circuitcellar.com
Magazine Support   Digital Library   Products & Services   Suppliers Directory 
 
 





 

September 2006, Issue 194

Multifuctional Wireless Alarm
Freescale Wireless Design Challenge 2005


by Carl Smith


StartSystem Overview Installation & Operation Garage Door Module Water Level Module Temperature Module Hardware Design Garage Door Sensor Water Level Sensor Temperature Sensor Software Design Improvements Peace of Mind Sources and PDF

HARDWARE DESIGN

The base unit is the most complicated design in the system. It must be able to receive and store the messages from the remote sensors, provide synthesized audio output (for status reports, prompts, and menu selections), recognize DTMF tones to respond to the menu selections, and, of course, interface to the telephone line in order to answer incoming calls.

The MC13192-EVB PCB that came with the contest kit is the main part of the base unit. It’s mounted on a prototyping perfboard, where the rest of the circuitry was constructed with point-to-point wiring techniques (see Photo 1).

cc

(Click here to enlarge)

Photo 1—The base unit’s MC13192-EVB and EMIC module were mounted on a standard perfboard. The rest of the unit was built with point-to-point wiring techniques on the perfboard.

The MC13192-EVB provides 5 V to the rest of the circuitry with a tap into the S106 power switch leads (see Figure 1, p. 32). The I/O from J107 is connected to the prototyping board via a ribbon cable. The serial port at J103 is connected to the perfboard with a D-sub connector to interface to the speech synthesizer module.

cc

(Click here to enlarge)

Figure 1—The main components in the base unit are a Freescale MC13192-EVB, an Emic text-to-speech module, and a Cermetek CH1837A direct access arrangement.

A Grand Idea Studio Emic text-to-speech module (distributed by Parallax) provides the speech synthesis functions. The module supplies high-quality speech synthesis, and it interfaces with a standard TTL serial interface at 2,400 bps. Its connections to the microcontroller include a SERIAL IN, a SERIAL OUT, and a BUSY line to indicate when the module is busy speaking and cannot respond to commands.

The Emic uses TTL voltage levels for the serial input and output rather than the full RS-232 voltage levels output by the MC13192-EVB. This was a good decision by the designers. In most applications, the module simply would be connected to the serial outputs of a microcontroller, but it makes level translation circuitry necessary to interface to the MC13192-EVB. I used a Linear Technologies LT1081 level translator chip for this job. Its function is similar to the popular MAX232 family and all of its derivatives, any of which would have been well suited for the job. I just happened to have an LT1081 on hand.

The Emic has an onboard audio amplifier that can power external speakers up to 300 mW, which is sufficient volume for connecting to an unamplified speaker. It also has a separate analog output, which is connected to the telephone interface circuitry to send audio signals over the telephone line. An analog input allows the Emic to amplify audio from the phone line, which is useful for debugging purposes.

Sending ASCII text commands to the serial input at 2,400 bps controls the Emic. Commands are available to convert text to speech, set the volume level, set the speed and pitch of the synthesized speech, check the Emic version, turn on the audio input, and open a Help menu. You can also store, delete, and recall abbreviations.

I used only the reset and text-to-speech conversion commands for this project. To reset, simply send a reset; command to the module and wait for the busy line to go inactive. To convert text to speech, send the text to be spoken to the module preceded by the “say=” command and follow up with a semicolon to finish the command. For example, to say “hello,” send the say=hello; command.

The Emic’s busy line is connected to a logic input on the MC13192-EVB. This allows the software to wait for the Emic to finish one command before starting the next. Note that all of the logic external to the MC13192-EVB is 5-V logic. The MC13192-EVB uses 3-V logic, so a ULN2003 IC consisting of seven open-collector transistors with integrated base resistors is used for logic level translation for all of the logic inputs to the MC13192-EVB. The internal pull-ups on the MC13192-EVB’s logic lines are enabled in the software to ensure a logic high when the open collector transistor is off.

The audio generated by the Emic module is injected into the phone line through a telephone direct access arrangement circuit built around a Cermetek Microelectronics CH1837 DAA module that provides some surge and protection circuitry, isolation, a hybrid two- to four-wire converter that separates the transmitted and received audio, ring indication, and hook control. External to the DAA is more robust line protection circuitry and EMI suppression circuitry that includes capacitors from tip and ring to ground to bypass EMI, fuses for surge protection, and a transient voltage suppression diode across the tip and ring. The transmit connections go to the Emic speech module. The receive line is routed to the DTMF detection circuit. The ring indicator and off hook control lines are connected to the microcontroller through 5- to 3-V logic level translation.

DTMF detection is performed by a BG Micro SSI204 DTMF detection IC. Given more development time, I would have been able to implement a DTMF detection routine in software, but I would’ve needed hardware to amplify the incoming phone line audio to a range suitable for the ADC on the GT60 processor. The SSI204 chip interfaces directly to the DAA audio output with nothing more than a 0.1-µF capacitor.