Large installation thoughts ...

Hi people, has anyone been playing with deploying these systems as a bulk deployment in a fairly large area using the standard configuration?

 

I am looking at running a test in a 50 unit facility to really find out what issues I can find!

 

I would like to use wireless for simplicity of install - I can just mark each one with unit #, send the sparkies out and wait for the units to come on line ..

 

However, not only do I hate the idea of battery failures, etc, but also the potential problems with the communications, especially if I not only wanted to run a central emonCMS server (or cascaded group), but also want to run the emonGLCD in some or most of the units ...

 

Thoughts?

 

Or am I better to head straight down the path of PoE, run an ethernet shield with a full arduino shield (and not the TX - or maybe a hybrid) and be done with it .. then maybe run a wireless network for the GLCD modules (or again maybe have them on PoE) ....

Mattia Rossi's picture

Re: Large installation thoughts ...

Hi,

what kind of data are you planning on measuring?

If they include current consumption, then the battery issue shouldn't exist since you will probably be able (with some forward thinking) to install them in a place where you have power.

 

As for communications, it really depends on how far the base/bases are from the sensors. To be on the safe side I would probably go with and emontx shield / nanodeRF config, and then test whether I can send data back to the server/s with wireless. If not, running a cable to the nanodeRFs would be your best option as this way you would be able to at least spare the wired part for the lcd.

Also, depending on distances, you could try with some proxies that collect data from a group of nodes and forward them to other proxies until the messages reach the server (a bit tricky)

Again, what kind of data would you like to monitor ?

Robert Wall's picture

Re: Large installation thoughts ...

Wearing my systems/applications engineer's hat, I don't know nearly enough to offer any useful advice. We need to know the layout of your plant, distances, if the radio frequency is used by anything else in the locality, possible sources of interference and any other factors that might affect radio propagation. My feeling is you could end up with a very 'iffy' system unless the radio is 100% solid. If cabling does not present a major installation headache, I would go for the wired Ethernet option. It will cost more initially but stands to be cheaper on the 'buggeration factor'.

Mal Harwood's picture

Re: Large installation thoughts ...

Thanks guys ..

It is unlikely that the installation will include true power, but in some installs may include CT's for low accuracy current flow.

I wasn't too worried about standard radio issues as I have a few different frequencies available and can likely design a repeater system if required ... I was thinking more about unit id problems and potential to use a single server to receive from multiple separate transmitters ... looking at the code, it appears I could setup a unique network number and then have each send a device I'd ... But I can't find doco say what these magic numbers could be without looking deeply into libraries ..

I also can't see much about security .. so if I did go this path, what is to prevent one unit from accessing the feed from another .. May not really be a security issue as it wouldn't really matter .. But some authentication process would help with misconfiguration .

Cost is not really an issue at this stage as it is a theoretical problem .. Once I can find a solution, then that may become a concern ..

StuntMonkeh's picture

Re: Large installation thoughts ...

Like Robert said, its difficult for an engineer to give a definitive answer without a great deal more information.

Experience tells me that It is notoriously difficult to predict wireless communication performance within buildings.  It would be extremely naive to assume that you could send the devices out into the field and expect them to appear on the dashboard!

Another communication solution that has not been mentioned is a wired serial based comms system using an appropriate protocol.  There are a number of pros and cons to this method but it really depends entirely on the situation and again there just isn't enough information to go on.

Anything is possible with enough time and/or money.

At some point you will be limited on one of those factors.

You need to sit down and define what you want to achieve.  In reality your own detailed specification.

From this point you will be able to assess the options available to you.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.