WO1992005639A1

WO1992005639A1 - Apparatus and method for compensating for loss in co-axial cable

Info

Publication number: WO1992005639A1
Application number: PCT/CA1991/000335
Authority: WO
Inventors: Stephen Webster
Original assignee: Gennum Corporation
Priority date: 1990-09-20
Filing date: 1991-09-20
Publication date: 1992-04-02
Also published as: CA2025797A1

Abstract

In a coaxial cable, the loss varies with frequency. To compensate for this, a frequency domain transfer function is generated which is a constant multiplied by the square of the frequency. This is then added to the original input signal, to generate a compensated signal. The constant of proportionality is varied depending upon the co-axial cable type and length. The constant proportionality can be varied automatically depending upon the peak level of the compensated output signal. The technique can be relatively insensitive to variations in component parameters and hence can be implemented in a monolithic integrated circuit.

Description

Title: Apparatus and Method for Compensating for Loss in Co-Axial Cable FIELD OF THE INVENTION

This invention relates to an apparatus for compensating or providing equalization for the different loss/frequency characteristics of commercially available co-axial cable.

BACKGROUND OF THE INVENTION

For commercial available co-axial cables, the loss in the cable varies with frequency. Generally, higher frequencies are attenuated more than lower frequencies. However, all frequencies are propagated at sensibly the same velocity, hence the phase relationship between the various harmonics and components of a waveform transmitted through a cable are maintained constant, i.e. these characteristics are roughly the same at both the receiving and transmitting ends. The transmission of serial encoded binary data inherently requires a transmission medium with a broad bandwidth, since the spectrum content of such signals is wide and substantially flat. Distortion can arise if the transmission medium does not have a reasonably flat response across the necessary bandwidth. At some point, this distortion will be large enough to lead to errors in the received data. It is known that generally, for frequencies greater than several megahertz, the loss/frequency characteristic of such commercially available co-axial cables is approximated , fairly closely, by the following equation:

G = exp - (KC.X. Vf) where: KC is a constant for a particular type of cable.

X is the cable length f is the signal frequencies

Accordingly, it is desirable to provide some sort of apparatus or device, which can compensate or equalize for the variation of loss characteristic with frequency for co-axial cable.

One known cable equalizer, has been proposed by the Sony Corporation. This is a high frequency equalizer, which is built into a decoder chip.

An operating principle of this circuit is to maintain the peak level of the signal constant after equalization. For this purpose, a variation of the peak level, at the output is first smoothed and preserved by a hold capacitor. The potential across the hold capacitor affects the equalization characteristics, which again changes the peak level. This provides a feedback loop so that the circuit reaches a stable state with the peak at the desired level. For this to be successful, the output level of the transmitter must be tightly specified so that

SUBSTITUTE SHEET the loss characteristics of the co-axial cable and the curve of equalization practically match at any length within the design specification, in this case 0 to 300 metres. In this proposal, the original input signal is augmented twice, in two separate stages. Each stage includes a variable gain amplifier that is connected, through a buffer amplifier, to the hold capacitor, so that the gain is controlled in dependence upon the output. Each stage further has a buffer amplifier which is connected to a summation unit and to the variable gain amplifier. The variable gain amplifier has an output connected to a high frequency amplifier which produces a signal that is added to the original signal from buffer amplifier in the summation unit. As there are two separate stages, it is essential to match the time constant of the two stages in order to obtain the desired output. Matching time constants requires matching component values which is difficult to achieve in a monolithic circuit.

SUMMARY OF THE PRESENT INVENTION

As discussed above, it is known that there is a standard formula or equation for determining the loss characteristic of co-axial cable.

The present invention is based on the realization that, in order to correct for this loss, one

SUBSTITUTE SHEET needs to apply an inverse of the transfer function given by the above equation. Further, it has been realized that the inverse function can be simplified, so as not to require precise time constants. This in turn enables it to be readily implemented in monolithic form, where precise component values are not readily obtained. As detailed below, it has been realized that this approximation requires adding to the original signal, the original signal multiplied by a constant and the square of the frequency.

Thus, in accordance with the present invention, there is provided an apparatus, for compensating for the variation of loss with frequency of a co-axial cable, the apparatus comprising, means for generating a frequency domain transfer function, which is a constant multiplied by the square of the frequency, the constant corresponding to the characteristics of the respective co-axial cable, and for multiplying an input signal by that transfer function, to form an auxiliary signal; a straight through channel for the input signal; and a summation means having inputs connected to the generating means and the straight through channel for adding that auxiliary signal to the original input signal to produce a compensated output signal. The present invention also provides a method for compensating for the variation of loss with frequency of a co-axial cable, the method comprising the steps of:

SUBSTITUTE SHEET (1) generating a frequency domain transfer function, comprising a constant multiplied by the square of the frequency;

(2) multiplying an input signal by the frequency domain transfer function to obtain an auxiliary signal; and

(3) adding that auxiliary signal to the input signal, to obtain a compensated output signal. Preferably, the auxiliary signal is generated by first multiplying the input signal by a constant of proportionality, and then subjecting that signal to separate differentiating steps. The constant of proportionality is selected so that when multiplied by the gain applied during each differentiating step, one obtains the desired constant.

For a better understanding of the present invention, and to show more clearly how it may be carried into effect, reference will now be made, by way of example, to the accompanying drawings, in which;

Figure 1 is a block diagram of a first embodiment of the present invention; Figure 2 is a block diagram of a second embodiment of the present invention;

Figure 3 is a block diagram of a third

SUBSTITUTE SHEET embodiment of the present invention; and

Figure 4 is a schematic of a circuit for implementing the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

As discussed above, the loss/frequency characteristic for commercially available co-axial cable, for frequencies above several megahertz, can be approximated by the following equation:

G = exp - (KC.X. Vf)

It has been realized that, to compensate for this loss, it is necessary to apply a system of apparatus with a response which is a reciprocal of the cable loss characteristic, with zero phase shift.

It has further been realized that the inverse of the cable characteristic can be simplified, to the following equation.

G^"1 = exp (KC.X. Vf) = 1+Kf² Where K is a constant to proportionality which is chosen to optimise fit within a range of frequencies, and for various types and lengths of cable. An advantage of this formula is that it simply requires the adding to the original signal of an auxiliary signal, which is dependant upon the constant and square of the frequency. As detailed below, this can be implemented

SUBSTITUTE SHEET simply, in a manner which is highly suitable for incorporation in a monolithic circuit. There is no necessity to match different circuit components to give comparable characteristics. Referring to Figure 1, the first embodiment of the apparatus, generally denoted by the reference 1 includes an input 2 and a straight through channel 4 connected to a summation unit 6. The summation unit 6 has an output 8 forming the output of the whole apparatus 1. A high pass channel 10 includes a first amplifer or gain unit 12 which amplifies the signal by a constant of proportionality K' . Unit 12 is connected to a first differential unit 14 where the function CRd/dt is applied to the signal. A second differential unit 16 similarly applies the function CRd/dt to the signal.

In a practical implementation, each of the differential units, 14, 16 will apply a 90 degree wideband phase shift so that the overall phase shift through the channel 10 will be 180 degrees. This is compensated for by a phase inversion at the summation unit 6, as indicated.

Accordingly, the overall transfer function affected through both channels 4 and 10 is given by the following equation. G^"1 = 1 + K.f²

It can be noted then it is not necessary for the CR values in the units 14 and 16 to be exactly identical.

SUBSTITUTE SHEET Any variations in these values can be compensated for by the constant of proportionality K'to give the desired constant K. Accordingly, it is expected that the apparatus device shown in Figure 1 should be readily implemented in a monolithic circuit.

Figure 1 shows a unit 10 with a fixed constant of proportionality. In many cases, to ensure a more accurate output, it is desirable to provide a self- adjusting capability. An apparatus including this facility is shown in Figure 2.

In Figure 2 , components that are common to those of Figure 1 are given the same reference numeral. The description of these components is not repeated for simplicity. The apparatus of Figure 2 is generally denoted by the reference 20.

Instead of the amplifier or gain unit 12, a variable gain unit 22 is provided. The output 8 is connected by a line 24 to a peak detection unit 26. The output from this peak detection unit 26 is connected to the negative input of a second summation unit 28 which includes a positive input for a DC reference level 30. The output of the summation unit 28 is connected to an error amplifier 32 which in turn has an output connected to the variable gain unit 22. Assuming the output signal, at the output 8, has been fully compensated, the peak amplitude of the signal should be the same as the first signal at the transmitting end of the co-axial cable. An excessive peak level is then indicative of over

SUBSTITUTE SHttϊ compensation and conversely a deficiency is the peak level is indicative of insufficient compensation. The reference DC level 30 is therefore set equal to the peak level of the transmitted signal, so that the apparatus 20 will automatically adjust the gain of the high pass channel 10 to provide the correct amount of compensation. This arrangement is therefore capable of adjusting automatically for different cable loss factors and for different lengths of cable. Further, the manufacturing tolerance for components used in the second embodiment 20 is reduced, since the value of these components are also compensated for by the negative feedback action (this assumes that there is always surplus gain available at the variable gain unit 22). This second embodiment is therefore even better suited for implementation as a monolithic integrated circuit, employing on-chip capacitors and resistors to define the time constants of the high pass channel 10. Neither the absolute value nor the matching of the time constants is critical to the operation of the apparatus.

An actual physical implementation of the differentiator functions of the units 14 ,16 will almost certainly introduce a finite propagation delay into the high-pass channel. Provided this delay is reasonably constant for all frequencies of interest, the high pass characteristic is not adversely affected. However, if the delay is not insignificant in comparison to the bit-period

ITUTE SHEET of the transmitted data, then it becomes necessary to equalize the delay for both the high pass channel 10 and the straight through channel 4 so that they are synchronized at the summation unit 6. Accordingly a third embodiment of the invention is discussed in relation to Figure 3. Again, components similar to those in Figure 2 are given the same reference numeral and the description of these components is not repeated.

This third embodiment is generally denoted by the reference 40. The apparatus 40 includes a delay unit 42 in the channel 4, to ensure that the time delay for the two channels 4, 10 is substantially identical, for all frequencies of interest. For data transmission, it is sufficient if the difference in the delays in the two channels is not significant compared to the bit-period for transmitted data. In a practical implementation of the apparatus, severe noise problems would result if the bandwidth of the high pass channel 10 was not limited. This can be achieve either by the insertion of a low-pass filter into the channel 10 or, more conveniently, by utilizing the natural high frequency roll-off of the differentiators if this is appropriate. The bandwidth of the high pass channel 10 need only extend as far as the highest fundamental frequency of the transmitted data (which is one-half of the bit-rate) . For best noise immunity, the high pass response should be rolled off beyond that frequency. It could also be noted that this roll-off, essential for good noise performance, is largely

SUBSTITUTE SHEE responsible for the propagation delay through the high- pass channel of a practical embodiment of the apparatus.

Figure 5 shows a circuit, indicated at 50 for implementing the present invention with variable gain. The circuit 50 additionally includes a comparaton section 52, not shown in the earlier block diagrams. The circuit 50 has an input 54 and an output 56. Additionally, it has an input 58 for an Automatic Gain Control Capaciton 56.

SUBSTITUTE SHEET

Claims

I CLAIM;

1. An apparatus, for compensating for the variation of loss with frequency for a co-axial cable, the apparatus comprising: generating means for generating a frequency domain transfer function, which is a constant multiplied by the square of the frequency, the constant corresponding to the characteristics of the respective co-axial cable, and for multiplying an input signal by that transfer function, to form an auxiliary signal; a straight through channel for the input signal; and a summation means, having inputs connected to the generating means and the straight through channel, for adding that auxiliary signal to the original input signal to produce a compensated output signal.

2. An apparatus as claimed in claim 1, wherein the generating means includes a gain unit for multiplying the input signal by a constant of proportionality.

3. An apparatus as claimed in claim 2, wherein the generating means further includes first and second differential units connected in series, for differentiating the input signal twice, to obtain the square of the frequency.

SUBSTITUTE SHEET

4. An apparatus as claimed in claim 3, wherein eac differential unit is such as to effect a 90 degre wideband phase shift, and wherein the output of th differential units is connected to a negative input of th summation means to compensate for the phase shift.

5. An apparatus as claimed in claim 2, 3, or wherein the gain unit comprises a variable gain unit, an wherein the apparatus includes control means fo controlling the output of the variable gain unit, i dependence upon the output of the summation means, t maintain the peak output detected at the summation means at a desired level;

6. An apparatus as claimed in claim 5 wherein the control means comprises a peak detection unit connected to the output of the summation means, a second summation means having a negative input connected to the output of the peak detection unit and a positive input for a DC reference level, and an error amplifier having an input connected to an output of the second summation means, the error amplifier being connected to and controlling the variable gain unit.

7. An apparatus claimed in claim 2, 3, or 4, which includes a delay unit in the straight through channel;

8. An apparatus as claimed in claim 6, which includes a delay unit in the straight through channel;

9. A method of compensating for the variation of loss with frequency in a co-axial cable, the method comprising the steps of:

(1) generating a frequency domain transfer function, which is a constant multiplied by the square of the frequency, with the constant being chosen to correspond to the characteristics of the respective co¬ axial cable;

(2) multiplying an input signal by the frequency domain transfer function to obtain an auxiliary signal.

(3) adding the input signal to the auxiliary signal to obtain a compensated output signal.

10. A method as claimed in claim 9 wherein step (1) comprises:

a) multiplying the input signal by a constant of proportionality b) effecting a first differentiation of the signal so as to multiply the signal by the frequency; c) effecting a second differentiation of the signal so as to multiply the signal by

SUBSTITUTE SHEET the frequency, with the gain of th differentiating steps of (b) & (c) bein chosen so that together with the constan proportionality, they give the desire constant.

11. A method claimed in claim 9 or 10, wherein th compensated output signal is monitored and where th constant of proportionality is varied in dependence upo peak detected for the compensated output signal.

12. The method as claimed in claim 11 wherein th peak detected for the compensated output signal is compared to a reference DC level, to give an error signal, which is amplified and controls a variable gain amplifier generating the constant of proportionality.

13. A method as claimed in claim 9 or 10, wherein the auxiliary signal and the input signal are summed in a summation unit, and wherein the input signal is transmitted to the summation unit through a delay unit to compensate for delay in generation of the frequency domain transfer function.

14. A method is claimed in claim 12, wherein the auxiliary signal and the input signal are summed in a summation unit, and wherein the input signal is transmitted to the summation unit through a delay unit to

TE SHEET compensate for delay in generation of the frequency domain transfer function.

SUBSTITUTE SHEET