Tour : click picture below for previous

Back to Mike's Electric Stuff

Click picture below for next

 Casio HBX-100 'PC-Unite' databank watch

00106034.JPG (254254 bytes)


I expected this unit to be built with some serious minaturisation techniques - chip-on board assembly and microscopic components. I was rather   surprised to see that apart from the main CPU, very conventional assembly was in fact used. I suspect this may have been to reduce development costs for what is probably a relatively low-volume product for Casio. Maybe future versions will be more miniaturised, reducing the overall size. I was also surprised that it used a flash device for data storage rather than battery-backed static RAM. Again this may have been due to availability of standard parts, the low volumes not justifying custom chips.

Check out TimexFun - Info on this and other high-tech watches.

00106006.JPG (112800 bytes) PCB front.

Grey circles on left are rubber buttons (conductive rubber pads on rear).
Large black square is main processor. Not sure if this is chip-on-board or a BGA (Ball-grid array) package.

Chip below it is a 16kbyte serial flash memory, Xicor X24F128. This presents some interesting questions - the watch specs state that memory capacity is approximately 8K, so what is the other 8K used for ? As it's a serial device, it can't be code, unless the CPU has a large on-chip RAM to copy it to. Also, as it's a flash device, and therefore nonvolatile, why does the watch trash the stored data when you change the battery and reset it?  Unfortunately my device programmer doesn't support this device, otherwise I would have read the contents!
I assume the use of a standard SO package is that the die is too big for a smaller package - this is often the case with larger eeprom/serial flash parts.

Black object on right is the infra-red IRDA transceiver. I was surprised to see one of these used instead of a smaller descrete infra-red LED/Receiver pair, which could have been a lot smaller.

Note the circular solder-bridgeable pad pair to the left of the second grey button from the top. When delving inside things you often find these factory-settable links. Of course I had to try linking it.... When this pad is linked, the PC software reports that the watch type is unrecognised, and the text entry uses only Japanese alphabetic characters.

00106011.JPG (94734 bytes)PCB rear.

Going round the board clockwise starting at 12 O'clock....

Inductor for switching power supply chip (top right), I think this is the supply for the infra-red transceiver. I suspect this chip is one of the Torex range

Bottom right - crystal (probably 32.768KHz) for timekeeping (obscured by black felt pad stuck to the top) , with fine-adjust trimmer above-right.

Bottom centre - 3.68MHz Ceramic resonator main CPU clock. The CPU will spend most of its time asleep or running at the lower watch crystal frequency to save power, and use this high-frequency clock when it needs to do anything more demending than show the time.

Chip at top-left (Panasonic M803 83) drives the electroluminescent backlight. It uses the round coil above it to step up the 3v battery to the high voltage (typically 50-100V) required by the backlight.

00106017.JPG (354957 bytes)This is the PCB from the PAD-1 PC to IR interface supplied with the watch.

I suspect this is a standard OEM product which has been badged by Casio.It uses a standard Hewlett-Packard IRDA interface chip, pluse some standard logic devices to interface the PC RS-232 port to an IRDA transceiver.

The packaging warns that the interface will not work on a laptop - this is because it takes its power from the RS-232 port, and laptops tend to have less power available from serial ports due to the low-power circuitry they typically use to generate the RS-232 voltage levels.


Click picture above for Previous

Back to Mike's Electric Stuff

Click picture above for next