GB2194866A - Timing signals - Google Patents

Abstract

Timing signal is distributed by amplitude modulating coded timing information on an audio frequency carrier signal. The modulation depth is sufficiently low for the carrier signal always to be detectable, and the modulation is time coherent with the cycles of the carrier frequency. Preferably each bit of the timing information corresponds to an integral number of cycles with the carrier signal. <IMAGE>

Description

SPECIFICATION Timing signals This invention relates to the generation and use of timing signals, this term including time and date reference signals.

In a broadcasting organisation there is a particular need reliably to maintain correct time iniormation throughout the system in order that programme material arising from several scattered sources is brought together smoothly and in sequence. There is also a need for accurate time difference measurement (e.g. stopwatches) in assembling components of an edited programme. There are some instances where time signals are broadcast directly (such as the Greenwich Time signal, the Ceefax clock, and the television clock) and, as in other organisations, there are applications involving data processing and control of services and security where timed and date information is required for recording events.

Conventionai solutions to the problems involve the use of impulse clocks, driven at one second, half-minute or minute intervals, together with the use of modern free-running quartz clocks and watches which have good short-term accuracy but which require occasional adjustment or resetting at unpredictable times. These latter result in wasted time taken to find new batteries and then to iind a reliable source to reset the time and they can only be relied upon at all timnes if there is a regular programme of replenishment which is labour-intensive and administratively burdensome.

It is already known to distribute time and date codes as serial data streams such as by using conventional data modulation methods such as frequency shift keying between frequencies of 1200Hz and 2400Hz.

It is also known to distribute standard frequencies or tone signals and to use these as a sort of time reference.

We have appreciated that these two concepts can be combined in a particular way to provide an especially convenient method of time signal distribution.

In accordance with this invention we propose the distribution of a timing signal which comprises an audio frequency carrier signal, on which the coded timing information is amplitude modulated, the modulation depth being sufficiently low for the carrier signal always to be detectably present, and the modulation being time-coherent with the cycles of the carrier frequency.

With the use of this invention it is possible to generate and distribute a standard signal to carry all the information necessary for timekeeping purposes within the broadcasting organisation. This signal serves all purposes where accuracies of a few milliseconds are sufficient; it excludes the particular needs of waveform generation and synchronisation, where accuracies of the order of a few nanoseconds are required but where adequate techniques already exist.

One embodiment of the invention will now be described by way of example with reference to the accompanying drawings in which: Figure 1 is a block diagram of a time signal generator or encoder; and Figure 2 is a block diagram of a decoder for decoding information from the signal produced by the circuit of Figure 1.

In the example illustrated a continuous and coherent carrier frequency, here a 1kHz sinewave, is used to distribute a regular and easily-recovered frequency reference. Amplitude modulation is used to convey the data stream. The amplitude modulation is not between 100 percent and 0 percent, as with for example the Greenwich Time signal, but is typically between 100 percent and 35 percent. This ensures that the carrier signal is always present and its zero crossings can be used as a regular time base for decoding the modulation. Using a relatively low depth of modulation preserves the zero-crossings of the carrier and allows simple recovery of the carrier.

The modulation is coherent with the carrier, that is the data boundaries are phased absolutely with respect to cycles of the carrier. For example a IkHz carrier can be modulated with data at 100 bits per second so that every bit corresponds to exactly ten cycles of the carrier waveform.

In this way it is possible to construct a system with several desirable features. It is economic because low-cost decoders can be used and there is no need for local locked oscillators. It is efficient in terms of bandwidth and travels through audio lines. It is rugged in that it is immune to noise and is not itself a source of interference.

Figure 1 shows an encoder system 10 comprising a frequency reference 12 which may be calibrated by an external frequency and timing reference 14. The internal frequency reference 12 drives a master counter 16 which supplies two feeds to an amplitude modulator 18. The first is a 1kHz carrier tone and the second comprises serial timing data at 100 bit/s, the content of which is given below. The carrier is modulated with the timing data and fed to distribution amplifier circuitry 20 for distribution as required.

The decoder 30 of Figure 2 has an input 1kHz bandpass filter 32 which feeds an automatic gain control (AGC) amplifier 34. From the output of this amplifier a first circuit 36 recovers the 1kHz carrier which is outputted for local use as a reference tone. The 1kHz signal is also applied to a demodulator 38 which supplies serial data-and a 100 Hz clock signal to serial data decoder and storage registers 40. The way the stored data is used depends upon the local need. As shown, a local display selector 42 controls which elements of the data are applied to a display 44.

An example of the coded signal content will now be given. The cycle repeats every second and it should be noted that the numeric data sent during each second relates to the second indicated by the next following second mark.

Bits Content 00-06 Seconds marker - always l's.

07-11 Minute marker - all l's at the first second of each minute.

12 Space 13-15 Source identification.

16 If this signal carries UTC then 1, otherwise 0 18 If daylight saving applies now, then 1, otherwise 0.

Indicator 19 If daylight saving applies Codes this time next week, then 1, otherwise 0.

20 If a positive leap second is due at the end of this UTC quarter year, then 1, otherwise 0.

21 If a negative leap second is due at the end of this UTC quarter year, then 1, otherwise 0.

22-25 Sign and magnitude of DUT1 26-45 Modified Julian date (BCD) 46-52 Week number (BCD) Date 1 53-55 Day of week 56-63 Year number (two digits,BCD) 64-68 Month number (BCD) 69-74 Day of month (BCD) 75-80 Hour (BCD) Time 4 81-87 Minute (BCD) 88-94 Second (BCD) 95-99 Pause Some of these terms need a little explanation.UTC is a universal time code based on atomic time which differs from Greenwich Mean Time (GMT) by a small amount known as DUT1. Every so often an extra second is added or one omitted to keep the two in step, this being known as a positive or negative leap second. The modified Julian Date (MJD) is a cumulative day count, independent of calender conventions. For example, Ist January 1987 is MJD 46,795.

The above code is such that no more than five consecutive ones can occur in the data except during the seconds pulse (length seven bits) or minutes pulse (length 12 bits). This is achieved in part by using binary coded decimal (BCD) format for most of the time and date information, as indicated above. some of the advantage of the system resides in the considerable redundancy in the transmitted information as to date which enables the date or day to be retrieved at the decoder in the most convenient manner for the particular application for which it is required.

The modulated clock in the example described is designed to allow simple recovery of seconds and minutes information. The information sent every second is sufficient completely to specify all that is likely to be required, including time, date, week number and summertime offset. The principle is that all the necessary calculations are performed once at the encoder, rather than in the decoder. This reduces the cost of the decoder, which does not then need any arithmetic capability, and it ensures that all decoders give the same answer. A decider plugged into the system will always respond within two seconds.

It is envisaged that there can be simple time and date display decoders, and special decoders offering extra facilities (e.g., week number, or time and date in other countries). standard decoding interfaces could be devised to allow computer systems (e.g., Ceefax, Datacast, Radiodata) to accept this data, as indicated on Figure 2.

The time and date reference signal can be distributed as an audio signal, either through standard audio lines, at a low level, as the signal is proof against noise and interference and should not itself cause interference, or through telephone-line wiring. If necessary the signal could be made available by telephone or even broadcast if required.

In order to distribute the signal nationwide it could be modulated on an out-of-band carrier (e.g., 21kHz) and carried within a high-quality audio channel, preferably preserving the coherence between the data modulation and the carrier phase. Such a signal could then be regenerated at audio frequency for use at the destination. If necessary, the information could be coupled into a data link, perhaps as part of a digital sound or video multiplex, in which case the interface would, ideally, add information on the relative timing of the data and the multiplex in order to preserve accuracy.

Claims (9)

1. A method of distributing a timing signal, comprising generating an audio frequency carrier signal, and amplitude modulating coded timing information on the carrier signal, the modulation depth being sufficiently low for the carrier signal always to be detectably present, and the modulation being time-coherent with the cycles of the carrier frequency.

2. A method according to Claim 1, in which the modulation is between 100% and X% where X is not less than

3. A method according to Claim 1 or 2, in which each bit of the timing information corresponds to an integral number of cycles of the carrier signal.

4. A method according to any of Claims 1 to 3, in which the timing information is arranged so as to be incapable of producing more than Y successive 1's or 0's except as a seconds timing marker pulse.

5. A method according to any of Claims 1 to 4, in which at least some of the timing information is in binary-coded decimal (BCD) format.

6. A method according to any of Claims 1 to 5, in which the information comprised in the timing information has an element of redundancy to facilitate selection of the desired information in the form most suitable for a particular application.

7. Apparatus for distributing a timing signal, comprising a generator for generating an audio frequency carrier signal, and a modulator for amplitude modulating coded timing information on the carrier signal, the modulation depth being sufficiently low for the carrier signal always to be detectably present, and the modulation being time-coherent with the cycles of the carrier frequency.

8. A method of distributing a timing signal, substantially as herein described with reference to the accompanying drawings.

9. Apparatus for distributing a timing signal, substantially as herein described with reference to the accompanyting drawings.