RFXCOM monitoring electricity and temperature

[CeeBee]

If anyone's interested, I've now got some monitoring of my electricity (generation etc.) and temperatures on the web (well it's temperature in the greenhouse, but solar thermal temps maybe coming soon). All 'rough and ready' at the moment.

Kit consists of:

RFXCOM ethernet receiver for 433MHz http://www.rfxcom.com/
RFXMeter module with 3xRFXPulse attachments counting flashes on electricity meter LEDs
An Electrisave/OWL whole house meter
Some Oregon Scientific temperature sensors

Software: xpl-perl running on Debian Linux.

Results at http://www.viridis.net/rrd/ (I said it's rough and ready, but only started doing this 3 days ago), and http://www.viridis.net/rrd/combined.html for hopefully 'nicer' graphs combining electrical generation/import/export.

Some explanation:

electrisave.09 is the Electrisave/OWL (house consumption, amps). The .1/.2/.3 components are for if you had up to 3 current-clamps attached. I've only got one, so .1 is identical to the summed reading, and .2 and .3 are zero.

The three rfxmeter things are from the 3 electricity meters. The graph units need a bit more work. Basically if you multiply the reading by 3.6, then the answer is kW. And if the y-axis labels include 'm', then that's rrdtool's representation of 'milli', so for example '100 m' is 360W.

rfxmeter.00f0 is from the _import_ meter
rfxmeter.01f1 is from the _export_ meter
rfxmeter.02f2 is from the _generation_ meter (PV panels)

thr128.08 is the greenhouse temperature (Oregon scientific sensor)
thr128.0c might be the outside temp, for for reasons known to itself only works if temp>10

Other 'thr' are sensors that briefly put in an appearance, but have stopped working, so please ignore.

I've also got an RFXSensor device, which is going to do temperatures for the solar thermal, but I haven't connected it up yet.

That'll do for now. I think it's great that this lot more or less works 'out of the box', albeit relying on me knowing a lot about computers/networks/software - lots of credit to Mark Hindess ('Beanz'), the author of xpl-perl. I've been talking to him about a few tweaks/fixes/etc.

[Jonathan]

Hi CeeBee,
This looks good stuff. Avoids the need for wires everywhere.
Can I ask, for the electrisave/OWL, I assume your using the rfxcom to 'intercept' the signal from the sensor - which protocol does it use? http://www.rfxcom.com/rfxcom.htm

[CeeBee]

Hi Jonathan

Yes - the RFXCOM picks up the signal from the OWL transmitter, so the OWL receiver/display isn't involved (similar with the Oregon stuff). Indeed the OWL/Electrisave doesn't seem to feature on that chart. It's mentioned on http://www.rfxcom.com/receivers.htm (and the 'Capabilities' link from there). Probably has its own protocol at 433Mhz (and is also known as cent-a-meter), but you don't get to see what magic the RFXCOM is doing to receive the stuff - once it's been received and decoded, it's fed to the computer in RFXCOM's own format.

I can imagine problems going forward with these non-RFXCOM devices, as their manufacturers come up with new ones with different 433MHz protocols - after all they only intend them to communicate with their own receiver. Even the OWL/Electrisave is largely replaced with a newer 'better' one - think that's the 'OWL CM119' mentioned on the RFXCOM pages. Depends how fast Bert (who I assume designs/programs the RFXCOM stuff, and he's even posted on here from time to time) can keep up. And I mentioned that it seems a bit 'flaky' as receiving my Oregon THR128 temperature sensors (only 1 out of 3 reliable at present) - must ask RFXCOM support for ideas.

For some reason, RFXCOM have recently become a lot more secretive about protocols - mention of non-disclosure agreements etc. There _used_ to be a lot more on their website. Not sure I understand, as you'd think the more people knew the protocol and wrote software, then the more hardware they'd sell. I assume _something_ was being ripped off, or distributed inappropriately.

Now I must get round to trying to set up the software on a Linksys NSLU/2 (Slug), so as to try to reduce the energy being used by the monitoring computer.

[CeeBee]

My RFXCOM stuff above now _is_ running on a Linksys NSLU2 ('Slug') so I've been leaving it on overnight, which I didn't want to do with the full-size computer.

Installation of Debian 'lenny' Linux on the Slug went smoothly. Its 32MB memory is a bit full with the software I'm running, and it's swapped about 16MB out to the 8GB USB stick that I'm using as the 'disk'. But there's no active swapping taking place - those 16MB must be stuff that's never (or hardly) needed again.

The data is still being sent over the internet to my hosted 'proper' computer, where the graphs are being drawn, and the web-pages served. These _might_ be possible on the Slug, but could push the memory use above reasonable limits.

I reckon my power-consumption involved in doing this lot is 30W or so. There are 5 power-bricks turned on which wouldn't previously have been, i.e. RFXCOM receiver, RFXMeter device, NSLU2, router, cable modem. Obviously some of these would be on anyway if I was wanting to use another computer.

Still must set up the RFXSensor to give some readings from the solar-thermal, and definitely create some prettier graphs (should be just different arguments to rrdtool graph...). And a reading on the web-page of e.g. solar PV kWh so far today would be good, since the graphs are just plotting power.

[whiteduck]

The RFXCOM devices look nice but there must be a cost effective way of picking up and decoding these 433MHz signals - the power monitoring devices are cheap to start with so all we are adding is RS232...

I found the following http://www.rfsolutions.co.uk/acatalog/Radio_Modules.html but don't have the knowledge of how to build something to plug into USB / RS232... any takers?

Cheers
Ian

[MikeM]

Hi,

My first post, prompted by this topic ! On & off I have been working on decoding the Electrisave/Owl protocol which is based on Oregon Scientific format and it has proved 'taxing' to say the least ! I am a hardware electronics engineer by trade and have an interest in low power radio, powerline comms, home automation and just recently home energy monitoring. This was re-invigourated by the installation of a condensing boiler this summer which has rewarded me with my biggest ever quarterly gas bill and now I want to look at my overall energy consumption.

To that end I bought an Electrisave and within moments decided I wanted to decode the protocol and somehow hook it into a PC. As previous posters have noted the RFXCom product will do just this but being an engineer with the right skills thought I could knock up a 'home brew' solution. The RF bits are easy (for me) but so far I haven't consistently cracked the convoluted protocol. I am more than willing to share what I have discovered with others though. The attraction is that the sensor/transmitters are dead cheap and there is no point reinventing this end. It is a shame the Electrisave receiver/display doesn't have a data output ?

Another half started project is a gas meter pulse counting interface which yesterday (after my enormous British Gas bill arrived) I decided I must finish ! This uses an photo reflective sensor and hooks into a PIC (been using these for years - sorry Arduino fans!)). My thoughts on this have varied from battery powered self contained logging to just sending 'pulses' over an RF link to a remote reader/logger/PC and I have noted the various comments on this forum as to the pros and cons of these solutions. The self contained solution does require a Real Time Clock which needs to be synchronised in some way whereas the simple pulse transmitter requires something running at the other end to receive and log the pulses.

I have also been studying various simple home networks and have been attracted to the Dallas One Wire system and have the bits for the DS1820 'Arbiter' and yesterday ordered (at great expense) some DS1820s ! During my summer boiler installation I 'wired' up all my radiators with a multiway cable and 24V power cables with the intention of adding sensing and zone control in order to economise on energy usage - the time has come to do it I feel !

Gentlemen, I am open to your guidance and suggestions ?

[CeeBee]

Hi Ian and Mike - glad to see a couple of new posters on this subject. Good luck with your DIY approach - it's not for me, since I do software, and haven't too much of a clue as to how to do electronics. So for me, the fact that Bert at RFXCOM had done it already was the right approach.

That said, my Electrisave isn't actually involved in my 'nicer' graphs at http://www.viridis.net/rrd/combined.html - since I already had real electricity meters, which hopefully are as good as 100% accurate, with flashing LEDs which could be counted, then it seemed best to use them. [Also, a 'current transformer' sensor would have sensed both my import and export as the same thing, which would have confused the situation somewhat.]

If you're going to pursue the Electrisave/OWL, bear in mind that I believe there's a newer model with different protocol (mentioned on RFXCOM site). Also someone seems to have deciphered the protocol before, see http://www.ukha-archive.com/ml/ukhad/2007-jan/msg00653.html.

Or another possibility is the CurrentCost meter which spits out XML data from a serial port. I think you can get one free from Southern Electric (and other Scottish & Southern suppliers) by joining the 'better plan' tariff.

[MikeM]

Thanks CEEBee. A Mr. I.C.B.Goodhew did report on the UK Home Automation forum that he had deciphered the Electrisave protocol but didn't say how he had actually done it ! The raw (RF) data is Differential Manchester but I just can't make sense of it  ! Other published information refers to the early Oregon Scientific format. Other posts in this forum have commented that the RFXCom info on this protocol has been removed.

You have however set me off in a new direction with the CurrentCost meter that does all that I want and at £50.00 ish seems a great deal. I am going to order one from the Ebay store. Nice software for it too ! Thanks.

Also got me thinking how I could use this device to 'read' my gas meter !  Not so daft as I could convert gas meter revolutions per minute into kWH -presented to the CurrentCost as an analogue voltage into the port where the current clamp  plugs !!

[Ivan]

Ceebee - Looks really good. Apologies if I've missed the answer, as reading threads in this section of the forum is a bit like reading foreign languages to me. My question is: Is the OWL able to differentiate between imported and exported electricity? The reason I ask is that the power meters I've seen aren't able to tell which way the power is 'travelling', so you get a positive power reading regardless of whether it is imported or exported power.

[whiteduck]

Thanks CeeBee and Mike!

I think you are right Mike - the current cost option is probably the easiest... until we see a more "open" standard decoding these protocols looks too much like hard work!   

Roll on the "VHS vs Betamax" battle of Home Automation protocols and then the fridge will understand what the microwave (and currentcost) is gabbling on about on 433Mhz and tell the hoover to keep the noise down... 

Cheers
Ian

[KenB]

MikeM, List,

Exactly 6 months after your original post - but I have been a bit busy recently.

Back in 1999, I developed a wireless gas meter pulse counter, for meters fitted with the R5 pulse output - basically a reed relay whose contacts closed every time the index wheel for 10 litres of gas passed by.

The microcontroller had to do no more than count the pulses, retain the accumulated total and form a short wireless packet that it broadcast over a 433MHz link every 1 to 15 seconds.

The low power micro ran off a 32kHz watch crystal which kept the power right down.  It was designed to last 10 years from an AA size lithium thionyl chloride cell, and so had a mean current consumption of just 25uA.

Later on I devised a photoreflective adaptor that fitted onto the index dials when the gas utility swapped out my old meter for a more modern one. 

The problem with photo reflective systems is that you have to keep the LED illuminated which will take at least a couple of mA - thus shortening the battery life by a factor of 100.   

If however you know how many mS the reflective digit takes to pass the reader at the maximum flow rate, you can devise a pulsed LED system that can get your LED mean current down a fair bit.  You could also use a knowledge of when your boiler is on and off to control whether the LED comes on - though this is a little more complicated, and no good if its a leaking pipe or your gas-oven that's using all the gas.

It would also be possible to measure the time between consecutive pulses,  - this could be used to calculate the gas flow rate in litres per second, and hence the burn rate or the power consumption of the boiler/cooker being investigated.

The key to the design is keep the pulsecounter/transmitter as simple as possible and put the bulk of the intelligence into the receiver, or further down the chain in the PC.

Have you made any further progress with the Electrisave protocol.  I know that a certain Iain Goodhew managed to crack it for the Home Automation BB.



regards,



Ken

[MikeM]

KenB,

Great minds ! I sort of lost interest in this project but strangely picked up my notes just a week ago - as the nights got colder and I considered switching on the (gas) central heating at enormous cost !

I too built a reflective sensor but every 250mS wake up PIC , switch on IR LED, monitor IR input, count, save to EEPROM then sleep to save power. My original plan was to log time and actual count as well then dump data via RS232 every so often.

I am now considering just counting pulses and collecting data at regular time intervals using a stand alone logger with Real time Clock to separate the counting and RTC/logging functions that consume power. This also avoids loss of count during data collection.

In answer to your question I gave up with the Electrisave protocol - just couldn't hack it ! I could dig out my notes if you were interested in persuing it.

I ended up buying an Efergy electricity monitor which I highly recommend - it works well and the PC data collection is neat. Pity they don't do a gas conversion kit !!

[ericw]

Ken,
Is something like this bright enough to give you a usable signal?

http://www.betalight.com/index_product.htm

[StBarnabas]

Interesting stuff!
spent a few hours today working on EricW's PIC one wire slave. Still the same problem device seen and code read but problem with data read. Anyway will be in the land of the data boys/girls for the next few months+. So very interested..

[KenB]

EricW,

Not sure about the Betalight glowing device - no experience of them. 

I had an idea that you could use some cheap fibre optic filament (Avago fibre optic evaluation kit from Farnell 327750 has everything you need to make a link) and site the LED transmitter and the photodiode receiver remotely from the meter - to avoid electrical connection.  The kit's supplied with 5m of fibre optic - so you can get 2.5m away from the meter - which is well enough away to satisfy the ATEX regulations.

http://uk.farnell.com/avago-technologies/hfbr-0501/fibre-optic-evaluation-kit/dp/327750?Ntt=327750

We use these kits at work for opto-isolating our serial data and they will run at up to 1Mb per second.  The wo ends of the fibre could be glued together to point at the reflective digit on the meter dial.   I've managed this with an opto-reflective sensor  - so the fibre optic which uses visible red light ought to work OK.  I'll have a quick experiment one evening.

So for £15.50 for the fibreoptic kit and £13 for the Arduino or PIC dev-kit, you could have a fairly lowcost solution for about £30 +VAT - and if the man from British gas complains - tell him confidently that there really is no electrical connection to the meter.

Lowest cost Arduinos from these people in Isleworth

http://www.nuelectronics.com/estore/index.php?main_page=index&cPath=1&zenid=56f28d9dc4033174fc08a74230ace969



Ken