[rrd-users] Energy Metering

Alex van den Bogaerdt alex at ergens.op.het.net
Tue Sep 4 18:40:58 CEST 2007

On Tue, Sep 04, 2007 at 05:02:42PM +0100, Simon Hobson wrote:

> >I may be asking for help myself soon; I'm thinking about a
> >webcam, a led (for illumination) and some ocr software to
> >read my own meters.  That, or a smarter idea on how to automate
> >reading those meters.
> I assume then that you still have the old 
> mechanical "Ferraris Disk" meter. I would look to 

Same brand and remarkably similar to this one:

It has two tariff registers.  I think I have a CM140.

> see if there's any way you could focus a sensor 
> on the disk which will have some markings on it - 
> IIRC there are lots of stripes around the edge 
> (high frequency, high resolution) and some larger 
> markings (such as one/revolution). Ideally you 
> would like to mount a reflective sensor close to 
> the disk surface, but your utility supply company 
> will not be amused if you break the seal and 
> start fiddling with their meter !

A big no-no indeed.

> If you could achieve this then you'd have just 
> the same system as the OP - one pulse every X 
> number of watt-hours.

480 revolutions per kWh.  But I would rather look at the numbers
and use those.  I could use the dark marking as a trigger to read
the numbers. This way if I miss a (couple of) revolution(s) there's
still reliable data to work with.

> WARNING - DO NOT try this without adequate knowledge/training !

Once upon a time I was shown, by someone which should have
known better, that those three strips were too close together
to be carrying current. They had to be ground.  I stepped back
3 meters (10 feet) and still my shirt got burned when he put in
his screwdriver to demonstrate...

Nice demonstration :-)

I usually prefer hands-on experience but in this case I
instinctively knew hands-off experience is better. Thanks anyway.

[snip replacement or addition]

Biggest problem there seems to be how to reliably detect which
of the two registers is counting.  I write _is_, not _should be_.

If you feel this is interesting enough for the list, please do
continue the discussion here.  If not, feel free to contact me
directy.  And of course feel free to stop the conversation if
you feel you have no more to say.

Alex van den Bogaerdt

More information about the rrd-users mailing list