The phrase ‘low’ (‘limited’ or ‘restricted’) duty cycle appears in a large num- ber of ISM band allocation descriptions, as a favourite method of allowing multiple users without calling for the complicated radio hardware that other access schemes (multiple channel, spread spectrum or listen before talk) require.
The principle behind a low duty cycle system is very simple: by restricting the duration and rate of repetition of any one user’s transmissions it becomes statisti- cally unlikely that they will coincide (collide) with those of other, similarly limited, systems. In practice, the actual restriction is specified as a ‘duty cycle’, usually between 0.1% and 10%, depending on the band and the intended usage. In layman’s terms this reduces to “you stab the button at random and hope like hell that you’re the only one transmitting at that instant”.
There are certain implications that result from this type of operation: The transmitter must be designed for efficient use of the available time. It must switch on and off cleanly (not generating AM splatter, like a GSM cellphone does) and the switching needs to be as fast as possible to maximise the time allowed for data transmission. The data coding must minimise the amount of time lost in preamble and synchronis- BiM3B Transceiver ing sequences, and the overall data rate should be as high as practical (within other limitations, such as channel bandwidth).
Where more sophisticated systems are considered, and data integrity must be maintained, then a more effective means of dealing with data loss is needed. Sending the same data more than once will reduce (but not eliminate) the probability of completely losing a vital piece of information. In such a case, the interval between re-transmissions must not be constant (a randomising function is required), or there is the risk of another, identical system in close proximity falling ‘in step’ and the retransmissions always colliding.
Where no data loss is allowable (forexample, in a data download application from a data logger), then a transmit/acknowledge protocol must be used. This requires transceivers at both ends of the link (opposed to the transmitter/receiver pair used by the simpler methods) and the transmission of a ‘data received suc- cessfully’ message by the destination end after each burst. The failure to receive such a message within a given time window will then cause the originating unit to re-send the same data. In this way, the transmission will be repeated until success- fully received.
Additionally, once the complexity of a transceiver system has been implemented, then the user has the option of using other protocols beyond transmitting at ran- dom, as required (such as master/slave polling or master synchronised time slots), provided the actual transmission durations are within the spec limits.
by Myk Dormer, for Radiometrix Ltd
First published in Electronic World