DIY Smart Meter?

[Dave Liquorice]

The CurrentCost one

has a TTL serial output from the display unit. You can either interface to a serial port (be that RS232, 5 V, 3.3 V, etc) yourself or get a TTL to USB serial convertor. The sensor is a clip on current transformer to a 3 x D cell battery powered unit that links to the display. Needs a bit of care fitting around one tail but it's capable or being pretty accurate. The data stream is XML and easy to decode, containing instantaneiouos power readings and history.

I have the TTL to USB convertor plugged into a HP Microserver and a Perl script that logs the instananeous min/max and calculates the average power level over 5 minute intervals. Bit of PHP and GNUPlot produces a web page for any given day.

The 5 min plots can spot things like SHMBO'd forgetting to switch the iron off or energy wasting gadgets like the "keep warm" kettle or filter coffee maker stewing the coffee...

A Pi Zero is perfectly capable of the data logging and web page stuff using the XML. Would need to convert the TTL from a CurrentCost to

3.3 V Pi GPIO levels. It would be trickier to to interface directly with a suitable current transformer as Pi's don't have ADC's. Aurdino's do, so a bit of signal conditioning from the CT and away you go.

[Roger Mills]

Is that showing true watts, or simply volt-amps by measuring the current and multiplying it by nominal voltage without taking any account of power factor? I notice that there's also an optical version which counts flashes from the meter - which should give true watt-hours. Does that version have the same logging ability?

[Andy Burns]

I got one of them free from e.on (together with USB/serial converter cable)

You don't get billed for watts, just volt-amps

[Andy Bennet]

No, you get billed for volts * Amps * Cos(phase angle between them). If you stick a large lossless capacitor across the mains and draw loads of current it will not cost you anything.

[Dave Liquorice]

It has no means of measuring the voltage unless there is a sneaky means of getting that via a current transformer. When I last checked what it was saying against the bills it was less than 1 kWHr out over a days 20+ kWhr consumption.

None of these things are intended to be an accurate check meter just a reasonable indicator of consumption. It could be scaled in bananas, more bananas - bad, less bananas - good. B-)

It's just an optical sensor and sender to the display unit. It's the display unit that does the logging and history etc.

[Andy Burns]

isn't it powered from a wall-wart? if that is a straight transformer type with internal regulator, it would have a scaled-down voltage reference

[Dave Liquorice]

It is.

But judging by it's size and weight its not got much iron in it...

[Robin]

Not with any I've seen. Some suppliers may offer them. But AFAIK the general rule is still that anyone wanting to access data from the meter direct has to (a) buy a consumer access device and (b) get their supplier to give it access to the meter's Zigbee network.

Meanwhile IIRC the Dutch simply provided a socket on smart meters so consumers could plug in and get their data - including of course plugging in something that connected wirelessly to their home network.

[The Natural Philosopher]

Absolutely wrong, you get billed for watt (hours).

The electricity meter is a very curious device. The mechanical one...

"The disc is acted upon by two sets of induction coils, which form, in effect, a two phase linear induction motor. One coil is connected in such a way that it produces a magnetic flux in proportion to the voltage and the other produces a magnetic flux in proportion to the current. The field of the voltage coil is delayed by 90 degrees, due to the coil's inductive nature, and calibrated using a lag coil.[16] This produces eddy currents in the disc and the effect is such that a force is exerted on the disc in proportion to the product of the *instantaneous* current and *instantaneous* voltage. A permanent magnet acts as an eddy current brake, exerting an opposing force proportional to the speed of rotation of the disc. The equilibrium between these two opposing forces results in the disc rotating at a speed proportional to the power or rate of energy usage. The disc drives a register mechanism which counts revolutions, much like the odometer in a car, in order to render a measurement of the total *energy* used."

Obviously *electronic* meters can do far far more also than just measuring current, as well! Rapid sampling of voltage and current enables true energy consumption to be calculated.

[The Natural Philosopher]

Doesn't need to have. Not to just provide a voltage proportional to the mains., Lots of turns of very thin wire on the primary of a tiddly iron transformer will do nicely...

[Martin Brown]

Not dissimilar to an oil tank Watchman and the batteries typically last a few years so it isn't much of a hassle.

Official Watchman replacement batteries are a ripoff though but tear down instructions for rep[lacing one are available on Utube.

[Dave Liquorice]

Only a two wire connection, at 3 V, to the display unit. No "sense" wire. It looks just like any other generic wall wart supply. No CurrentCost branding or "use only with...".

[DJC]

My Watchman. instruction leaflet included battery replacement instructions. Just a standard CR2430 from eBay.

[Martin Brown]

The older Watchman models used a custom built set of AA batteries in a copper pipe and sold for an extortionate amount of money eg.

Appear to be discontinued so DIY is now the only way to do it.

[alan_m]

I replaced a battery in a sensor a few years ago. The property was in a rural village where the majority of people have an oil or LPG tank (no mains gas). Those with oil usually have the same type of level sensors so probably a reasonable market for the type of coin cell they take. I had to resort to Ebay as the battery seemed not to be stocked by any hardware/agricultural store anywhere in a wide locality, even though most stocked other types of coin cell.

[Harry Bloomfield, Esq.]

Martin Brown has brought this to us :

Not that I have heard, mostly it seems to be a matter of hacking into the data sent to the indoor display.

[Harry Bloomfield, Esq.]

It happens that Dave Liquorice formulated :

The modern ones switch off after 20 minute-ish. Pointless them remaining on after 20 minutes, the coffee is undrinkable.