators 56, 58 together control the amplitude and phase of the higher sideband frequency component 2f₂-fι.

The process of calibrating the test signal device is equivalent to calibration process described with reference to Figure 4.

Referring to Figure 6, there is shown a block diagram of another system 60 in accordance with the invention for producing a relatively clean two-tone test signal T'. In this embodiment, the amplitude and angle modulators 52, 92 shown in Figure 5 have been

replaced by an IQ modulator 62, and the amplitude and angle modulators 53, 93 have

similarly been replaced by an IQ modulator 63. Again the two variable attenuators 55, 57 together control the amplitude and phase of the lower sideband frequency component 2fι-f₂

whilst the two variable attenuators 56, 58 together control the amplitude and phase of the

higher sideband frequency component 2f₂-fι.

Referring to Figure 8, there is shown a block diagram of a system 80 in accordance with the

invention for producing a relatively clean two-tone test signal T. In this embodiment, the

sideband components 2f f₂ and 2f₂-fι are produced respectively by two synthesisers 81 , 82 and are added to the main signals ft and f₂ by the adders 83, 84. The output of synthesiser 81 is fed to the adder 83 via a variable phase adjuster 86 and a variable attenuator 88, and the

output of the synthesiser 82 is fed to the adder 84 via a variable phase adjuster 87 and a

variable attenuator 89. The variable phase adjusters 86, 87 and variable attenuators 88, 89 are

adjusted as in the previous embodiments to provide optimum cancellation of intermodulation

products produced during the step of combining the signals fi and f₂.

It will be evident in view of the foregoing description that various modifications may be made

within the scope of the present invention. For example, the two-tone test signal may be a

three, four or multiple tone test signal. Also, although the embodiments have been described with reference to discrete components, it is equally applicable for the components to be

embodied as functional elements of a digital signal processing (DSP) chip or as programmed functions in a programmable device. In the DSP implementation the adders 83 and 84 and

the combiner 41 may be integrated as one functional element

Figure 9 shows a block diagram of a system in which two signal generators S and S₂ generate

signals fi and f₂, respectively, one of these signals F, being modulated in a modulator 90 before both signals are combined in a combiner 91 to produce a two-tone test signal T'. The

modulating signal comprises two modulating frequency components f_s and 2f_s combined in a combiner 92. The effect of the modulator of the signal fi is illustrated in Figure 10, which

shows the resulting two sets of sidebands at f_sι-f_s and 2f_s]-2f_s. The modulating signal f_s is

selected to be equal frequency to the difference between the frequencies Fi and f₂ so that the

intermodulation components of Fi and f₂ are substantially cancelled by the sidebands -f_s, as

shown by broken lines in Figure 10. The lower sideband component -2f_s remains in the composite test signal T\ but does not cause a problem because it is removed from the two

tones Fi and F₂.

Claims

1. A method of generating a test signal, the method comprising the steps of generating at

first predetermined frequency, and the second main signal having a frequency component at a

second predetermined frequency, combining said main signals to form a composite test signal containing one or more intermodulation components, characterised in that the method further comprises the step of introducing at least one cancelling signal at a frequency substantially

equal to that of one of the intermodulation components so as to suppress that intermodulation component in the composite signal.

2. A method as claimed in claim 1, wherein the test signal is used for measuring the linearity of a system.

3. A method as claimed in claim 1 or claim 2, wherein each main signal is generated in a

respective signal path, and the composite test signal contains at least frequency components of the first and second predetermined frequencies.

4. A method as claimed in any one of the preceding claims, wherein each cancelling

signal is introduced at a suitable amplitude and phase so as to reduce or cancel a respective

one of the intermodulation components in the composite signal.

5. A method as claimed in any one of the preceding claims, wherein the step of

introducing each cancelling signal comprises generating a signal at a frequency substantially equal to that of one of the intermodulation components, and combining this signal with at least one of the main signals.

6. A method as claimed in claim 5 wherein at least one cancelling signal is combined

with each of two or more of the main signals to suppress a respective intermodulation component.

7. A method as claimed in any one of claims 1 to 4, wherein the step of introducing each cancelling signal comprises generating a signal at a frequency substantially equal to that of one of the intermodulation components, and combining this with the composite test signal.

8. A method as claimed in any one of claims 1 to 4, wherein each cancelling signal is

introduced by modulating one of said main signals with a modulating signal.

9. A method as claimed in claim 8, wherein one of said main signals is modulated with a composite modulating signal containing at least two frequency components so as to generate

two or more cancelling signals

10. A method as claimed claim 8 wherein each of two or more of said main signals is modulated with a respective modulating signal to generate one or more cancelling signals.

11. A method as claimed in any one of claims 8 to 10, wherein the step of modulating one

main signal generates a cancelling signal and an additional sideband signal at substantially the

frequency of another of said main signals.

12. A method as claimed in any of claims 8 to 11 wherein each of said first and second

main signals is modulated by a modulating signal substantially equal to the difference in frequency between them.

13. A method as claimed in any one of the preceding claims, wherein the intermodulation

components comprise third-order intermodulation components.

14. A method as claimed in any one of claims 8 to 12, wherein modulation of the main signal is achieved using amplitude modulation (AM).

15. A method as claimed in any one of claims 8 to 12, wherein modulation of the main signal is achieved using angle modulation.

16. A method as claimed in claim 14, wherein the angle modulation is achieved through

frequency modulation.

17. A method as claimed in any one of claims 8 to 12, wherein modulation of the main

signal is achieved using IQ modulation.

18. An apparatus for generating a test signal, the apparatus comprising signal generator

means for generating at least first and second main signals, the first main signal having a

frequency component at a first predetermined frequency, and the second main signal having a frequency component at a second predetermined frequency, and means for combining said

main signals to form a composite test signal containing one or more intermodulation

components, wherein the apparatus further comprises means for introducing at least one cancelling signal at a frequency substantially equal to that of one of the intermodulation

components so as to suppress that intermodulation component in the composite signal.

19. An apparatus for generating a test signal substantially as herein described with

reference to any of Figures 4 to 8.