About ST7538

[yongfiona]

Posted on February 06, 2004 at 07:49

About ST7538

[yongfiona]

Posted on January 27, 2004 at 04:54

hi! i would like to ask if any of u know whether ST7538 can be used on dc power line???

i'm doing my final year project and need to use a modem on DC power line communication. thanks!

[dix]

Posted on February 06, 2004 at 07:05

The 7538 can be used with a DC power line, as the modem simply works by coupling an FSK AC sinewave signal (132.5kHz - default) via a coupling network directly onto the power line. It doesn't care if the power line is carrying 240Vrms or some other low DC power source.

The key is to ensure that the coupling networks matches your requirement. As you're using DC then this actually makes the task much simpler, as you don't have the problem of having to establish a good 50/60Hz rejecting ratio. You might even get away with a simple capacitor as the coupling component. Have you a source of the chips, as I'm having problems getting hold of the device here in the UK. It wouldn't be a good idea to go down this path and then to find that you couldn't get hold of a few smaples. The other big problem is making up a prototype, as the 7538P is the only package and this is a TQFP44 device. Not ideal for mounting on a prototype board. The crystal oscillator is also a real pain in the bum, as the circuit uses a sub-threshold oscillator. This is to reduce the supply current of the chip, but it does make the oscillator a bit hit or miss, (frequency wise). If you use the recommended PCB layout then I'm sure its fine, but if you're trying to make up your own prototype, then its unlikely to be a real PCB, because of the cost limitations. The other major issue is that the 7538 needs an MCU to operate. Unlike other simpler design, the chip wont work without the SPI serial comms link to the MCU. The SPI comms link is also a real pain, as the modem must act as the master device and the MCU as the slave. The timing is non-standard SPI as well. Take care to ensure that the MCU supports the right phase and polarity for the serial link. The MicroChip PICs don't work too well with this configuration. The older ST7537 was much simpler to get up and running. As is the Philips TDA5051A. The 5051A is a much more basic part and uses ASK modulation, but it might be worth a look. Regards [ This message was edited by: Steveboy on 09-02-2004 17:25 ]

[dix]

Posted on February 06, 2004 at 07:49

Oh I missed one very important issue.

As the you're using a DC power source, then the output smoothing capacitor will shunt the FSK comms signal. [i.e. it looks like a low impedance at 132.5kHz]. So you need to find someway of adding a blocking network at the terminals of the power supply. If you're shifting loads of power around your DC supply network then this can be a major problem. [Inductive core saturation !] Foundation Fieldbus uses a similar idea, but gets around this idea by using a bespoke power supply. The supply has a funny dymanic response, so that its output looks like a high impedance at the comms frequency. Perhaps a little too complex for a final year project. You could consider using an low voltage 50/60Hz network and then rectify the AD to DC at the local power points along your network. If you use bulk transformers to stepdown the 240/ 120Vrms mains, then the impedance of the copper winding should be sufficiently high at the carrier freq. (132.5kHz). Check the impedance of your power source at 132.5kHz. This will tell you if the 7538 will be able to drive it. [ The Tx section will drive short peaks of up to 1A into the load. The Rx section is looking for a signal >1mV.] Some linear supplies us cheap output capacitors. After all they're only seeing 100/ 120Hz ripple, thus they don't have to have a low impedance at HF. The problem comes when you're using a switch-mode supply, as they use much better output capacitors. Regards [ This message was edited by: on 06-02-2004 12:22 ][ This message was edited by: Steveboy on 09-02-2004 17:22 ]