On/Off Mechanism

[sravankb]

Hello,

I would like to know if there is any sort of a programmable on/off mechanism available for the nodes. For example, I would like a node to be off for a few days, turn back on and sample at 1 Hz for a few hours, and then turn off again for the next few days.  Is it possible to do this? (When I say, it turns off, I am not referring to the sleep mode, where it uses power by sending a "heartbeat" signal every 61 seconds)

Another question that I have is -  if I cut the gateway for x minutes, and then turn it back on again, will the end-nodes/routers connect to the network automatically? Is there an upper limit to x?

Thanks,

Sravan 

[Online_Courses · Accepted solution]

To answer your first question, there is no way to have the WSN node be powered off and then power itself on again.  The only thing you can do is have it sleep.  I you, however, have a separate system that will poer the node on and off, but that is a system you would have to build.

The second question is answered in this KB: http://digital.ni.com/public.nsf/allkb/88B2957808343185862575EC00120872?OpenDocument

"After failing to join a network, an end node will sleep before trying to reconnect. This is done to help conserve power. The node will sleep an increased amount of time between failed attempts. The nodes use the Fibonacci sequence (1, 1, 2, 3, 5, 8, 13, 21, 34, 55) to determine wait times. The nodes can wait up to 55 minutes between connection attempts."

I hope that answers your questons.

[sravankb]

Hello,

Yes. Thanks, that does answer my questions. I would like to know about a few additional things as well: (when I say "usual activity", I mean a sample interval of 1 minute)

1. Does the power used by the nodes/routers increase when they detect a lower link quality? If yes, how large is the change from their usual activity?

2. Do they consume more power than usual activity when attempting to connect to the network? 

3. Finally, you mentioned that an external on/off mechanism can be set up for the nodes. Is there any industry-standard technique for doing this?  We were hoping to use some sort of RFID-based system to turn them on/off.

Thanks for the help,

Sravan 

[GoCougs]

1. Does the power used by the nodes/routers increase when they detect a lower link quality? If yes, how large is the change from their usual activity?

      -No, we dont' dynamically change the transmit power;  however, due to packet losses there may be an increase in the packets transmitted.  This would cause an increase in power usage because the node will send the signal until it receives an acknowledgement that the gateway/node has received it.  The amount will be completly random, depending upon the interference that causes the weak signal.

2. Do they consume more power than usual activity when attempting to connect to the network? 

         -In order to connect to the gateway there is more data transfering back and forth than a simple transfer from the acquisition.  The main power hog is the radio and there is much more radio communication during the connection process than during 1 minute data acquisition.

3. Finally, you mentioned that an external on/off mechanism can be set up for the nodes. Is there any industry-standard technique for doing this?  We were hoping to use some sort of RFID-based system to turn them on/off.

         -You could use a relay switch, but that would have to be externally powered by a separate device.

Mostly, I wanted to bring to your attention that the battery power in a WSN under normal activity will last for 3 years.  Do you need a longer period than 3 years for battery life?

[Kevin_H]

Hey sravan,

I just wanted to add a few more notes.

-When connecting, an end node essentially uses the same amount of power as an always on router.  This is because the end node stays on to receive and transmit connection messages back and forth to the gateway.  Therefore, the connection process is more power hungary than a standard one minute sample interval.

-Could you give me some info on your overall application and what exactly you are trying to accomplish with the on/off mechanism?  I think there are some interesting things you potentially do with LabVIEW WSN for each of the nodes.  For example, you could create a LabVIEW WSN VI that makes use of user messages.  These user messages can be thought of as simple string commands.  You can implement logic on the node to enter different states based on what string you send.  For example, you could create logic that changes the node sample interval to 1 Hz when a certain command is sent.  You could then take your data, and send a command to change the sample interval back.  This would give you some programatic control of the events of the node.  In addition, the new NI 9792 can programmatically configure the node attributes from LabVIEW.  This feature requires you to be running the VI on the 9792, as there is no remote support for this feature yet.

Thanks,

Kevin 

[sravankb]

Hello Kevin,

Sorry for the late reply. The application we are trying to implement involves using the node for a few hours, and then turning it off (not sleep mode) and then on again after a few days for a few hours of activity, after which the cycle repeats. Our aim is to waste no amount of power when the nodes aren't supposed to gather any data. 

However, it seems that the nodes cannot be remotely turned off, which means now that we have to figure out a mechanism to do turn off their power when they're not needed.

Thanks,

Sravan

[Kevin_H]

Hey Sravan,

That sounds very interesting.  With the current hardware/software and the 61 seconds heartbeat, the only solution would be some sort of external on/off mechanism.  

If it were possible to manually adjust the heartbeat interval (say to something really long like days or weeks), would this be something that would work for you? Keep in mind the drawback to such a setup would be you lose responsiveness from a node.  However, for a deployed solution, this may be acceptable.

On the other hand, do you think you would prefer a physical on/off switch?  This of course would require a user to be at the node to turn it off when sampling is finished. 

Let me know your throughts.

Thanks,

Kevin 

[sravankb]

Hello Kevin,

Sorry for the delayed response. Yes, having a manually adjustable heartbeat signal would definitely help us. We would actually prefer a programmable on/off mode, though. A physical switch would not be too great for our purpose. 

Thanks,

Sravan