BIM-433-40

Radiometrix Hartcran House, Gibbs Couch, Watford, WD19 5EZ, England Issue 3, 13 July 2001 Tel: +44 (0) 20 8428 1220, Fax: +44 (0) 20 8428 1221 BiM-UHF Low Power UHF Data Transceiver Module UK Version - BiM-418-40 Euro Version - BiM-433-40 The BiM-418-40 and BiM-433-40 are miniature UHF radio modules capable of half duplex data transmission at speeds upto 40 Kbit/s over distances of 30 metres "inbuilding" and 120 metres open ground BiM-UHF transceiver Features Miniature PCB Mounting module Licence Exempt operation in UK on 418MHz, MPT 1340 (BiM-418-40) ETS 300-220 tested for European use on 433.92 MHz (BiM-433-40) SAW controlled FM transmission at -6dBm ERP. Double conversion Superhet receiver -107dBm receive sensitivity Single 4.5 to 5.5V supply < 15mA (tx or rx) Half duplex data at upto 40 kbit/s Reliable 30 metre in-building range Direct interface to 5V CMOS logic Fast 1ms power up enable for duty cycle power saving On board data slicer, supply switches and antenna change over. The module integrates a low power UHF FM transmitter and matching superhet receiver together with the data recovery and TX/RX change over circuits to provide a low cost solution to implementing a Bidirectional short range radio data link. The high data rates (upto 40kbit/s) and fast TX/RX changeover ( printer links High integrity wireless Fire / Security alarms Building environment control / monitoring Vehicle alarm systems Remote meter reading Authorisation / Access control Radiometrix Ltd, BiM Data Sheet page 1 figure 1: block diagram figure 2: mechanical dimensions Radiometrix Ltd, BiM Data Sheet page 2 Pin Description pin 1 & 3 RF GND These pins should be connected to the ground plane against which the integral antenna radiates . Internally connected to pins 9,10,18 . RF input / RF output for connection to an integral antenna. It has a nominal RF impedance of 50Ω and is capacitively isolated from the internal circuit. 0V connection for the modulation and supply Carrier Detect - When the receiver is enabled, a low indicates a signal above the detection threshold is being received. The output is high impedance (50kΩ) and should only be used to drive a CMOS logic input. This digital output from the internal data slicer is a squared version of the signal on pin 13 (AF). This signal is used to drive external digital decoders, it is true data (i.e. as fed to the transmitters data input). The 10kΩ output impedance is suitable for driving CMOS logic. Note: this output contain squared noise when no signal is being received. This is the FM demodulator output. It has a standing DC bias of approximately 1.5V and may be used to drive analogue data decoders such as modems or DTMF decoders. Output impedance is 10kΩ. Signal level approx. 0.4V pk to pk. We recommend this signal always be available on a convenient test point for diagnostic purposes. Note: unlike the RXD output which is always true data, this output is true data on the BiM-418 and inverted on the BiM-433. Should be driven directly by a CMOS logic device running on the same supply voltage as the module. Analogue drive may be used but must not drive this input above Vcc or below 0V. This input should be held at 3ms of preamble (55h or AAh hex) to allow the data slicer in the BiM to settle, followed by 1 or 2 FFh bytes to allow the receive UART to lock, followed by a unique start of message byte, (01h), then the data bytes and finally terminated by a CRC or check sum. The receiver data slicer provides the best bit error rate performance on codes with a 50:50 mark:space average over a 4ms period, a string of FFh or 00h is a very asymmetric code and will give poor error rates where reception is marginal. Only 50:50 codes may be used at data rates above 20kbit/s. We recommend 3 methods of improving mark:space ratio of serial codes, all 3 coding methods are suitable for transmission at 40kbit/s :• Method 1 - Bit coding Bit rate , Max 40kbit/s , Min 250bit/s Redundancy (per bit) 100% (Bi-phase), 200% (1/3 : 2/3) Each bit to be sent is divided in half, the first half is the bit to be sent and the second half, it's compliment. Thus each bit has a guaranteed transition in the centre and a mark:space of 50:50 . This is Bi-phase or Manchester coding and gives good results, however the 100% redundancy will give a true throughput of 20kbit/s. A less efficient, variation of Bi-phase is 1/3 : 2/3 bit coding. Each bit to be sent is divided into 3 parts, the first 1/3 is a low, mid 1/3 is the data bit and final 1/3 is high. This code is easy to decode since each bit always starts with a negative transition. This code should not be sent faster than 100µs per bit (10kbit/s) since the mark/space can vary for 33 to 67%. Radiometrix Ltd, BiM Data Sheet page 11 • Method 2 - Byte coding Bitrate , Max 40kbit/s , Min 2kbit/s Redundancy (per byte) 25% (synchronous), 50% (async) If only a subset of the ASCII code is required (e.g. 0-9 , A-Z and a few control codes) then translate (via. a look up table) the required ASCII codes into the 8 bit codes below. These codes all have a 50:50 mark/space when sent serially. Of the 256 possible 8 bit codes, 70 contain 4 ones & 4 zeros. The 68 Hex codes below have a 50:50 mark:space and may either be sent/received from a standard serial port (UART) using 1 start, 1 stop and no parity or as bytes of a synchronous code. Use for this subset also allows simple byte error checking on reception as all received codes must contain exactly 4 one's and 4 zero's. 17 4E 87 B2 1B 53 8B B4 1D 55 8D B8 1E 56 8E C3 27 59 93 C5 2B 5A 95 C6 2D 5C 96 C9 2E 63 99 CA 33 65 9A CC 35 66 9C D1 36 69 A3 D2 39 6A A5 D4 3A 6C A6 D8 3C 71 A9 E1 47 72 AA E2 4B 74 AC E4 4D 78 B1 E8 (note 0Fh & F0h have been omitted to minimise consecutive 0 or 1's) Other subsets are also possible e.g. a 10bit code has 1024 differs, 252 of which have 5 one's and 5 zero's i.e. a 50:50 M:S ratio. • Method 3 - FEC coding Bit rate , Max 40kbit/s , Min 4.8kbit/s Redundancy (per byte) 100% Each byte is sent twice; true then it's logical compliment. e.g. even bytes are true and odd bytes are inverted. this preserves a 50:50 balance. A refinement of this simple balancing method is to increase the stagger between the true and the inverted data streams and add parity to each byte. Thus the decoder may determine the integrity of each even byte received and on a parity failure select the subsequent inverted odd byte. The greater the stagger the higher the immunity to isolated burst errors. Radiometrix Ltd, BiM Data Sheet page 12 Linear operation A pair of transceivers may also be viewed as a linear analogue channel with a pass baseband of 100Hz to 17kHz with 40dB (see quieting curves v RF input). The test circuit shows the TXD input biased for linear operation and a simple digital filter to shape the transmit data to a raise cosine wave shape. The 22kΩ resistor linear- biases the TXD input. The drive voltage should be between 3.5 and 5V pk to pk to achieve full modulation (greatest S/N at receiver) figure 9: linear drive Raised cosine shaping may be applied externally to any serial data stream and will yield better error performance than unshaped data at high data rates (up to 40kbit/s) for data steams with 50:50 mark:space (4ms averaging period). Several excellent modem chips (FX 589 & FX 909) are available for Consumer Microcircuits Ltd (CML Tel 44 1376 513833). These chips employ GMSK (shaped data and matched receive filters) and enable operation up to 40kbit/s. figure 10: raised cosine generator Digitised analogue data Linear operation of BiM transceivers will allow direct transfer of analogue data, however in many applications the distortion and low frequency roll off are too high (e.g. bio-medical data such as ECG). The use of delta modulation is an excellent solution for analogue data in the range 1Hz up to 4KHz with less than 1% distortion. A number of propitiatory IC's such as Motorola's MC3517/8 provide CVSD Delta mod/demod on a single chip. Where the signal bandwidth extends down to DC , such as strain gauges, level sensing, load cells etc. then V-F / F-V chips (such as Nat Semi LM331) provide a simple means of digitising. Packet data Radiometrix Ltd, BiM Data Sheet page 13 In general, data to be sent via a radio link is formed into a serial "packet" of the form :Preamble - Control - Address - Data - CRC Where: Preamble: This is mandatory for the receiver in the BiM to stabilise. The BiM will be stable after 3ms. Additional preamble time may be desired for decoder bit sync., software carrier detection or receiver wake up. The minimum requirement is a single bit or unique bit pattern to indicate the start of message (frame sync.). Additionally, decoder information is often placed here such as: packet count, byte count, flow control bits (e.g. ACK, repeat count), repeater control, scrambler information etc. This information is used for identification purposes and would at least contain a 16/24 bit source address, additionally - destination address, site / system code , unit number and repeater address's may be placed here. User data , generally limited to 256 bytes of less (very long packets should be avoided to minimise repeat overheads on CRC failure and channel hogging). 16/24 Bit CRC or Checksum of control-address-data fields used by the decoder to verify the integrity of the packet. Control: Address: Data: CRC: The exact makeup of the packet depends upon the system requirements and may involve some complex air-traffic density statistics to optimise through-put in large networked systems. Networks BiM's may be used in many different configurations from simple pair's to multi-node random access networks. The BiM is a single frequency device thus in a multi node system the signalling protocol must use Time Division Multiple Access. In a TDMA network only one transmitter may be on at a time, "clash" occurs when two or more transmitters are on at the same time and will often cause data loss at the receivers. TDMA networks may be configured in several ways - Synchronous (time slots), Polling (master-slave) or Random access (async packet switching e.g. X25). Networked BiM's allow several techniques for range / reliability enhancement: Store and forward Repeaters: If the operating protocol of the network is designed to allow data path control then data may be routed "via" intermediate nodes. The inclusion of a repeating function in the network protocol either via dedicated repeater/router nodes or simply utilising existing nodes allows limitless network expansion. Spacial Diversity: In buildings multi-path signals create null spots in the coverage pattern as a result of signal cancellation. In master-slave networks it is cost effective to provide 2 BiM's with separate antenna at the master station. The null spot patterns will be different for the two BiM's . This technique 'fills in' the null spots, i.e. a handshake failure on the first BiM due to a signal null is likely to succeed on the 2nd BiM. Radiometrix Ltd, BiM Data Sheet page 14 Receiver Battery Saving In many applications the receiver need not be always waiting for a signal (i.e. drawing 15mA). Often it is only required to turn the RX on after a transmission to receive handshake data , thereafter it may be deselected (i.e.