programming logic help for level sensors

[jcraft]

We have some tanks that have have discrete level sensors in them. The hysteresis is very low in them and depending on the type of liquid we have in the tank the sensors can be twitchy at times which causes a dip axis to abort if the sensor is chattering at all. The manufacturer advises to either add some timed hysteresis into the program loop. My question is to do this portion of it I basically added an off delay timer to the fill valve in the tank and set for a couple seconds. this way the tank will fill for a few seconds after the sensor is satisfied. Does this sound like what the manufacture is suggesting to increase the hysteresis? If not how else could it be done. Also, there second reccomendation is to add a sensitivity filter on a monitor before the dip starts. This part is confusing to me so maybe someone may know how to do this portion. I was invisioning setting a regular timer in series with the level sensor bit and set for 3 seconds maybe and that would actuate a coil. That way the sensor would have to stay transitioned for a certain time before it set the output coil. Chattering wouldn't have any effect then. Does this sound like a filter. Programming is in GE Versapro. Thanks for any help

 

[OkiePC]

As long as a couple seconds is not enough to cause problems like overfilling the tank, then that sounds good. It is commonly called a debounce timer.

I am not 100% sure I understand "Also, their second reccomendation is to add a sensitivity filter on a monitor before the dip starts."

But if they're talking about discrete switches, I don't know what else it could mean aside from a debounce timer.

 

[TConnolly]

I've done several cooling towers where I've added debounce timers to the floats. Cooling towers can be extremely turbulent, to the point that even submerged floats can bounce. Simple on delay timers solve the problem.

 

[Kidblue]

Not for GE but it's pretty generic...(as long as you don't need different ON/OFF times)

Hope this helps.

 

[daba]

Kidblue - Your debounce logic suffers the same problem of many similar designs.

If your Input is ON, and the debounce timer has timed out, what happens when you turn the processor to Program mode, and back to Run, or the power goes off and is restored.

You guessed it, the output state is Reset, and the Timer has to time again. During this time of course, the debounce circuit is telling the rest of your process that the probe is uncovered.

What is needed is retention, where the state is maintained through a power-loss, or mode change.

I use the following, which has that retention, but also has immediate switching if the probe state has changed while the processor has been "down".