Electricity Monitoring - Which Photodiode?

[cliffwright]

Hi Folks,

Just reading the Gumbrell blog about power monitoring, I have a few questions;

1) can someone point me in the direction of which photodiode I need to achieve the same electricity meter monitoring as described here?

2) My electric meter is outside and I'm planning on utilising the digital input on an SLD module that is under my stairs (so about 6-10 meters between the 2 as the cable runs); Are there any issues of cable length to consider? Is there a specific cable type I should use? Was planning on using a couple of joined pairs of a Cat5 run . . .

3) Is there anything else I need to consider given I'm using the digital input of an SLD module rather than a dedicated ODI module?

Thanks folks

Cliff

[Karam]

1) Probably David is best to advise you on what he knows works, but if you are looking at a pulsing LED I would imagine a phototransistor such as SFH310 might do the job. The 'when lit' current needs to be >1mA (which of course depends on the intensity of the light source).

2) Not ideal length but should work. The main problem is potential for noise pickup. Shielded cable would be best with shield attached to IDRANet shield, if used, at the module connector or failing that to 0V at the module digital input. Otherwise if just using CAT5 cable then use a twisted pair not wires from separate pairs. You may additionally need to bolster the capacitive filtering at the digital input eg. 1uF between input and ground but don't go crazy as the pulses are probably quite short.

3) Unfortunately the dimmer modules don't support pulse counting on their digital inputs so you'll have to try and use one on another module such as PLH or DRB

[cliffwright]

1) Probably David is best to advise you on what he knows works, but if you are looking at a pulsing LED I would imagine a phototransistor such as SFH310 might do the job. The 'when lit' current needs to be >1mA (which of course depends on the intensity of the light source).

2) Not ideal length but should work. The main problem is potential for noise pickup. Shielded cable would be best with shield attached to IDRANet shield, if used, at the module connector or failing that to 0V at the module digital input. Otherwise if just using CAT5 cable then use a twisted pair not wires from separate pairs. You may additionally need to bolster the capacitive filtering at the digital input eg. 1uF between input and ground but don't go crazy as the pulses are probably quite short.

3) Unfortunately the dimmer modules don't support pulse counting on their digital inputs so you'll have to try and use one on another module such as PLH or DRB

Thanks Karam

I'll be using a lenght of C-Bus cable so hopefully won't pickup too much in the way of noise . . . we'll see I guess . . .

David - would appreciate the benefits of your experience if you're around?

[Gumby]

Sorry - was on holiday in a hole with sporadic connectivity...

err.. well, I *should* have carefully selected the correct device to use based on my knowledge of electronics and examining the characteristics of the ODI and guessing the output of the meter LED, but what I actually did is wandered into Maplin and bought the only visible light photodiode available ... can't remember the model number I'm afraid.

I put an ODI right next to the meter to avoid long cable runs. So you could extend your Idranet to the meter instead of trying to avoid noise pickup in photodiode cabling. Partly because I had a spare ODI... and partly because it started out as a complete experiment and I didn't want to worry about pickup and extra filters and screening etc.

[cliffwright]

Sorry - was on holiday in a hole with sporadic connectivity...

err.. well, I *should* have carefully selected the correct device to use based on my knowledge of electronics and examining the characteristics of the ODI and guessing the output of the meter LED, but what I actually did is wandered into Maplin and bought the only visible light photodiode available ... can't remember the model number I'm afraid.

Cheers - appreciate the help; that sounds like my kind of approach to electronics (but possibly with less electronics knowledge to start with)

By the same approach, I'm guessing this might do the job then?

http://www.maplin.co.uk/sfh2030-silicon-photodiode--2242

I put an ODI right next to the meter to avoid long cable runs. So you could extend your Idranet to the meter instead of trying to avoid noise pickup in photodiode cabling. Partly because I had a spare ODI... and partly because it started out as a complete experiment and I didn't want to worry about pickup and extra filters and screening etc.

My 'leccy meter is outside in a location that is pretty stand-alone in terms of Idranet needs so for me to 'site a dedicated module near it it would more or less have to be for the sole and total purpose of meter reading which, like you, this is a little experiment (a) to teach me a bit more about Cortex's uses and (b) see if once I've captured this data it's actually of any tangible use.

Given this, I was hoping to make use of a 'spare' digital input rather than fork out ~£50 for a module like an ODI.

Might give the 'long cable run' approach a go first to see what happens and then think about how I could 'work in' a dedicated unit more usefully.

[chris_j_hunter]

checking our new digital meter (*) just now, which we hope shortly to read in a similar way, I see there are two LEDs behind a small window that's marked 1000 impulses per kWh (ie: max. pulse rate could be 100 x 250 / 1000 = 25Hz) ... hmm, wonder which of the two LEDs we should read via the photodiode ??

we're still wiring the house, including for Idranet (using appropriately orange Cat-5e FTP - third box just started) - and (thanks to this thread) it looks now like we should be saving any off-cuts for sensor spurs !

(*) a Landys + Gyr type 5235A / 20mA-100A / 2W 230W / CI 2 2009 - with a bar-code & an LCD digital display, as well as the two LEDs


PS: have just found the datasheet, and it looks like the B version of our meter would have terminals for the pulse output :

http://www.landisgyr.com/za/apps/products/data/pdf1/5235_Technical_Specification11.pdf

unfortunately, that's not the one we have, but it's interesting that some meters out there could have terminals for this function & so not need a photo-diode to be added ...

OTOH, it also looks like the supposed two LEDs are actually one LED plus a photo-diode for two-way comm's using this protocol :

http://en.wikipedia.org/wiki/IEC_62056

ie: if we add a photo-diode, the utility company would be prevented from reading our meter with a hand-held device ...

hmm, that seems to put a spanner in the works !

[Karam]

By the same approach, I'm guessing this might do the job then?

http://www.maplin.co.uk/sfh2030-silicon-photodiode--2242

I don't think the photodiode will pass sufficient current when illuminated. A photo transistor would be more sensitive and fast enough for this application.

[Paul_B]

Hi all rather than start a new thread I thought I should update an old.

For those with an electricity meter that produces a LED "blink" for a certain usage, for example an Ampy Automation electricity meter (below), you may wish to consider the Enica Opti-Pulse LED sensor. At £24 it was a case of "plug and play" the sensor and connect with a Idratek ODI.

http://www.enica.co.uk/optipulse.html

WP_20130525_002.jpg

Hope that might help others

Paul

[chris_j_hunter]

Paul -

good find ... we've some QRI on the other side of the wall from our (external) meter cupboard, so it's tempting to order one !

Distance-wise, the run might be a bit under 2m - not sure how far the cable would run (website didn't seem to say), but hopefully it could easily be extended ... (?)

Chris

[Paul_B]

The cable that comes with the sensor is about 1m in length. I extended mine with my newly acquired soldering skills and used the shielded twisted pair CAT5 cable that I use for the wiring of Idranet. I used two pairs, orange / white, white / orange connected to the sensors red cable and blue / white, white / blue to the blue cable. Blue pair goes to 0V and Orange pair (red sensor) went to channel input on the ODI. The sensor has a shielding cable which I connected to the shielding wire on of the STP which links back to 0V on Idranet at the power supply.

Paul

[Karam]

For those wishing to experiment, I have conducted a quick test of the sensitivity of an inexpensive BPW85C opto transistor (Farnell ref: 1045524). This can connect directly into a digital input (or via an extending cable as described by Paul above) with the long pin (E) going to 0V and the short pin (C) going to the relevant digital input on the digital input terminal block. The sensitivity appears high enough to detect a reasonably bright red LED but unfortunately I can't vouch for how well it would work on a real meter LED since I didn't have access to one and don't know if they vary or are subject to some brightness standards. Some method of affixing the sensor and preferably occlude ambient lighting (eg. some thick enough black material drilled through for transistor and with ds tape on meter side) would also be required.

[chris_j_hunter]

Paul -

thanks, will take some covers off & check our routing first - IIRC, it would have to come through the same hole as our 100A feed into the house (!)

Chris

[chris_j_hunter]

Paul -

have just received our sensor - and it came without means of fixing (installation notes mention adhesive tape, but no sign of such on ours) - so, how did you do yours, please ?

Chris

[Paul_B]

Chris,

Sorry I posted a response last night but it seems to have gone "missing".

To me it sounds like they have forgotten to add the adhesive tape to your unit. On mine it came with an adhesive pad that covers all of the plastic area of the sensor (everything bar the cable), it then had a hole in the centre cut out which allowed the photo-diode to "see" the diode of the meter. I ripped off the protective paper, lined up the central hole with the LED on the meter and just pressed it home. Job done.

Paul

[chris_j_hunter]

Paul -

thanks, in which case we'll be dropping them an e-mail for them to send what's missing (especially since it was far from cheap) - we need to avoid damaging the meter, and Blu-tak might not have the longevity ...

we've pulse meters on major items, but the rest still adds up to quite a lot, so it'll be good to have the total, as well - will help no end with the analysis ...

Chris