GB2166556A

GB2166556A - Measuring frequency response of digital transmission systems

Info

Publication number: GB2166556A
Application number: GB08526498A
Authority: GB
Inventors: Peter Wagner
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 1984-11-02
Filing date: 1985-10-28
Publication date: 1986-05-08
Also published as: GB2166556B; JPS61112972A; GB8526498D0; DE3439918C2; DE3439918A1

Abstract

In a method for measuring the frequency response of digital transmission systems, e.g. digital filters in magnetic tape recorders, periodically recurring pulses (Fig. 1) or pulse trains (Fig. 2) are produced (1) (Fig. 3) in which each pulse has a width corresponding to the reciprocal value of the bit frequency (bit period) of the digital signal and in which the recurrence frequency of the pulses or pulse trains is small compared with the bit frequency. The test signal is supplied to the transmission system (2) and the frequency spectrum of the output signal of the transmission system is analyzed (5). <IMAGE>

Description

SPECIFICATION Method for measuring the frequency response of a transmission system for digital signals This invention relates to a method for measuring the frequency response of a transmission system for digital signals.

Methods are already known for the measurement of the frequency response of transmission systems for digital signals, in which a test signal produced at the output of a wobbulator is supplied to the transmission system to be tested via an analog/digital converter, and undergoes digital/analog conversion following the transmission system and is finally made visible with the aid of a spectral analyzer. It is also known for the purpose of testing transmission channels, particularly the recording and/or playback channel of a magnetic tape recorder, to record signals with a given pulse shape and represent the same oscillographically after passing through the transmission channel following playback from a magnetic tape.The disadvantage occurs when using the known method that the played-back signals are often difficult to evaluate oscillographically, so that it is e.g. difficult to balance a recording/playback channel, such operation calling for considerable experience.

The object of the present invention is to provide a method for measuring the frequency response of a transmission system for digital signals, which requires little effort and expenditure and which can be performed in a simple manner.

Accordingly, the invention provides a method for the measurement of the frequency response of a transmission system for digital signals, comprising supplying to the transmission system a test signal in the form of periodically recurring pulses or pulse trains in which each pulse has a width corresponding to the reciprocal value of the bit frequency (bit period) of the digital signals, and in which the recurrence frequency of the pulses or pulse trains is small compared with the bit frequency, and analysing the frequency spectrum of the output signal of the transmission system.

The invention has the advantages that the test signal can be produced with simple digital circuitry, and that the frequency response is directly indicated.

It is particularly advantageous for testing the recording/playback channel of digital filters in magnetic tape recorders to insert a test signal according to the invention into a signal to be recorded during time intervals in which no information is present. This permits checking of the recording/playback channel during actual operation.

An embodiment of the invention is de scribed in greater detail hereinafter with reference to the accompanying drawings, wherein: Figure 1 is a voltage-time diagram of a first test signal which may be used in performing the invention, Figure 2 is a voltage-time diagram of a second test signal which may be used in performing the invention, Figure 3 is an embodiment of an arrangement for performing the method according to the invention, Figure 4 is the spectral density of a test signal used in performing the invention, Figure 5 is a detail of the diagram of Fig. 4, Figure 6 is an embodiment of an arrangement for performing the method according to the invention on the recording/playback channel of a magnetic tape recorder, and Figure 7 is the spectral distribution of signals occurring in the arrangement according to Fig. 6.

Fig. 1 shows a test signal usable in the invention, such signal comprising periodically recurring pulses having a high level B and a width equal to the bit period T of the digital signals, the test signal being at a low level A between pulses for a time amounting to an integral multiple of the bit period. The period of the test signal is NXT, in which N is significantly higher than 1. As N approaches infinity, the frequency spectrum of such a signal consists of lines having identical amplitudes and which recur with a spacing of 1 (NXT).

The spectrum of Fig. 4 is obtained because in practice N is not equal to infinity.

A similar spectrum occurs in the case of a test signal according to Fig. 2, in which a pulse-form oscillation with half the bit frequency of the digital signals is modified at intervals of NXT by intermittently suppressing one pulse. This second test signal therefore comprises consecutive trains of pulses in each of which the pulses have a width of one bit period and are spaced apart by one bit period, and wherein between consecutive pulse trains there is an interval of one pulse period. Such a test signal may be generated by the logical interconnection (exclusive OR) of a signal according to Fig. 1 and an uninterrupted pulseform oscillation with half the bit frequency of the digital signals.The spectrum of the test signal according to Fig. 2 differs from that of the signal according to Fig. 1 in that there is additionally a strong spectral line at half the bit frequency (shown in broken line form in Fig. 4). A signal according to Fig. 2 is particularly suitable for the checking of transmission channels not permitting the transmission of d.c. voltages because the d.c. voltage part is relatively small compared with the a.c. voltage part of the signal.

When selecting N, it must be borne in mind that for smaller values of N there will be fewer spectral lines within the frequency range up to the first zero possible for evaluation (1/T in Fig. 4). However, as N becomes lar ger, the energy content of the test signal to be evaluated by the spectral analyzer decreases, so that limits occur where the test signal does not differ significantly from the noise. Values for N between 64 and 256 have proved advantageous when testing transmission channels for digital signals with a 50 MHz bit frequency.

Fig. 3 is a block circuit diagram of an arrangement for performing a method according to the invention. Test generator 1 produces a test signal according to Fig. 1 or 2, which is supplied to the transmission system 2 to be tested. The test signal passing through the transmission system is supplied via a changeover switch 3 to a digital/analog converter 4.

The resulting analog signal is then represented in a spectrum with the aid of a commercially available spectral analyzer 5.

Reference is made to the following in connection with the digital/analog converter 4. In the transmission of digital signals, it is important that the transmission channel for these signals has an adequate band width. If the band width is too small, the digital signals are defomed in such a way that a trouble free regeneration and conversion of the transmitted signals is no longer possible. For testing such a channel by the method according to the invention, the transmitted signal is therefore supplied directly to the spectral analyzer without conversion into analog form. The effect of the frequency response of the transmission channel of the digital signal is important for this particular use.However, there are also transmisson channels in the wider sense in which the frequency response of the analog signals represented by digital signal is influenced, this mainly relating to digital filters. For the measurement of such frequency responses it is necessary, as shown in Fig. 3, to place a digital/analog converter 4 upstream of the spectral analyzer 5. In the system according to Fig. 3, the digital signals are also transmitted in parallel form, i.e. for example with eight parallel lines. In order to be able to carry out a comparison measurement the digital/analog converter 4 can be directly connected to the test generator with the aid of a switch 3.

Fig. 5 shows spectra with a frequency scale modified compared with Fig. 4. Curve a corresponds to the spectrum of the test signal Curve b represents the spectrum of the output signal of the digital low-pass filter following digital/analog conversion.

Claims

1. A method for the measurement of the frequency response of a transmission system for digital signals, comprising supplying to the transmission system a test signal in the form of periodically recurring pulses or pulse trains in which each pulse has a width corresponding to the reciprocal value of the bit frequency (bit period) of the digital signals, and in which the recurrence frequency of the pulses or pulse trains is small compared with the bit frequency, and analysing the frequency spectrum of the output signal of the transmission system.

2. A method according to claim 1, wherein the test signal comprises consecutive trains of pulses in each of which the pulses have a width of one bit period and are spaced apart by one bit period, and wherein between consecutive pulse trains there is an interval of one pulse period.

3. A method according to claim 1 or 2, wherein the output signal is digital/analog converted prior to spectral analysis.