DE2634357C3 - Circuit arrangement for converting pulse code modulated signals into carrier frequency signals - Google Patents

Description

J ₁ Ci

/, Γ2

Z,

where C I is the capacitance of the input side

The invention relates to a circuit arrangement for converting PCM signals into TF signals, which has a Have a frequency shift compared to a reference frequency,

with a digital-to-analog converter connected to the PCM signal source,

connected with a plurality of de ^r t output of the digital-to-analog converter and be switched in time series switches,

jo with the first bandpasses, one of which is connected downstream of each switch, each of which is used for Separation of a carrier-frequency single sideband signal is used and its passbands are selected in this way are that the frequency shift of the pass band

J5 is an integer multiple of the sampling frequency used to generate the PCM signals compared to the reference frequency position,
with modulators, each of which is connected on the input side to the output of one of the bandpass filters and which are fed with different carrier frequencies and

with a second bandpass filter connected downstream of the modulators combined on the output side.
In telephone networks, carrier frequency signals are transmitted in the upper network level and LF signals in the lower network levels. The carrier frequency technology or TF technology enables multiple use of the transmission paths through frequency division multiplexing. In the lower network levels,

W in particular on the terminal trunk lines or connecting lines between the node exchange and the end exchange, the pulse code modulation or PCM is growing Meaning. The pulse code modulation enables multiple use of the existing low-frequency

Y or NF cables through time division multiplexing.

With the introduction of pulse code modulation on the terminal trunk lines, the introduction of PCM switching attractive in the knot offices. To adapt the carrier-frequency transmission technology in the

μ upper network levels and the PCM switching technology Corresponding converter devices are required.

The object of the invention is to find an arrangement that is as simple as possible, the PCM signals in distortion

h> converts poor TF signals.

F i g. I shows an arrangement which can be used to convert PCM signals into TF signals. There are a PCM and a TF transmission system

NF-wise interconnected. The PCM multiplex signal is converted in a digital-to-analog converter 1 into a pulse-amplitude-modulated multiplex signal PAM , in which, for. B. 30 PAM signals, each having a pulse repetition frequency / 4 = 8 kHz, are interleaved.

A channel switch 2, which is controlled by a clock pulse of the pulse repetition frequency / 4 = 8 kHz with a pulse width of approximately 2.5 us, fades out a pulse-amplitude-modulated single-channel signal from the pulse-amplitude-modulated multiplex signal and feeds it to a low-pass filter 71w . The low-pass filter 71 converts converts the pulse-amplitude-modulated signal into an LF signal, which is amplified with the aid of the amplifier 72 and output via a transmitter interface 73.

F i g. 2 shows for the in FIG. 1 shows a schematic representation of the spectra of the signals in the individual processing stages. For the sake of simplicity, it is assumed that the LF signal occupies the frequency range from 0 to 4 kHz, in fact it only occupies the frequency range from 300 Hz to 3400 Hz. The spectrum of the pulse-amplitude-modulated signal consists of spectral zones, each of which contains the spectrum of the LF signal shifted by a whole multiple of the pulse repetition frequency in normal and inverted positions. Several such signals, interleaved in time, form the multiplex signal PAM. A pulse-amplitude-modulated signal is converted into an LF signal in that a low-pass filter 71 suppresses all spectral components above the frequency f _A / 2 = 4 kHz.

The LF signal reaches a first modulator 83 via a transformer interface 73, 81 and a low-pass filter 82, to which a carrier signal with the carrier frequency hr \ = 48 kHz is fed.

The first modulator outputs a double sideband amplitude-modulated signal AM, the center frequency of which is 48 kHz. The following filter 4 or channel filter suppresses a sideband and thus converts the AM signal into a single sideband modulated signal EM . The spectrum of the EM signal results from the spectrum of the LF signal by shifting it by 48 kHz.

In a second modulator 5, the EM signal is transmitted with one of the carrier frequencies / Vr2 = 112 kHz ... 156kHz inverted and put in its place in the Basic primary group moved. The TF signal of the basic primary group contains twelve NF signals in the Frequency division multiplex. The undesired sidebands arising during the second modulation become suppressed by a band filter 6.

There is a finding within the scope of the invention now in the fact that in an arrangement of the basis of FIG. 1 kind of explained the spectrum of the in the first The EM signal generated by the modulator is already included in the spectrum of the PAM signal. The EM signal can in particular with a band-pass filter whose passband, to put it simply, 48 kHz to 52 kHz. is exactly 48.3 kHz to 51.4 kHz, can be generated directly from the PAM signal.

From the French patent application 73I6Ö59 (Publication No. 22 58 056) a method is already known which makes use of this knowledge.

The pulse width of the scanning signals should, however, be small compared to the period of the filtered out frequency be. Even small fluctuations in the pulse width lead to fief.ience-dependent attenuation distortions.

According to the invention, the circuit arrangement is designed such that in the connection each Switch with the downstream bandpass filter, a matching circuit is inserted in each case that the incoming pulse-amplitude-modulated signal is converted into an impressed pulse stream in such a way that the from charge transferred to the individual current pulses to the output of the matching circuit is proportional to the amplitude of the pulse amplitude modulated signal.

Here you can advantageously use the receiving part of a PCM device in this way Connect the transmission part of a channel converter together so that the effort for the LF circuits is saved. In particular, the PCM low-pass filter and the first modulation stage in the TF device are saved.

Further advantageous refinements of the invention emerge from the subclaims.

The invention is based on the in FIGS. 3 and 4 and based on the in F i g. 5 illustrated frequency plan Sews explained.

It shows

1, as already explained in more detail at the outset, a circuit arrangement for converting PCM signals into TF signals;

F i g. 2 a frequency plan, also already explained, for the circuit arrangement according to FIG. 1:

F i g. 3 shows a circuit arrangement for converting PCM signals into TF signals without generating the Low frequency;

F i g. 4 shows an adaptation circuit for the circuit arrangement according to FIG. 3;

F i g. 5 shows a frequency plan from which the conditions can be derived when using a sampling low-pass filter.

The in F i g. 3 is used to convert PCM multiplex signals into TF multiplex signals. The PCM multiplex signals are over the digital-to-analog converter 1 is supplied to a branch point. This branch point is over several parallel branches, of which only one is shown in more detail in the figure, to the input of the Filter bandpass 6 out.

The digital-to-analog converter 1 converts a PCM multiplex signal into a PAM multiplex signal, each individual PAM signal having the pulse repetition rate Fa .

The branch shown in more detail in the figure contains the switch connected to the digital-to-analog converter 2, which together with the switches of the branches not shown in detail forms a distribution arrangement with whose help sends the pulses from the digital-to-analog converter in chronological order to the assigned channel filters are distributed, the pulsampiitudenmodulier th signals of the individual in the PCM multiplex signal enchai.enen channels to reach the individual branches.

The branch shown in more detail in the figure contains the channel switch 2 at the input, via which the digital-to-analog converter 1 is connected to the input Ei of the matching circuit 3. The output A 1 of the Anpassungsschaltu ⁿ g is provided with a first band-pass filter 4 and channel filters of the TF system connected. A clock pulse with the frequency / 4 is fed to the input E1 of the adaptation circuit 3.

The first band filter 4 is followed by the modulator 5, which is fed with the carrier frequency fm. The outputs of the module motors 5 are jointly led to the input of the second band filter 6.

The band filters 4 are designed in the same way for all branches and have one in the example described

Passband from 48 kl Iz to 52 kHz. so that d; is Spectrum of the appearing at the output of the band filter 4 EM signal results from the spectrum of the LF signal, in turn, by shifting by 48 kHz.

The EM signal, which has the same frequency length in all branches, is generated with the help of the modulator 5 each converted into a different frequency position, with the help of carrier frequencies 112 kHz to 156 kHz, in such a way that, as already shown on the basis of FIG. 2 described, results in a basic primary group.

The amount by which the EM signals are shifted with respect to the reference frequency - or the original frequency position, in particular with respect to the frequency range of the coded LF signals - can optionally also have other values. A prerequisite is, however, that the EM signal or the pass frequency range of the band filter 4 is shifted by the amount in ■ Λ. where m is an integer and f _{A is} the sampling frequency.

! Τ !! ΐ d? R the beautiful?.! *? ¹ "? And Hnr ArmlntJ-Diciital-Wanrllrr work.

In the example according to FIG. 3, the individual signals are telephone signals with a bandwidth of 4 kHz. Appropriately, the sampling frequency U = 8 kHz and the whole number mb, so that circuit parts can be taken over from conventional channel converters in an advantageous manner.

In the circuit arrangement according to FIG. 3, the digital-to-analog converter I is assigned to several channels. In this case it is necessary to have 4 switches between the digital-to-analog converter and the channel filters 2 to be arranged via which the output of the digital-to-analog converter is applied in sequence to the inputs of the channel filter 4. These switches have im Principle a pure distribution function and are for the conversion of individual PCM signals into TF signals not mandatory. However, they can also take on additional functions, in particular those of definition the pulse duration.

In the event that with the help of the circuit arrangement according to FIG. 3 PCM individual signals in TF individual signals are to be converted, it is sufficient to only use the A branch shown in detail in the figure without providing the switch 2. In this case the Digital-to-analog converter converts a single PCM signal into a single PAM signal.

The power of the converted signal emitted by the channel filter is approximately proportional to the Duration of the amplitude-modulated pulses present at the input. The pulse duration should therefore be very constant be. With a permissible influence of the pulse duration on the level stability of t%, the pulse duration may by a maximum of Fluctuate by 1%.

It can therefore be useful not to control the channel filter directly, but rather via an adapter circuit that is insensitive to fluctuations in the pulse duration. Another advantage of the matching circuit is that. With a suitable design, the channel filters of the TF systems, which are usually not dimensioned for control with pulses, can be used directly without additional frequency-dependent attenuation distortions occurring.

According to FIG. 3 provided matching circuit 3 is therefore useful when particularly high demands are made on the transmission quality. If this is not the case, it may be omitted.

The adaptation circuit converts the incoming PAM signal into a largely impressed pulse stream . so that the charge transferred by the individual current pulses / around output A I is proportional to the amplitude of the PAM signal. The duration of the pulse currents should be small compared to the period of the frequency m · f. \, Which for a frequency scheme according to FIG. 4 is the reciprocal of 48 kHz. Such a matching circuit is shown in more detail in FIG. The switch 2 is across the longitudinal resistance; 31 led to the capacitor 32 located in an input-side shunt branch. The connection of the capacitor 32 carrying voltage against reference potential is led via the sampling switch 33 to the output A \ . The resistor 34 is located in an output-side shunt branch, parallel to it a parallel resonance circuit of the following bandpass filter 4. The resistor 31 symbolizes the internal resistance of the digital-to-analog converter I, supplemented by the resistance of the wiring between the analog-to-digital converter 1 and the capacitor 32 Switches 2 and 33 close at different Z ^p '* C ⁿ , and 7war .Srhaltpr 2 zur / pil t] + n- T. where T = 1 // ignition switch 33 at time ii + n'T. their contact times must not overlap. Advantageously, an integer multiple of the reciprocal of the clock frequency of the pulse-amplitude-modulated multiplex signal PAM is chosen for the time difference / \ t \ - t \\ and the duration of a channel sampling pulse is selected for the closing duration ri of switch 2 and Γ2 of switch 33. Appropriately applies

R, C ₁

and

SaCI

Z.

where R _{1 is} the value of resistor 31. Cl is the capacitance of capacitor 32, C2 is the capacitance of capacitor 41 and Z is the nominal value of the bandpass terminating resistor. The receipt £ 2 of the

In particular, a clock pulse with the frequency /! a = 8 kHz and the pulse width 2.5 jis is fed to the matching circuit 3. The sampling switch 3i is controlled with the aid of this clock pulse.

In the arrangement according to FIG. 3 must be taken into account.

υ that the blocking attenuation of the channel filter or bandpass 4 in the frequency range above the converted LF signal, in the example above 52 kHz. sufficient to meet the noise and crosstalk requirements placed on the system. This doesn't have to be the case since

jo according to FIG. 2 with this arrangement there are significantly more sidebands at the input of the KanalF'ers than in the TF device. A missing selection can be brought up by inserting a sampling low-pass filter, not shown in the figures, in front of the matching circuit 3 , the clock frequency f _{T of} which is selected so that it is an integral multiple of the sampling frequency f _A. in particular of 8 kHz, and also the carrier frequency, in the example 48 kHz, is an integer multiple of or equal to the clock frequency

For an implementation with frequencies according to FIG. 2, the clock frequency Fr can therefore assume the values 16 kHz, 24 kHz or 48 kHz.

Since the PCM system delivers values in a time interval corresponding to / χ = 8 kHz, the sampling low-pass filter lacks intermediate values. The value NuK is expediently fed in as intermediate values. The frequency spectrum is shown schematically in FIG. 5 for a clock frequency Zr = 16 kHz. When using a

2b 34 357

\ b '. ¹ M. ':!,!' ^, ". Is; (in · Schnl ι! Ι t <] u ^ i \ / des ViMsi ¹ 'ml (ι I) (' i.irgt: '' iipled IK" . κ e) -l here. I ¹ / «.ils I Ili

■ ¹ I s I ^: .'K ¹ HH der [,! Kllii '(| iieii /' mi .nliiih'sr · ■ ι [) ρ ι · 11 c η! Elements, realized '.> Ιτ ιΙιτι

I. ■ ■ / ι / ι ;;! vim 1 '".' M niifli-ink'n. I: k.iiui iibei, um h vnr (K ¹ In Diyilal-.An.ilii.L 'SVindlc'- I

I) .r- ■ ¹¹ IiIiCIV'-;'!. ki: ill ' ¹ , mi I in ^ MiiL''C- \ • ^l .i>. | ii' .-! p.i'- ■ i'iiiu '.' i " i'.ci (U ¹ Ii and .iK I) iL "i.ilfillcr, irhi, - !!« Λ; ι

- ' ¹ ^.!''' MT s: He k, inn er fur mchrci-- Kim.ilc l'K ⁱⁱ i -. nij:

'.Γ ■' ■ .In SpCT (I.IHlpllil !! '(Ιο · - Λ l ·', 1>. I! I ι '' \ >, Γ ■ ■■■, lOS'CMIII / ". V \ ι '' Ί 1 '11, VV (I It CI lKnill JlC I, 1''! ICl], CIl / (K-- ¹ ,' ■ ι 'Il''ι' ι Ί: 'I, lk · U; et j! [C; / '. Ι ■ <■ 1 hk H /! U ¹ J' ■ -. Ι '.!'! Ρ.ι S '.''. 11 *. Ρ f C ^ - h C 1 '■ ': ■ < !! S U '! L'!! * IC ^v '. '. '. !!!'

ιΊι '. !.! ■ · I i -'.- iinc: / -i '' - inn ", im \ '-L'ai-. · De \ Μ ^1. ! Ι. ■ ■■ ■' ■ ■ S <'; .i-! l'I "s 2 ci'ü'ih: sich ιΙ, ιηρ, ■>'.I. ·

'-' I "'■!.: K'. I, .CV ij '■ \ bl, ISl'! Cip,! SSCS | ll '() I l'V ^J i .!' Μ !.>

! ' ■ \ bi, i \ niclp, i '■■ (>' ■ dci \ iisw · .: ' ¹ S (Ii.illu-i!' -.:, I '■ - ■ ^Mi i ..' \ i " h, i! 'i ^; - sicil'I Lc I tkllrcqm "/ des ;> - ■ _ ■ 11: <I

'■', MU- 'Xi''** (! IC I lll.l kl! ΙΠ' Π ¹ - '"CS' ir- 'it ι'',!\',! - 'i ¹ V ,, | ['. iM' sj

Claims (8)

Patent claims: 1. Circuit arrangement for converting PCM signals into TF signals that have a frequency shift compared to a reference frequency, with a connected to the PCM signal source Digital-to-analog converter, with several connected to the output of the digital-to-analog converter and in chronological order switchable switches, with first bandpass filters, one of which is connected downstream of each switch, each of which serves to separate a carrier-frequency single-sideband signal and whose passbands are selected in such a way that the frequency shift of the passband with respect to the reference frequency position is an integral multiple of that used to form the PCM signal / used sampling frequency,
BUt modulators, each of which is connected on the input side to the output of one of the bandpass filters and which are fed with different carrier frequencies and with a second bandpass filter connected downstream of the modulators combined on the output side,
characterized in that a matching circuit (3) is inserted in the connection of each switch (2) to the downstream bandpass filter (4), which converts the incoming pulse-amplitude-modulated signal (PAM) into an impressed pulse current, in such a way that the individual current pulses the charge transferred to the output (A 1) of the matching circuit (3) is proportional to the amplitude of the pulsari.plitudenmodulierlen signal. 2. Circuit arrangement according to claim 1, characterized in that the TF multiplex signal a Basic primary group around the first bandpass channel filters of a TF transmission system with pre-modulation are. 3. Circuit arrangement according to claim 2, characterized in that the pass band of the Channel filter is 48 to 52 kHz, and that the sampling frequency has the value 8 kHz. 4. Circuit arrangement according to claim 1, characterized in that the duration of the pulse currents is small against which is shifted from the frequency by which the frequency band is shifted from the reference frequency position is. resulting period is. 5. Circuit arrangement according to claim I or 4, characterized in that the matching circuit (3) one in an input side branch Contains capacitor (32) to which a sampling switch (M) with subsequent transverse resistance is attached (3 *) follows. 6. Circuit arrangement according to claim 5, characterized by such a dimensioning that the product of the internal resistance of the digital-to-analog converter and deer capacitance (Ci) of the capacitor lying in the input-side shunt branch is very much smaller than the closing time of the upstream sampling switch (2), and that Cross-branch capacitor and C2 is the capacitance of the capacitor located in an input-side cross-branch of the bandpass filter and Z is the nominal value of the BandpaQ terminating resistor, and that the value of the cross-resistor (34) following the sampling switch (33) is equal to Z is 7. Circuit arrangement according to one of the preceding claims, characterized in that a sampling low-pass filter is inserted upstream of the matching circuit (3), the clock frequency of which is an integral multiple of the sampling frequency (fa) , and that the carrier frequency (fr _r ) is an integral multiple of or equal to is the sampling frequency (f _A ) 8. Circuit arrangement according to one of the preceding claims, characterized in that a sampling low-pass filter is inserted in front of the digital-to-analog converter (1)