JPS5831063B2 - Fault detection method for multiple signal converter - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a single sideband frequency division multiplexed signal (hereinafter referred to as SSB).
- Called FDM signal.

) as a pulse code modulated time division multiplexed signal (hereinafter referred to as POM-T
It is called a DM signal.

The present invention relates to a failure detection method in a multiplex signal converter that converts into (), particularly a multiplex signal converter using a fast Fourier transform (or fast inverse Fourier transform) circuit.

Fast Fourier Transform
ns-form: FFT) and fast inverse Fourier transform (
In-verse Fast Fourier Tra
ns form: IFFT) method is J,
It has long been devised by W. Cooley and J.W. Tuckey (1965).

Since this invention, calculation time has been significantly reduced in the field of numerical analysis, such as calculation of frequency spectra and correlation functions, which previously required long calculations even when using an electronic computer.

Later, along with advances in integrated circuits, calculation time was further reduced significantly, resulting in FFT and IFFT (hereinafter simply FFT).
It's called FT.

) has been put to practical use not only in the field of numerical analysis but also in the field of digital filters, which was conventionally considered impossible to process in real time.

In recent years, research has been conducted to apply FFT to multiplex signal conversion devices (for example, ``Digital TDM-
"Study on the system configuration of FDM converter (trans multiplexer)" Institute of Electronics and Communication Engineers Communication System Study Group Material C37
5-190).

The intention of using digital processing such as FFT in multiplex signal conversion equipment is that it has many advantages such as reduction in equipment scale through integration, simplification, ease of maintenance, improved transmission quality, and economic efficiency. It is.

When configuring a multiplex signal conversion device for converting an SSB-FDM signal into a PCM-TDM signal using this FFT, the present invention provides a method for combining an SSB-FDM signal before conversion and a PCM-TDM signal after conversion. The present invention focuses on the occurrence of special conditions in between, and provides a fault detection method that can detect faults in the multiplex signal converter based on an extremely simple principle.

Hereinafter, the present invention will be described in detail with reference to the drawings.

In order to facilitate understanding of the present invention, FIG. 1 shows block interception of a multiplex signal conversion device using FFT.

In Figure 1, 1 is an input signal, analog SSB-
The FDM signal is converted into a high-speed serial digital signal, and the next stage polyphase circuit 2 and the following
An analog-to-digital converter circuit 2 has a function of converting the high-speed serial signal to a parallel signal so that the FFT circuit 3 can operate at a low speed.
This is a polyphase circuit consisting of a plurality of digital filters for converting the T circuit 3 into a signal suitable for the human input signal format.

3 is an FFT circuit that performs fast Fourier transform or fast inverse Fourier transform; 4 is a parallel circuit that is the output of FFT circuit 3;
This is a parallel/serial signal conversion circuit for converting a digital signal into a serial signal having a PCM-TDM signal format, and is generally used in a PCM encoder.

That is, with the above configuration, the SSB/FDM signal is
The frequency arrangement of the normal SSB-FDM signal is shown in FIG. 2, which is converted into a CM-TDM signal.

In Figure 2, fL is the lower limit frequency of the SSB-FDM signal, and fH is the SSB
- the upper limit frequency of the FDM signal, B is the band of the SSB-FDM signal, f is the band of each channel signal (i.e. audio frequency band, typically 4 KHz), f is the sampling frequency of the SSB-FDM signal, n is the number of multiplexed channels represents.

That is, the SSB-FDM signal is within the audio frequency band (
It is obtained by frequency division multiplexing the signals of each channel, which normally exist at f = 4 KHz, with each channel shifted by a different frequency.

Such a signal is converted into PCM/TD by the converter shown in Fig.
The PCM-TDM signal is obtained by time-division multiplexing the signals of each channel within the audio frequency band after sampling them at a sampling frequency of f'8 (usually 8 KHz-2f). It is something.

That is, the sampling frequency at each output terminal of the FFT circuit 3 is f'.

In that case, the size N (rows and columns) of the FFT needs to satisfy the conditions described below.

In other words, when sampling the SSB-FDM signal at f5, in order to prevent the sampled frequency spectra from overlapping each other, fH≦f, ≦2fL (1) f > 2B = 2nf = nf' (2) S-
-OS or f5≧2fH (3) It is necessary to satisfy either of the following conditions.

In this case, it is known that the scale N of FFT is given by the following equation.

(Refer to the above-mentioned Institute of Electronics and Communication Engineers materials). N=f/f
(4) S From these relational expressions, fH/f'8≦N≦2fL/f18 (5) N≧n (6) or N≧2fH/f'8 (7) can be obtained.

That is, in order to satisfy the conditions (1) and (2), it can be seen from (6) that N must be larger than n.

In addition, in order to satisfy the condition (3), fH≧B(8
), so (7) is N≧2fH/f′, ≧2B/f′, −2nf(y/f′
, -n(9), and again, N must be greater than n.

That is, when the present invention configures a multiplex signal conversion device using FFT, the number of channels of the multiplex signal n is as described above.
This method focuses on the fact that there is a difference between the FFT scale N and the FFT scale N, and uses this difference to detect a failure.

Note that, as is clear from the above equations, N=n also holds true, but this case is excluded from the present invention because it is unrealistic for the reasons described below.

That is, when the conditions of equations (1) and (2) are satisfied, the frequency array of the SSB-FDM signal is fL-f8/2. as shown in FIG. fH-f8, and when the condition of equation (3) is satisfied, it becomes fL-0, as shown in Fig. 4.
fH-f8/2, and in either case, the arrangement is fixed and generality is lost.

Furthermore, (1) As far as the existing hierarchical structure (hierarchy, see Table 1) is concerned, the SSB
There is no FDM signal.

(ii) When the SSB-FDM signal is manually input to a multiplex signal converter via a transmission line, an analog bandpass filter is installed before sampling because it contains unnecessary components such as noise outside the band. , N=n, the bandpass filter must have steep characteristics that allow only frequency components from fL to fH to pass, and it is difficult to realize such a filter.

If the SSB contains frequency components other than fL to fH,
-F When a DM signal is sampled, its noise will be superimposed on the information signal.

On the other hand, if N is made larger than n, it becomes possible to soften the characteristics of the bandpass filter, making it easier to realize the filter.

Song) In order to use FFT as this kind of multiple signal conversion device, it is necessary that N can be decomposed into prime factors, and N is n
Become bigger.

FFT processing is easiest when N is a power of two.

Next, each analog signal Xo tX constituting the FDM signal
t""...XN t, and each digital signal X'o, X'1... appearing at the output end of the FFT circuit 3 of scale N
For example, the reaction with x'N is as shown in Figure 5,
Xo, Xl and X'1. ...XN-1 and X'N-
1 contain the same information.

That is, as shown in FIG. 5a, the FDM signal occupying the band from f/2 to f5 is divided into N bands for each fo, and the signal is separated and generated at the output end of the FFT circuit 3 shown in FIG. 5b. It turns out.

Therefore, an empty channel containing no information is generated at the output end of the FFT circuit 3 by the difference between the number of multiplexed channels n and the FFT scale N, and an arbitrary analog pilot signal is added to a specific band of the FDM signal. Then, a digital signal having a specific code pattern corresponding to the pilot signal can be generated in the empty channel.

Note that there are various hierarchical structures (hierarchies) for multiplexed signals in analog and digital lines, so there are various possible conversion levels for multiplexed signals, but the existing hierarchy in Japan is shown in Table 1. .

Based on Table 1, for example, if we consider converting a 2XG SSB/FDM signal to a 1.5M PCM-TDM signal, the lowest frequency f that satisfies equations (1) and (2) is 112KHz. Since f/ is 8 KE(z, N is 14 from equation (4).

The actual block configuration of the multiplex signal conversion apparatus in this case uses two sets of FFT circuits 7 and 7', as shown in FIG.

However, this is true for SSB-F of the channel arrangement shown in Fig. 7a.
The DM signal is input, and the digital signal having the frequency spectrum shown in FIG. 7b is sent to the output terminal of the FF and T circuit 7.
It is assumed that the value is extracted to ~0.

In this case, two empty channels will be generated at the output ends of both FFT circuits.

Therefore, for example, the analog-to-digital conversion circuit 5
, for the polyphase circuit 6 and FFT circuit 7 system, the SSB-FDM signal on the input side is applied to the analog-to-digital conversion circuit 5 so that a digital signal for fault monitoring appears on the empty channel at the output end of the FFT circuit 7. By adding a fault detection analog signal to a specific band, which will be described later, it is possible to easily detect that a fault has occurred in the system.

Examples of the present invention are shown in FIGS. 8 and 9.

First, FIG. 8 shows that the failure of the device is detected at the output terminal [phase] of the FFT circuit 7
This is attempted to be detected by monitoring the code pattern of the digital signal.

Generally, a sine wave is used as a pilot signal, but in this case, the sampling frequency f (112 KHz) of the analog-to-digital conversion circuit 5 is divided and FFT is performed.
The frequency of the pilot signal at output terminal 0 of circuit 7 is f
'(8KHz), a delay circuit 10 that applies an appropriate delay to the output of the frequency divider 9, and a discrete output signal from the delay circuit 10. A filter 1 that extracts an appropriate component from the frequency acid component to produce an analog signal, and a synthesis circuit 12 that superimposes the output of the filter 1 on the input SSB-FDM signal are required at the stage before the analog-to-digital conversion circuit 5.

Further, an FFT circuit 7 is provided at the subsequent stage of the output terminal 0 of the FFT circuit 7.
A code pattern detection circuit 13 is required to monitor the code pattern of the pilot signal.

In this example, the constituent countries of the FDM signal are as shown in FIG.

That is, in FIG. 10, for example, the shaded areas Xo and X1
3 is a frequency band to which a pilot signal can be added.

Therefore, when an analog pilot signal is added to Xo or In some cases, it is possible to detect that a failure has occurred in the device.

In the above, the pilot signal is X in the frequency band between f8/2 and fs.

Alternatively, although the pilot signal is added to X03, it is also possible to add the pilot signal to other frequency bands.

This will be explained below using FIG. 11.

That is, as shown in FIG.
When the signal is sampled at f,=112 KHz, it becomes b.

However, 60KHz to 108KHz shown in a.
Even if there is no SSB-FDM signal in the band of
, B, C, or D, its frequency spectrum will similarly be b when it is sampled.

Therefore, if one of the signals indicated by the solid line a is used as a pilot signal, that pilot signal will appear at the position indicated by the solid line after sampling.

This also applies to the pilot shown by the dotted line.

That is, the frequency band to which the analog signal as a pilot signal is actually added is X in FIG.

, X13, is not necessary.

For example, if you want to generate a 2KH2 pilot signal at output terminal 0 of the FFT circuit 7, the analog sine wave to be superimposed on the SSB/FDM signal is as shown in FIG.
110,114,=・”)IG(z becomes possible.

In addition, if you want to generate it at the output terminal ■, (54, 58
, 166, 170...) becomes a KHz sine wave.

2K having a certain positional relationship with the sampling frequency f
When superimposing a Hz analog signal, the frequency dividing circuit 9 in FIG. 8 sets the sampling frequency f.

(112KHz) may be divided into 1156.

The sampling frequency f s (112 KHz) of the analog-to-digital conversion circuit 5 is shown in FIG. 12a, and the signal obtained by dividing it is shown in FIG. 12b.

A signal obtained by applying an appropriate delay to this signal is shown in FIG. 12c.

As will be explained later, by changing this delay amount,
It becomes possible to fix the code pattern of the output terminal 0 of the FFT circuit 7 to several types of patterns.

The signal in FIG. 12C is further passed through a filter 11 to obtain an analog sine wave of 2KE-(z (FIG. 12d)
This signal is superimposed on the SSB/FDM signal by the combining circuit 12.

Furthermore, by adjusting the amount of delay using the delay circuit 10, it is possible to make the code pattern at the output terminal 0 of the FFT circuit 7 a specific pattern.
It is possible to do as shown below.

Although FIG. 126 shows a digital PCM signal expressed in level values, in reality, a binary code pattern expressed as 0 or 1 is generated.

If we ignore the polarity and only consider the level value, then A and Mouth will have one type of level value, and Ha will have two types of level value.

That is, if a level value other than these occurs, it can be detected that an abnormality has occurred in the device.

Next, in FIG. 9, a digital-to-analog conversion circuit 15 is provided at the output terminal 0 of the FFT circuit 7, and a failure in the device is detected by monitoring the analog signal.

In this case, for example, a 2 KHz sine wave (FIG. 12f) is generated from the signal generator 14, and the S
When superimposed on the SB-FDM signal, the same signal (FIG. 12g) will be obtained from the output terminal 0 of the FFT circuit 7 via the digital-to-analog conversion circuit 15 and the signal detection circuit 16.

Also, as mentioned above, 110KHz, 114KH
Even if we generate a sine wave of z,...
A 2KHz sine wave will appear.

In addition, if you try to generate it at the output terminal ■, 54KHz, 5
It becomes a sine wave of 8KHz, 166IG(z, 170KHz-).

As is clear from the above explanation, in a multiplex signal conversion device using FFT, there are as many empty channels that do not contain channel information at the output end of the FFT circuit as the number of channels equal to the difference between the number of multiplexed channels n and the scale N of FFT. will occur.

Therefore, an analog pilot signal is added to a band outside the required band (, ff, -fH) of the SSB-FDM signal corresponding to this empty channel, and the code pattern of this pilot signal that appears at the output end of the FFT circuit after conversion is The monitor can detect failures in the multiplex signal converter.

As described above in detail, the present invention provides S
In a multiplex signal converter that converts an SB-FDM signal into a PCM-TDM signal, an analog signal is added as a pilot signal to a specific band outside the SSB/FDM signal band, and after conversion, a specific signal is added to the output end of the FFT circuit. The pilot signal, which is generated as a digital signal having a code pattern, is monitored to detect a failure in the multiplex signal converter.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a multiplex signal conversion device to which the present invention is applied; FIGS. 2, 3, and 4 are diagrams showing channel arrangement examples of SSB-FDM signals that are manually operated by the device in FIG. 1; FIG. 5a , b is a frequency array diagram and a block diagram for explaining the correspondence between the input SSB/FDM signal and the output of the FFT circuit in the device of FIG. 1, and FIG. 6 is a diagram of a multiplex signal conversion device to which the present invention is applied A block diagram showing an example, FIG. 7a,
b is a frequency array diagram for explaining the human power and output of the device in FIG. 6, FIGS. 8 and 9 are block diagrams showing embodiments of the present invention, respectively, and FIGS. 10 and 11 are embodiments of FIGS. 8 and 9. FIG. 12 is a diagram showing the relationship between the pilot signal and the output code pattern for implementing the present invention in FIGS. 8 and 9. FIG.

Claims (1)

[Scope of Claims] 1. In a multiplex signal conversion device that converts an SSB-FDM signal into a PCM/TDM signal using fast Fourier transform means or fast inverse Fourier transform means, the fast Fourier transform means or the fast inverse Fourier transform means Under the condition that the scale of the SSB-FDM is larger than the multichannel
Inserting an analog signal as a pilot signal in a specific band outside the signal band, and monitoring the pilot signal generated as a digital signal with a specific code pattern at the output end of the fast Fourier transform means or the fast inverse Fourier transform means. A fault detection method for a multiplex signal converter, characterized in that a fault in the multiplex signal converter is detected by: 2. The fault detection system for a multiplex signal converter according to claim 1, wherein the pilot signal is determined to have a constant phase relationship with the multiplex sampling frequency of the SSB-FDM signal.