I have just setup a new system having received all the bits needed yesterday.
EmonTX V3 ready built with one CT sensor
Raspberry Pi running EMonCMS with hard disc
I haven't built up the EmonGLCD yet and I am running the EmonTX from pen cell batteries at the moment. The 9V AC transformer still has to be connected.
I have been running a Current Cost setup until now and thought it was fairly accurate, possibly reading a bit high. The reading I am getting from the Emon setup seems very high. At the moment it's showing 1212Watt and the Current Cost is showing 920Watt. Thats nearly 300 watt difference. Checking the actual amount consumed over an hour on the household meter it's more or less agreeing with the Current Cost reading. I am wondering if I have missed something in the setting up process. I know I haven't got the 9V AC power connected yet, but neither has the Current Cost unit. Any suggestions?
Re: Accuracy question
The emonTx V3 sketch assumes a system voltage of 230 V if it can't see an ac input. As you appear to be in the UK, your nominal system voltage is still likely to be 240 V, so that immediately accounts for an error of 4% or so but in the wrong direction. You're 30% adrift. The c.t. itself has a 3% tolerance, and the burden resistor 1%, and the analogue reference a further 2%. So you should be within 6%. Can you check the 3.3 V rail on your emonTx? That is also the reference for the ADC and if that is low, then the current measurements will be correspondingly high. I assume you are using the shop c.t. in CT1.
What load did you have? Without the voltage reference, the emonTx can only measure current and hence show an apparent power, with assumptions about the voltage, as I've mentioned, and the power factor. If the power factor of your load is particularly bad (it would have to be about 0.7) that would go a long way towards explaining the difference (except we don't know what the CurrentCost unit's assumptions are).
Re: Accuracy question
Thank's for the pointers. I am going to connect up the AC adapter tomorrow as I am think that the error may be due to voltage changes. I am still getting to grips with the system and really need to do some more definitive tests. One thought. I have the CT which incidentally is the shop one, on the live line and the Current Cost CT is on the neutral line. Both CT's are fairly close to each other. Is it possible for them to have some form of crosstalk and interfere with each other.
So the the only test was to watch the house meter consumption over an hour with what looked like a fairly constant load of about 1kW. It was then that I noticed the big discrepancy. However since then (a few hours later) both were showing the same reading. So I am even more confused at the moment.
What I would like to do is run for a day and make a note of the consumption on the house meter and find a way to compare it with the Emoncms readings. But at the moment I dont understand the system enough to do consumption measurement over a fixed period of time.
Re: Accuracy question
I don't think an error that big is down to one cause, but several all adding up the wrong way.
Until you get the voltage input working, the emonTx and the electricity company's meter are measuring different quantities, one real power/energy, the other apparent power; so unless all your loads are purely resistive, they should never agree since apparent power must always be the same or larger than real power.
The c.t's might well interact, but only to a very small degree because all the flux should be trapped inside the core, and very little should leak out. I've never tried to make any definitive measurements but I've also never noticed an effect - and I've had three on the same cable with only an inch or so between them.
If you look into what you can do with emonCMS, you'll see that kWh features fairly prominently, so you should be able to do a direct comparison, when you have the voltage input working that is.