How much time does a single transmission take?

[sravankb]

Hello,

I was wondering what is the time taken to transmit one sample takes. For example, if the sample interval is 5 seconds, does the actual transmission take 10ms? What does it do for the rest of the time?

Also, how much energy does a 3202 node take to transmit a single sample?

Thanks, 

Sravan 

[Smootastic]

Hi sravankb,

The node should sleep during the time that it's not actively measuring, transmitting, executing custom code (VI on the node) to preserve power.  Transmission of the data should be fairly quick (specced at 250 kbits/s in the manual: http://www.ni.com/pdf/manuals/372775c.pdf ).

The manual also mentions that you'll your power usage at 1 S/s intervals is about 13 mW.  Note that power requirements change when using router and programmable nodes, or lengthening the sample interval.

[sravankb]

Thanks Smootastic,

So, for example, if I transmit a single sample from the 3202 every five seconds, then only 64 us (16 bits/ 250 kbps) of this time is the actual transmission? How much energy does it consume the rest of the time? 

--Sravan

[sravankb]

The reason I am asking these questions is because I am trying to make sense of a simple test that I am carrying out -

I have made a battery pack of the 4 AA battery source, and connected this in series with a resistor, whose terminals are also connected to the AI0 input. I am essentially trying to find the voltage across the resistor using a 5 seconds sample interval. This will, in turn, tell me how much current is being supplied by the batteries. (Voltage across resistor / resistance ) Since I know the voltage provided by the battery pack, I can find a rough estimate of the power consumed. Does this seem reasonable?

[sravankb]

Sorry for the triple-post, but I have a few more questions - 

• What are the different states a node takes during every sample interval? For example, during a 5 second interval, it goes through - idle -> sensing -> transmission -> idle. By sensing, I mean it is simply measuring the analog input, not transmitting it.
• How long does each state last in a "n-second" sample interval?
• How much power does each state consume? 

Thanks,

Sravan 

[Smootastic]

Hi Sravan,

In every measurement cycle (not counting sleep and heartbeat intervals), you have the following states:

 

In total, each cycle will consume about 12 mJ of energy, the majority of which is consumed by transmitting the data via radio.  

As far as sleep is concerned, it can be estimated to about 7.5 mJ of energy consumed, assuming a 60 second interval; this value will be less depending on the sampling interval (plus the 3202 isn't asleep the whole time, since you factor in measurement cycles).

Timing-wise, the measurement cycle period should be the same regardless of the sampling interval so that the module sleeps as often as possible to preserve energy.  I don't know exact timing intervals, but I've seen about 20-40 ms for the entire measurement cycle. 

Hopefully this answers most of your questions, let me know if you want to clear anything up. 

[sravankb]

Hey Kyle,

I tried finding these specs from the 3202's spec sheet, but couldn't find the power specs. Could you provide me a link or the document that contains the images/info from your post? It would be very helpful.

Thanks,

Sravan

[Smootastic]

Hi Sravan,

These pictures were given to me by my colleagues and aren't found in the documents.  Do you have additional questions? 

[Malcolm]

Hi Sravan,

The data shown is from an internal R&D white paper.  We are working to expose some of this information more publicly, but we need to scrub it for accuracy and readability.  Unfortunately, it may be a while before we have the customer visible white paper ready.  I also hope we can provide a software tool which can help predict battery life based on a more granular sample interval input.  This should be a lot better than the life at 1 min and 1 s in the manual.

Please let me know if there are specific pieces of data you need and I can likely provide that.  Also, I'm curious what your end goal is.  Are you trying to validate that your node will run for three years?  Are you running at a rate more in the middle and need to determine the expected life?  Also, are you running custom VIs on the node with LV WSN?

Regards,

Malcolm Borgendale

Group Manager, WSN R&D

[Kevin_H]

Hey Sravan,

I wanted to provide some additional details to Kyle's post above. 

This is on average what we have seen with a 3212 node sampling at a 60 second sample interval:

Sleep Duration: 58.774 seconds

Sleep Power: 7.41 mJ

Acquisition Duration (All 4 channels): 1150 ms

Acquisition Energy: 2.84 mJ

Transmit Duration: 55 ms

Transmit Energy: 9.17 mJ

Receive Duration: 21 ms 

Receive Energy: 1.83 mJ 

 

Additionally, if your sample rate is above 61 seconds, then you will run into the Heartbeat Interval which fires after every 61 seconds of inactivity.  Below are average power measurements for the heartbeat interval:

Transmit Heartbeat Duration: 55ms

Transmit Heartbeat Energy: 7.79 mJ

Note: Each heartbeat will then also need to account for receiving that a response from the gateway (Receive information from above).

Taking all that into account, the graph below shows estimated battery life for sample intervals up to 300 seconds.  The dips in the graph represent a sample interval that is just larger than the heartbeat interval, which essentially means you are transmitting twice in a very short period.

Hopefully you find this information helpful.  Please let me know if I can clarify anything for you.