I want to help a customer monitor his 3 phase power, and help him balance his loads among the 3 phases in a 240V 3 phase, 3 wire system.
I have 3 questions:
1. Can I accurately monitor 3 phase power using 3 EmonTx V3's and 3 AC adapters (1 per phase)?
2. I see Martin's "Full Fat" sketch syncs the measurments between three EmonTX's to occur at the same point in time, but why is this important? If the 3 EmonTX's free run, won't it all average out over minutes or hours anyhow?
3. How do I do the math to combine the individual readings in emonCMS to come up with a total power number? Is it as simple as adding PH1 power + PH2 power + PH3 power?
The electrical panel to be monitored will have mostly single phase loads but also some 3 phase loads like air conditioners.
Thanks for your help!
Re: 3 Phase Monitoring using 3 EmonTX V3's - some questions
1. Yes, provided that you have no true 3-phase loads, and you prepared to accept that running the "Discrete Sample" sketches you'll have non-synchronous readings as you appreciate. All within the accuracy limitations of the "discrete samples" sketches, of course.
2. It's important only if you're not prepared to accept that running the "Discrete Sample" sketches you'll have non-synchronous readings as you appreciate.
3. Just add, as you say. If you do have 3 emonTx's running asynchronously, and a large 3-phase load switches on, and suppose the 3 emonTx's report 3⅓ s apart, you'll see the load come on in 3 stages.
Re: 3 Phase Monitoring using 3 EmonTX V3's - some questions
Thanks for the answers.
If I run the "continuous sampling" sketch, but don't sync the emonTX's, what ill effects will be noticed? If a large 3 phase load switches on, won't each emonTX pick it up within one cycle anyway, since they are sampling continuously? I guess I'm just trying to get my arms around the benefit of syncing them.
Re: 3 Phase Monitoring using 3 EmonTX V3's - some questions
No, the "continuous sampling" sketch will pick up the change within 1 sample ( ~ 1/50th mains cycle) but it won't report it for up to 10 s. If you need instant updates, you will notice, otherwise it should make no difference.
Martin's "Full Fat" sketch predates his PLL energy controller and Robin's "continuous sampling" sketches; does that help your understanding?
Re: 3 Phase Monitoring using 3 EmonTX V3's - some questions
>Martin's "Full Fat" sketch predates his PLL energy controller and Robin's "continuous sampling" sketches; does that help your understanding?<
Aha! The lightbulb has come on.
So if I'm monitoring 60 hz power, which sketch will provide the highest accuracy: PLL or "Continuous Sampling"?
Side note: I have had some trouble adapting the 50hz sketch "emonTxV3_3Phase_Voltage_4CTs_Temp" to 60hz. (which I have running on a single EmonTxV3 in this case). After changing samples delay for L2 and L3, calibrating with a resistive load, and more than a month of fiddling with Phasecal I am still not getting results that match up with the power company's smart meter. Thus my impetus to use 3 independent EmonTXV3's instead.
Re: 3 Phase Monitoring using 3 EmonTX V3's - some questions
My problem is I don't have 3 phase supply and I'm on a 50 Hz system, so all I can do is make sure it gives the correct "wrong" answers for me! You would of course need to change (move forward) the sample that's picked up for L2 & L3, and as far as I know there isn't a 3-phase PLL sketch in existence (another one on my to-do list!). But there are 60 Hz derivatives of Martin's PLL.
There ought to be no difference at all between the "continuous" and the PLL sketches. The PLL samples at exactly the same points on the wave each time, whereas the "continuous" is not phase locked and the sampling points will therefore drift along the wave. As long as there are no really high power high order harmonics that encroach into the area where aliasing is a problem, I think any difference is going to be down to calibration (as calibration and correction for phase error is done slightly differently).
As the notes in the 3-phase sketch explain, delaying the L1 wave and using it to represent the other two will always be a compromise that's beset with difficulties and has shortcomings that are inherent in the basic concept. For example, the "phase" isn't a phase delay, it's a timing delay, and the two can no longer be equated when the frequency changes. So I'd always advocate using 3 separate emonTx's where you desire better accuracy and you can justify the cost.
I've hinted a few times that a 3-phase emonTx would go down well outside the UK (indeed some poor unfortunates have assumed that the standard one IS three-phase), but there's no hint that a design might be under consideration. It would need at least 9 ADC channels (3 x V, 3 x 2 x I) and something better than the ATMega 328.