EZIO2x4 Debounce

[Anonymous September 5, 2012 at 9:26 pm Post count: 4]

#40344 |

I am using inputs 1 and 2 on the EZIO2x4 to monitor the status of our two garage doors. They are working well except that each time the doors move the relays are naturally bouncing a bit and I get a report of three status changes instead of just one. I have the “Debounce the Inputs” checkbox checked in the config register using the SimpleHomeNet Utility Suite V2.02e but I still see three status changes for each movement.

Question:
Does the debounce algorithm actually work?
Does it work for the opto-isolated inputs AND the analog/digital inputs?
What does the debounce algorithm look like? How does it function?

I have moved this circuit from a Z-wave low-voltage input device to the EZIO2x4 because the Z-wave was not as reliable as I required. Everything is exactly the same except for the input device and I never saw a bounce problem with the Z-wave device. I am using these garage door contacts from Smarthome. http://www.smarthome.com/7455/Garage-Door-Contacts-for-Closed-Circuits-SM-226L/p.aspx

[Anonymous September 5, 2012 at 11:58 pm Post count: 1001]

#43847

This is not likely a bounce in the magnetic reed switch. Can you post the Insteon traffic that is being evaluated for multiple Input state changes? What type of software/firmware is being triggered by the Input state change? Insteon Group protocol generates multiple Group related messages each of which could be interpreted as a state change. Also message retry can appear to be multiple messages.

[Anonymous September 6, 2012 at 12:37 am Post count: 4]

#43848

Here’s what I see in the SHN utility traffic monitor page.

4 from ID:1B.03.FE flags:8B cmd1:27 cmd2:2
4 from ID:1B.03.FE flags:8B cmd1:27 cmd2:3
4 from ID:1B.03.FE flags:8B cmd1:27 cmd2:2

That’s all. This is that I get when I move one of the garage doors so that the magnet and the reed switch are no longer close

I am using Homeseer v2.5.0.58.

I have no groups defined in any of my Insteon devices. Everything is recieved by Homeseer who makes all the decisions based on my triggers and events.

I’ll hook up my oscilloscope to the input pins of the EZIO to see what the relay chattrer looks like.

[Anonymous September 6, 2012 at 12:57 am Post count: 1001]

#43849

Do you have pullup resisters on inputs 3 & 4 to keep them from floating?

[Anonymous September 6, 2012 at 1:39 am Post count: 4]

#43850

yes, both inputs are pulled up with separate 10k resistors

[Anonymous September 6, 2012 at 5:34 pm Post count: 1001]

#43851

Use the Trace PLC/PLM Raw check box on the Insteon Traffic tab. This will display the raw PLM message which has a little more detail.

I ran the tests below using a mechanical switch to turn Input 1 On and Off. No Debounce option in effect. A mechanical switch is more likely to produce bounce than a reed switch.

I also noticed that in the three lines of your posted trace Input 2 is On, then Input 1 & 2 is On, then Input 2 only is On. Without knowing the context of the trace it looks like Input 2 turned On, then Input 1 turned On, then Input 1 turned Off. The trace does not depict a single input bouncing.

From my test

0 :11 0250 0CAE5F 070301 8B2701
0 from ID:0C.AE.5F flags:8B cmd1:27 cmd2:1 Input 1 On
0 :11 0250 0CAE5F 070301 8B2700
0 from ID:0C.AE.5F flags:8B cmd1:27 cmd2:0 Input 1 Off
0 :11 0250 0CAE5F 070301 8B2701
0 from ID:0C.AE.5F flags:8B cmd1:27 cmd2:1 Input 1 On
0 :11 0250 0CAE5F 070301 8B2700
0 from ID:0C.AE.5F flags:8B cmd1:27 cmd2:0 Input 1 Off
0 :11 0250 0CAE5F 070301 8B2701
0 from ID:0C.AE.5F flags:8B cmd1:27 cmd2:1 Input 1 On
0 :11 0250 0CAE5F 070301 8B2700
0 from ID:0C.AE.5F flags:8B cmd1:27 cmd2:0 Input 1 Off

From your post

4 from ID:1B.03.FE flags:8B cmd1:27 cmd2:2 Input 2 turned On
4 from ID:1B.03.FE flags:8B cmd1:27 cmd2:3 Input 1 turned On
4 from ID:1B.03.FE flags:8B cmd1:27 cmd2:2 Input 1 turned Off

[Anonymous September 10, 2012 at 9:22 pm Post count: 4]

#43852

I finally got around to looking at the bounce of the garage door contacts with my scope. Found upon the first movement of the door a 100ms impulse that cuased the contascts to change state immediately. This was followed by another 2ooms of chatter before it finally got stable.

Since it appears that the SHN EZIO device does not debounce the inputs I debounced with an external RC circuit. Everything works just fine now. The EZIO device only reports one opening or closing.

[Author]

Posts