Digital Debounce? Is Rockwell Just Plain Wrong?

BillRobinson · Nov 10, 2010

Ok,

I'm using Rockwell's PlantPax add on instructions. In the package there is a Digital-In function that has a "Debounce" option, which doesn't seem to work. Maybe I'm wrong, but I thought debouncing protected against quick false positives/negatives. Here is their code.



     input      B.TT                   A

|-----||--------|/|-----------------[TIMER]------|

|  |      A.TT       |                           |

|  |--------||-------|                           |

|                                                |

|                                                |

|    input      A.TT                   B         |

|-----|/|-------|/|-----------------[TIMER]------|

|  |      B.TT       |                           |

|  |--------||-------|                           |

|                                                |

|     A.EN                            OUT        |

|-----||------------------------------( )--------|

Here's my test signal/results/expected

input _____|-|_______|----------|_|---------|_________

output _____|----|____|----------|___|-------|_________

expect _________________|----------------------|_________

Which doesn't seem to be a debounced signal at all.

rdrast · Nov 10, 2010

It looks like it works just fine.
Timer A extends any valid input pulse to a minimum duration on the output, while allowing for the output to still be continuous.
Timer B requires that the input be off for a set time before re-triggering.

Traditional debounce circuits are intended to prevent literal contact bounce when opening or closing, which is generally exhibited with many transitions of true/false/true/false in a micro to millisecond range.

BillRobinson · Nov 10, 2010

OK I get it, a debounce doesn't check the validity of the initial instance, just the chatter after the initial instance.

Thanks rdrast!

Eric Nelson · Nov 10, 2010

To get your 'expected' result, this should work:



     input                             A

|-----| |---------------------------[TIMER]------|

|                                                |

|                                                |

|                                                |

|                                                |

|    input                             B         |

|-----|/|---------------------------[TIMER]------|

|                                                |

|                                                |

|                                                |

|    A.DN       B.DN                  OUT        |

|-----| |-------|/|-------------------( )--------|

   |        |

   |  OUT   |

   |--| |---|

-Eric

daba · Nov 10, 2010

Eric, that logic will not survive a mode change, or power-cycle.

The pre-scan will reset OUT, and could trigger other logic, while waiting for TIMER_A to re-establish OUT as a 1

Change the third rung to ...

XIC TIMER_A.DN OTL OUT

And add another rung

XIC TIMER_B.DN OTU OUT

Eric Nelson · Nov 10, 2010

daba said:
Eric, that logic will not survive a mode change, or power-cycle.

Excellent point, but was that a requirement in this application?...

-Eric

TConnolly · Nov 10, 2010

Bill,

Based on your timing diagram, what you appear to be looking for is something similar to what I use in cooling tower tanks which tend to be very turbulent.



Level_Switch

----] [----------------+----------+-(EN)

                       |TON TMR_A +-(DN)

                       +----------+



  TMR_A.DN

----] [----------------+----------+-(EN)

                       |TOF TMR_B +-(DN)

                       +----------+



 TMR_B.DN                Level_SW_DB

----] [-----------------------( )---

The TOF is not a commonly used instruction, so for the benefit of those unfamiliar with it, the TOF DN bit turns on as soon as the rung is true, then remains on after the rung goes false while the timer begins timing until ACC>=PRE. Whether this will work for you Bill or not depends on the application - keeping in mind that in a tank where I use this, some time factor is built in because it takes time for the average tank level to change and I'm using large presets. But hopefully it helps.

BillRobinson · Nov 12, 2010

Thanks Alaric,

I built something like that before except it was a slow opening relay and a slow release relay. The only problem with this is that you must make sure that:

TMR_B.PRE > TMR_A.PRE

Digital Debounce? Is Rockwell Just Plain Wrong?

BillRobinson

Member

rdrast

Lifetime Supporting Member

BillRobinson

Member

Eric Nelson

Lifetime Supporting Member + Moderator

daba

Lifetime Supporting Member

Eric Nelson

Lifetime Supporting Member + Moderator

TConnolly

Lifetime Supporting Member

BillRobinson

Member

Similar Topics