Edge Detection Circuits

[GeePaw]

Folks...

Here are three rising edge-detector circuits.

The first two came from the archives, and were fairly old. They're meant to detect a rising edge: when the switch goes from red to green, the output goes to green. All three circuits do this. The top one sends a single short pulse of 1 tick. The middle and bottom let you set the duration of the pulse using the filter's timing.

But.

The top and the middle have a problem: they also detect a *falling* edge, sending a tiny blip pulse when the switch goes green to red.

This is not a problem if you're using the output to control most hardware, a sweeper or an aquatuner, let's say. But if you're using it to control the reset port on a  memory toggle or the conveyor meter, you'll get a reset both on the rising and the falling edge, which likely isn't what you wanted.

It took me an embarrassingly long time to figure out the issue: the gates before the XOR take time to react, and on a falling edge, for one tick, the upper XOR input is green and the lower one is red, so the XOR fires green. The fix, synchronize them by inserting that ribbon-reader, which takes the same amount of time as the filter does to pass its signal along.

On the one hand, maybe this will help someone else. On the other hand, I absolutely suck at hardware logic: If someone uses a simpler and more compact perfect single-edge detector, please show it off.

Seeya -- @GeePaw

[Saturnus]

I think the below from an old post of mine are the ones most people use I reckon.

For a positive edge detector you can also use a counter set to 1 with output and reset connected which obviously also makes a negative edge detector by having a NOT in front of the counter.

 

[gabberworld]

2 hours ago, GeePaw said:

 

On the one hand, maybe this will help someone else. On the other hand, I absolutely suck at hardware logic: If someone uses a simpler and more compact perfect single-edge detector, please show it off.

 

one thing you are wrong at here. that is a software logic what we deal in here, you may not experience like that in hardware or it may act little differently

[Tuode]

7 hours ago, Saturnus said:

Aren't these builds also affected by the issue mentioned above? Sorry if this is a stupid question, for I don't play around with automation very often

[nakomaru]

1 hour ago, Tuode said:

Aren't these builds also affected by the issue mentioned above? Sorry if this is a stupid question, for I don't play around with automation very often

On the contrary - they specifically exploit the fact that gates take 0.1 seconds to update. That's part of the role of the NOT gate on the left - to allow a double green or double red signal to the center gate for only a moment.

[Sasza22]

Personally i use a buffer gate with a XoR for one edge and a filter with XoR for the other, then both a filter ande a buffer for both edges. Additionally a 0.2s filter at the edge to avoid false positives. There are probably better setups than that but this one seems the easiest for me to understand.

[Occam Blazer]

Top to bottom: Rising edge detector; falling edge detector; toggle detector. (two versions each)

The signal counters' outputs are connected to their reset port. Switch them to advanced mode and set to 1%.