sensor fail safe mode

what is meant by fail-safe mode of sensor. how to program when this sensor goes ON.
usually i put a NO contact for the sensor and connect to output coil and latch it.
but i donno about this fail-safe mode of sensor.

fail safe relates directly to a safety device function.
a standard sensor can either fail always ON of Always OFF.
neither of these faults can be detected until the machine fails to operate correctly.
with fail-safe sensors they normally are connected through an electronic safety relay. this device fully monitors the sensor and will lock out the operation if an error is detected.
the sensor usually has dual contacts (2 pairs)
it also detects both broken or shorted connections.

While I agree with iant's answer with respect to safety devices, there is also another more general answer, too.

Any time you are evaluating the "what-if" scenarios in a system design, you have to determine how you want your system to respond when a sensor fails in either condition (short circuit or open circuit). Iant addressed this in his answer above for the safety devices, however it also applies to normal non-safety devices too. In "most" of these cases you are more interested in how to respond when a wire is broken (open circuit) but you definitely need to evaluate the shorted circuit condition, too. The term "Fail-safe" indicates that such a failure will cause the machine to respond in a favorable fashion (do nothing, stop, stay where it is, shut off the motor, etc.) rather than respond in an unfavorable fashion (motor turns on, motor stays on, valve closes causing extreme pressure, etc.). Frequently there is quite a lot of additional programming required to handle these conditions, but that's all part of it.

As others have said in this forum many, many times, the easy part is figuring out how to make things work the way it should when everything is perfect. The hard part is effectively handling all the break-down situations. This part, and how efficiently and effectively you achieve it, is also the same one that defines you as a programmer.

These are accurate answers, no need to correct anything but I would like to add a little.
A classic example of where one needs a fail safe sensor (or circuit) is a photo eye that tells the PLC to shut down a valve when a tank is full. If the sensor fails and does not stop the flow there will be either a mess or a hazardous condition. Either of these conditions will cost money and time. Another thing that these systems do is minimize downtime by streamlining the troubleshooting process by flagging where the problem is.
In some cases you may find "smart" I/O also that can detect opens and shorts in any type of I/O, even old fashioned limit switches. Like these I/O modules from GEIP:
IC695MDL664 Smart Input Module
IC695MDL765 Smart Output Module

I/O like this will help the programmer by keeping the program simpler and shorter by putting most of the diagnostics in the module rather than in logic.

Sorry if this sounds like an advertisement, it wasn't meant to be.

In "most" of these cases you are more interested in how to respond when a wire is broken (open circuit) but you definitely need to evaluate the shorted circuit condition, too.

Click to expand...

The classic example of a "fail-safe" function is the simple STOP pushbutton. Theoretically and logically a STOP switch could be either normally open or normally closed. If you make it normally open, then any cut wire could prevent a dangerous machine from stopping (not "fail-safe"). However if you make the pusbutton logic normally closed, then a cut, broken, or disconnected wire will do the same thing as the stop button: stop the machine, a "fail-safe" function (the controls have failed in the safest manner).