DE3733967A1 - Quadrature superposition method for demodulating the carrier-frequency received signal in radio clock receivers - Google Patents

Abstract

The use of quadrature superposition methods in narrow band time signal receivers, in particular the use of the quadrature sampling derived therefrom, renders possible very compact receivers in which all that remains necessary as analog modules is the antenna and a preamplifier with prefiltering.

Description

The present invention relates to an arrangement for reception of AM-modulated time signal signals such. B. of DCF-77 signal. It is characterized in that a special mixing process is used.

Previously known receiver circuits (superimposed receivers) mostly use one or more in a row arranged mixer stages to match the input signal to corresponding Amplify intermediate frequencies without feedback and to be able to filter. Then z. B. by Rectification demodulated.

Another known method is that of synchronous demodulation. This mixes the input signal with a mixing frequency, which is exactly the carrier frequency of the one to be demodulated Signal corresponds. This method has over the first the advantage that high with very simple filters Filter effect can be achieved and already through that Mixing is demodulated. In addition, the overlay of the Useful signal through mirror frequencies using the simplest Filters are excluded. A disadvantage, however the effort (PLL) required to generate the exactly coupled Mixing frequency is required. In particular the reductions in the carrier signal complicate the establishment of a stable PLL circuit.

In contrast, in the arrangement according to the invention the use of a quadrature overlay method [4] the benefits of mixing down to the lowest frequencies be maintained without any special effort for the Mixed clock generation would be required. Through different Choice of mixing frequencies can be the different  available time signal transmitter are selected, with which universal time signal receiver can be realized (Claim 6) [5].

The basic principle of quadrature overlay is Mixing the input signal with two mixed frequency signals, the same frequency, but a phase shift of 90 ° to each other.

This technique is used in the literature of transmission technology [4] as a comfortable but difficult to implement Process described so that the less complicated Straight ahead and the overlay receiver preferred will.

The effort for the purely analog quadrature overlay on the one hand arises from the high demands on accuracy the multiplier, a quadrature overlay in digital representation, on the other hand, would be unrealistically high Require processing rates in the signal evaluation circuits.

Therefore, the quadrature overlay has so far also been used not used in time signal receivers where one can not afford a lot of equipment.

With the arrangement according to the invention and in particular with With the help of scanning mixers one arrives at one Receiver that is very suitable for very narrow reception bands is easy to implement and still has all the advantages of known known method.

In the following, it is first described how the Principle of quadrature overlay AM-modulated, carrier frequencies Signals mixed down and demodulated at the same time will. The arrangement on which the calculation is based corresponds the claim 1. The description is given in Time range in which the calculation is easy  leaves. Then the calculation of the quadrature scan presented in the arrangement of claims 2, 3, 4 and 5 corresponds.

1.1 The quadrature overlay in the time domain

In the following, the quadrature overlay is the same in the Application considered on the time signal.

If the useful signal y (t) is modulated onto a sine carrier of frequency f ₀, the function h (t) results, see Figure 1

h (t) = y (t) · cos (2 π f ₀ t) . (1)

This carrier-frequency signal h (t) is now not mixed in a known manner with a single frequency, but it is combined on the one hand with a signal sin (2 π f _m t) and on the other hand with a signal cos (2 π f _m t) multiplied. The frequency f _m can be chosen arbitrarily. The resulting signals are called x ₁ (t) x ₂ (t) . First, x ₁ (t) is calculated

x ₁ (t) = h (t) sin (2 π f _m t)
= y (t) cos (2 π f ₀ t ) sin (2 π f _m t)
= y (t) · (sin (2 π ( f _m - f₀) t ) + sin (2 π ( f _m + f₀) t )) / 2 (2)

The summand of the frequency f _m + f₀ can be eliminated by filtering with the simplest filters, especially if the mixed frequency f _m in the order of f ₀ has been chosen in accordance with claim 2. This particular choice of the mixing frequency also makes it possible to place the image frequencies occurring during the mixing down so close to the useful signal frequency that they become harmless. The filtered residual signal is called x _{1 r} (t). For the sake of simplicity, the forward gain of the filter function should always be chosen so that any pre-factors of the filtered signal can be shortened to one:

x _{1 r} (t) = y (t) sin (2 π ( f _m -f₀) t ). (3)

The same calculation for x ₂ (t) gives:

x ₂ (t) = h (t) cos (2 π f _m )
= y (t) cos (2 π f ₀ t ) cos (2 π f _m t)
= y (t) · (cos (2 π ( f _m -f₀) t ) + cos (2 π ( f _m + f ₀) t )) / 2 (4)

Here too, the summand of the frequency f _m + f ₀ can be eliminated by simple filtering. The residual signal is called x _{2 r} (t):

x _{2 r} (t) = y (t) cos (2 π ( f _m - f₀) t ). (5)

If the two signals x _{1 r} (t) and x _{2 r} (t) from equations (3) and (5) are squared and then summed, the terms that contain the difference frequencies add up to the value one:

x _{1 r} ² (t) + x _{2 r} ² (t) = y ² (t) · [sin² (2 π (f _m - f₀) t ) + cos² (2 π ( f _m - -f₀) t )] ² = y (t). (6)

If one takes the root of this expression, the desired useful signal y (t) results

The following is always the overlay with quadrature overlay of the carrier frequency useful signal with the orthogonal Signals, the subsequent squaring and the summation referred to both components. The following It is not necessary to square the resulting signal, because with binary time signs also from the squared Sum evaluates the time signal.

The multiplication with the phase-shifted mixed signals can be interpreted as complex multiplication in the time domain, where the cosine component is the real part and the Sinus component the imaginary part of the resulting mixed signal represents. The summation of the squared components and the subsequent etching then corresponds to the amount formation. If the complex components become the Calculated phase, even possible image frequency can be  Distinguish signals from the useful signal.

The calculation steps described above are illustrated once again in the electrical block diagram of the receiver in Figure 2, which corresponds to claim 1.

After reception by the antenna and after preamplification the signal becomes two mixing stages (multiplier stages) fed that with each other out of phase Signals of the same frequency are operated. The respective output signals are filtered, squared and added.

The squaring and summation can, as shown in Figure 3, already be carried out by a processor.

1.3 Quadrature scanning

In the block diagram in Figure 2, two mixing stages that generate the quadrature components are shown as multipliers, each with a subsequent bandpass. It will now be shown that the quadrature overlay, which requires two analog multiplication stages, can also be replaced by quadrature scanning. In block diagram Figure 4, which clarifies claim 3, two samplers with phase-shifted sampling clocks are provided for this. Like the mixing stages described above, the samplers are also operated at a sampling frequency in the order of the carrier frequency. As a scanner z. For example, so-called SC filters can be used, which can be used for scanning and filtering at the same time [3].

The necessary orthogonality between the scanning signals is generated by a phase or time shift of T _d = T / 4, where T indicates the period of the scanning frequency.

We now start again with the time function h (t) from equation (1). It is scanned with a sequence of pulses with an arbitrary phase position T _d . The scanning sequence is named as usual

where δ (t) is Dirac's needle pulse. Then the result x (t) of the sampled function h (t) is as follows

In order to find the Fourier transform X (f) , the transforms are first individually determined for the factors h (t) and s (t -T _d ). They are each named with the corresponding capital letters H (f) and S (f) . According to equation 1, two factors must be considered for H (f) , the transforms of which, as is known, must be folded in the frequency domain. Since the transformation of a sinusoidal oscillation of the frequency f Spektr results in two spectral lines at the frequencies ± f ₀, the spectrum H (f) belonging to h (t ) is obtained by convolution

This is the well-known spectrum of an AM-modulated signal. The transform to the sampling function s (t -T _d ) in Eq. 8 is

With equations (9), (10) and (11) the Fourier transform of x (t ) results in the convolution of the terms H (f) and S (f) :

If the sum is brought forward and the convolution is carried out, the sampling frequency f _abt = 1 / T follows:

The spectrum of the sampled signal is thus determined. Interfering mixed products are now filtered out with a bandpass of the resonance frequency f ₀- f _abt , so that only the components for n = 1 and n = -1 can pass through the filter. The remaining part remains

Transformed into the time domain:

= y (t) · [cos (2 π (f ₀ t - f _abt t + f _abt T _d ))] (15)

Now choose T _d = - T / 4 = -1/4 f _abt for the sine scan. Then you get the time function x _{1 r} (t)

x _{1 r} (t) = y (t) cos (2 π (f ₀ t - f _abt t) - π / 2)
= y (t) · sin (2 π (f ₀ t - f _abt t)). (16)

For the cosine scan, T _{d is} chosen to be 0:

x _{2 r} (t) = y (t) cos (2 π (f ₀ t - f _abt t)). (17)

If the two components are squared individually, added and the square root is finally drawn from the sum, so results as above:

The original signal is therefore also released from the carrier here restored.

The received interference signals, the frequency of which is a multiple of the carrier frequency f ₀, must be attenuated for the quadrature scanning by multiple pre-filtering. With the available carrier frequencies of the various time signal transmitters [1], however, this does not contradict the demand for universality.

Quadrature scanning can also be carried out directly by a fast microprocessor or a signal processor with the aid of two A / D converters of the required data width [2]. In Figure 5, two A / D converters operated with phase-shifted sampling clocks are used for this purpose.

The sampled signals can then be processed purely digitally, the processing steps corresponding to the arrangement of the block diagram in Figure 4.

Of course, only an A / D converter with a scanner can be used, which then digitizes the two quadrature components in time-division multiplexing (claim 5, Figure 6).

Various arrangements can thus be made with the described arrangements achieve pleasant properties:

In the methods of quadrature overlay and quadrature scanning can be mixed down at any frequency and be demodulated at the same time. There are only stability requirements to short-term stability for compliance the quadrature phase, the long-term stability is for the demodulation largely irrelevant.

Specular frequencies that could also pass through the filters are made harmless by the fact that the mixed frequency close to the carrier frequency to be demodulated is placed. In contrast to the known synchronous demodulation is neither a phase nor a frequency coupling here of the mixed signal with the carrier signal required.

If the mixed frequency is chosen so that the useful signal resulting from the superimposition has a very low resulting carrier frequency (f ₀ - f _m → 0), the simplest RC elements are sufficient to filter out undesired signal frequencies. By mixing down the useful signal, a large gain in quality can be achieved for the filters [3].

A sampling stage and a filter can be replaced by one compact SC filter to be replaced, creating a largely same behavior of both quadrature channels is achieved.

The entire quadrature overlay can be digital Area z. B. with the help of a signal processor. Then there is only a weak pre-filtering and one Amplitude adjustment required as analog circuit components.

In summary one can determine:
With quadrature scanning, a suitable method is obtained in order to enable a compact and, at the same time, flexible analog reception preamplifier for a largely digital universal receiver. The analog circuitry is limited to three operational amplifiers and two SC filters, even when using simple processors. This makes it possible to implement receivers the size of a matchbox.

Basically, even with the quadrature scanning Receiver can be realized, the analog part only consists of an antenna and an amplitude adjustment.

Corresponding to the different block diagrams Prototypes realized.  

2. Bibliography

[1] Bermbach, R .:
New radio clock concepts through close integration of receiver and microcomputer.
Darmstadt dissertation 1985
[2] Bermbach, R., J. Wietzke:
Use of digital signal processors in time signal receivers
Radio clock technology, Oldenbourg Verlag, 1987
[3] Bermbach, R., W. Hilberg:
Narrow band filtering using the alias effect.
Patent application 1985, patent granted 1987
[4] gap, HD:
Signal processing
Springer Verlag, Berlin 1975
[5] Wietzke, J .:
Radio clock with rod antenna
Radio clock, Oldenbourg Verlag, Munich, 1987

3. Captions

Figure 1: Carried time signal.
Figure 2: Block diagram of a quadrature overlay receiver.
Figure 3: Block diagram of a simplified quadrature overlay receiver.
Figure 4: Block diagram of a quadrature scanning receiver.
Figure 5: Block diagram of a simplified quadrature scanning receiver.
Figure 6: Block diagram of a quadrature overlay receiver with signal processor.

Claims (6)

1. Narrow band receiver for receiving the signals from time signal transmitters, in particular the transmitter DCF-77, containing an antenna, preamplifier, devices for mixing, filtering, amplification and signal evaluation, characterized in that the antenna signal after a preamplification simultaneously has two mixing devices ( Mixing stages) is supplied, in which it is mixed with sinusoidal signals that have the same frequency but a phase that is 90 ° different, that the intermediate frequency signals (quadrature components) that are produced are individually filtered and amplified, and that finally the results are squared and added, which results in the squared demodulated useful signal. 2. Arrangement according to claim 1, characterized in that as Mixing frequency a frequency is selected that is equal to the Carrier frequency is or near the carrier frequency lies, so that the resulting difference frequency is very small and thus the subsequent filtering and amplification is particularly easy to design. 3. Arrangement according to claims 1 and / or 2, characterized in that sampler for the two mixing devices be used (scanner mixer) and that these samplers  in violation of the known sampling theorem (accordingly the mixing in claim 2) with a sampling frequency in the Magnitude of the carrier frequency of the useful signal operated are, the sampling clocks of the two sampling mixers one Have a time difference of a quarter period to each other. 4. Arrangement according to one or more of claims 1, 2 or 3, characterized in that required for the two Probe mixer uses the sampler of two A / D converters and that the subsequent filtering and arithmetic operations already in digital representation of a processor be made (and thus gain and offset differences between the quadrature channels). 5. Arrangement according to one or more of claims 1, 2, 3 and 4, characterized in that instead of a scanning mixer the scanner of two A / D converters only a single A / D converter is used with the scanner in multiplex mode. 6. Arrangement according to one or more of claims 1, 2, 3, 4 and 5, characterized in that for realizing a universal time signal receiver the mixers or sampling mixers optionally different, on the different Time signal transmitter tuned mixing or sampling frequencies are fed.