Posted on March 17, 2004 at 09:19Hi,
I'm using the ST7FLite29 as my MCU and thus my code is ST7 assembler. I'm a hardware designer anyway, so you're probably a lot better off not seeing any of my code.
The software interface of the 7538 is very simple, its the hardware that's complex. The main serial interface is basically just an SPI port, with the 7538 acting as the master in the SPI comms system.
i.e. simply hook-up the following lines;
MISO (MCU) to TxD (modem)
MOSI (MCU) to RxD (modem)
SCK (MCU) to CLR/T (modem)
And configure your SPI port. [Clock = idle low, Clock Phase = Valid data on the 1st rising edge.]
You can use the 7538 in asynchronous mode, (i.e. without CLR/T) but I wouldn't recommend this, as you can't configure the chip without the clock. So unless you wish to use the default configuration you need to use synchronous mode. It also saves you having to recover the clock in Rx mode.
The other key control lines are REG_DATA and Rx/Tx. The Status lines BU and CD-PD come into play when you're using the chip for real. You can operate without them to start with.
Okay, so to configure the chip take REG_DATA high and Rx/Tx low, (see table on page 11 of the datasheet). Load the correct 3 byte configure into the chip using your SPI port. Ensure the correct timing. (as per the diagrams on page 14) as this is critical.
When you've done this take REG_DATA low, as its best to idle in DATA mode. {stops you writing to the REG in error}.
The half duplex comms access control is then via Rx/Tx. As it simply controls the mode of the chip. Watch out here, as there are timing issues with the PLI interface. This takes 3.2msec to power up after you take Rx/Tx low. (see pages 16 and 20).
In Tx mode you simply check the BU signal, for a free comms LAN. If it free then you can transmit you message data stream. The BU ildes low and goes high when the FSK carrier is detected. Its actaully releted to the CD_PD line.
In Rx mode you sit and wait for the CD_PD line to go low. In my code I use the SPIF interript flag (of my SPI port) to capture the data. The datasheet isn't too clear here, but I've found that the chip pulls the CD_PD low for the complete duration of the Rx cycle. So you can test for CD_PD low (start of Rx data) and wait until CD_PD goes high (end of Rx data).
If you use pre-amble detection, then the CD_PD flags for a byte string of AA, AA, AA, AA or 55, 55, 55, 55 at the start of the data stream. If you select pre-amble in your 7538 configuration, but forget to send it, then the first 4 (non-zero) bytes are ignored and the CD_PD flags as a carrier detection on the 5th Rx byte. When you actaully send the pre-amble the CD_PD goes low as soon as the message is received. A bit of digital magic here, as the pre-amble also appears in the output data stream. So CD_PD appears to go low before the string has arrived. This is just due to the internal digital filtering and CD_PD actually flags in-time with the 1st data edge. Nice trick..!
Look out for the relative timing issue betweem Rx/Tx and REG_DATA. These are critical and the chip doesn't work if you get them wrong.
The 7538 chip really is this simple to interface to. However the hardware coupling inteface is a lot more complex. This requires a good knowledge of analogue electroincs and filter design, (have a look at AN1714).
Out of interest what type of project are you working on? Don't worry if you can't say. I was just interested, as I'm using the chip on a 24Vac /24Vdc loop rather then 240Vrms. I'm having real problems with the interface, as the impedance matching is critical. I'm also having real problems establishing reliable comms.
Good luck.
Regards
Steve.
nullnull
