Problem with the fail-safe feature of rs485

[mitaka]

rs485 fail safe problem

i use sp491 (s*ip*ex) and fail-safe feature don't work
with open inputs some of the IC have a 0 on receiver output, others 1 on the same conditions.
is anyone have the same problem?

 

[zape]

rs485 fail safe problem

Let me understand. Are you implementing a fail safe network external to the receiver? or your problem comes just from leaving the inputs in high impedance?

 

[dumbfrog]

Re: rs485 fail safe problem

it should give u a logic "1" when the input is left unconnected

also, make sure the REB' is low to enable the receiver (if not, it might give u a logic "0".

 

[mitaka]

Re: rs485 fail safe problem

the fail-safe feature is embeded in the chip, but on some chip it works on the others it not. the receiver is enabled always.

 

[mitaka]

Re: rs485 fail safe problem

this is the answer that i receive:
The receiver failsafe SP485E and SP491E have receivers that will output
logic 1 when the inputs are left unconnected. In order for this to function
reliably, there must be nothing at all connected to the inputs.
Looking at your attached schematic it shows there is a 120ohm termination
resistor across the inputs, also some TVS diodes and series resistors. In
that configuration the receivers are not guaranteed to failsafe to logic 1.
The termination resistor will short the A and B lines to zero volts
differential. The weak internal pull-ups for failsafe are not enough to
overcome the shorting from a termination resistor.
Any other components, even in high impedance modes, may leak enough to push the receiver over its threshold. Even long traces can act as an antenna to generate noise signal enough to overcome the failsafe.
If you require a defined idle state it is best to use an external
bias-resistor network that can maintain sufficient differential.