RS 485 wake up problem

[Inkwaterman]

Hi !

I have the following question:

I have to do a network of devices based on PICmicro.
This network in a RS485 and the device is battery powered.

For save the battery I put in sleep mode the PIC micro and I want that when there's a RS485 comunication all pic wake up for process the incoming packets and return in sleep mode.

I read the feature of many RS485 tranceiver and I saw that they use a current of quite 1 mA.

I saw also that is possible put in sleep mode some kind of tranceiver.

Is possible to put the tranceiver in sleep mode and wake it up , with the PIC , when the RS485 line become busy ?

Which kind of IC I have to use ?


Thanks .

[Ttelmah]

You can't put the transceiver to sleep, but you can turn off it's transmit component. Look at a perfectly 'standard' 485 driver like the MAX485. The 'quiescent', with the driver off, is 120uA typical. This has two control pins, /RE which enables the receiver, and DE, which enables the driver. Whenever you are not sending data, drop DE, and the power consumption drops as well. Some of the Texas transceivers have an even lower consumption in this mode.
Now on the standard PICs, the USART, cannot wake you from sleep. It relies on the presence of the oscillator, to work. However there are latter chips like the 18F1320, which can run the USART, from the internal clock, while the main processor is asleep. On these the consumption of the USART block is a few uA. Then latter chips still, with the EUSART module, and 'nanoWatt' technology, support the ability to wake if the incoming data line drops. This is designed to work in synchronous mode, but can be used with care to wake the processor. The alternative (available on all the PICs), is to wire the incoming serial line to both the RX pin on the PIC, and an interrupt input, and use this interrupt to wake the PIC.
There is a 'caveat', with all sleep modes, that if the oscillator is a crystal, it'll take a significant time to wake up. On a HS crystal oscillator, this can be in the order of 1mSec,nd for a *4 PLL on the same source, a total of about 3mSec is needed. The code needs to be designed to send a 'wake' character, and then pause for long enough to allow the chip to be ready, before actually sending data.

Best Wishes

[Humberto]

Inkwaterman,

In the pointed thread, you will find another good solution for RS485 remote wake up, posted by SherpaDoug:

http://www.ccsinfo.com/forum/viewtopic.php?t=18055&highlight=rs485


Humberto