Archived Forum

The systems - both emonTx and emonPi - sniff the voltage at startup. If there's a sensible voltage there, it uses it. Otherwise, it assumes 230 V.

(If you're in the UK, although the 'nominal' voltage is 230 V, the centre voltage is 240 V and is likely to remain so for the foreseeable future, because to change it would mean changing almost every distribution transformer on the network. It's highly probable that your voltage will always be 240 V ±6%, and very rarely venture into the 230 V - 10% area.)

Your optical sensor is by definition correct, but only if it does not miss pulses and (probably) only when you are not exporting power (because most meters - there are exceptions - don't flash the LED when exporting, and how do you discriminate between an import flash and an export flash? ! ! !).

Your emonPi (Node 5), without a voltage input, assumes 230 V so it is the emonPi that would be reading low if you were using it and the emonTx should probably be closer to the true power.

But having said that, the calibration uses the nominal component values, so under very worst-case conditions it could be wrong by a significant amount. There's a Building Blocks article that explains the sources of error and their effect.

What I suggest, if you have a programmer, is to calibrate the emonTx according to the instructions in Building Blocks. If you don't have a programmer, then you can calibrate inside emonCMS using the times "×" operator on the inputs before you log to feed. (Voltage and power are independent at this point - the multiplication is done inside the emonTx.) You'll need to derive power from your meters by timing the pulses to give you the pulse rate = energy rate = power.

If you come to use the current inputs on the emonPi, I'd suggest you probably need a second ac adapter unless the emonTx is physically close enough to share, as you'll only be able to have apparent power if you use the emonPi current and the emonTx voltage multiplied together.

Thanks for that. Just wondering then, what is the AC-AC input for "out of the box" then? From the suggestions to calibrate stuff it sounds like the input voltage is not used for this.

So, in a related scratching of head, when is the phase stuff used from the AC-AC input?

I am now reading EmonLib.cpp - makes more sense if I see the code :)

I work so much better when reading code... I am getting it better now. It's the actually electronics / electricity that is not my area;

So, as a bit of a helper to me - are my grid and generation meters measuring power (sorry, probably the wrong terminology already) based on a nominal voltage or do they do the same as the emonPi/TX and watch the voltage as well as the current? Apologies if this sounds very basic, again.

If you are referring to the generation meters (for instance an Elster AC100) unit or similar, then these do both read the real voltage and current and then calculate the power from that. 

As they are used for billing purposes, they are normally within 1% accuracy (probably better!)

If you have another "energy meter" which doesn't have a direct AC power supply, and just a simple clamp then its going to be making assumptions about the voltage - and guess at 230 or 240v.

I've now found and used the multimap visualisation and once I flatten the meter power (in my head) to ignore the pulses above and below a median over time, the CT clamp and the grid meter are actually quite close - certainly single percentages overall.

I think this is just "operator error" and inexperience on my part!

Peter - I think a read of Building Blocks - "An introduction to AC Power" is called for.

You'll be relieved to know that "power" was the right word in that context: they measure power and accumulate the total energy. (Energy = power × time, or more precisely, the integral of power over time. Or the inverse of that: power is the rate of using energy. And there's a BB article on meters too.)

And as Stuart says, many/most of the consumer energy displays only use the nominal voltage or (perhaps slightly better) a voltage you tell them, so they are OK for showing you've left something on but won't give accurate readings all the time, even though by carefully tweaking the voltage ("calibrating") you might get reasonable agreement with your supplier's meter over a long period.

Thanks. I did read (OK, slow scan read) most of the building block pages but my brain works in a way that it makes no sense until I can apply the information. Now I've got an installed system it's starting to sync in - but with these gaps.