I am building and installing a Emon based monitoring board for my central 110 V split-phase breaker box. I will obviously need to do voltage monitoring in the box as well. What I would like to know is if I have other current sensors (CTs probably) local to specific loads (e.g. range/oven, refrigerator, television, window AC unit, and possibly even more) to provided greater granularity how important would it be that I take a Vref at each of those loads? How much accuracy would I be giving up if I used the Vref taken at the main panel than if I were to do a Vref at each local load?
Thanks!
Re: Centralized voltage monitoring with distributed current monitors in a 110 V singe phase system
I should mention why I'm interested. If I don't gain much in the way of accuracy from having a Vref at every metering point it simplifies the design of the various sensors that I distribute around my house. A simplified design means no AC-AC transformer, shorter BOM, less IO, and smaller boards. All good things.
If you're wondering why I'd be interested in having load level monitoring when I already will have circuit/breaker level monitoring it's simply because I'm slightly unbalanced and have a thing for large quantities of data. Okay, maybe that's not entirely true. I'm simply interested in getting as good of a picture of what's going on in my place as possible. Plus, it sounds like fun.
Re: Centralized voltage monitoring with distributed current monitors in a 110 V singe phase system
The accuracy you lose depends on the voltage drop in your wiring. You give no clue as to where you are in the world, in the UK there are rules that state the maximum drop allowable at the furthest point. You need to find out what the rules are where you live, or what the actual voltage drops are on load, and decide whether that loss of accuracy is acceptable to you.
However, the emonTx system doesn't have a mechanism for transmitting the voltage reference between sensor nodes, so you will only be able to calculate apparent power (VA) if your voltage and current sensors are in different places. To get real power (W), you need local voltage sensing. (You need to look at the maths to see why that should be.)
Re: Centralized voltage monitoring with distributed current monitors in a 110 V singe phase system
I'm sorry, I should have included a far more information than I did yesterday.
I'm in the the US. Portland Oregon to be exact. The the building I live in was built in 1968-69. I'm not 100% positive exactly what type and gauge of wire was used in the construction. Everything that I've encountered has been three conductor (hot, neutral, and earth ground), solid, copper romex. My unit is supplied with US standard 110 V single phase power. My breaker panel contains seven 15 amp breakers, and six 30 amp bridged dual breakers (I think those are the numbers, I'm not at home so I can't check for sure). There is no primary breaker, switch, or disconnect between the panel's primary bus and the power company's meter outside (0.5 meters away).
Also, the system I'm working is only partial based on the emon designs but it's not exactly the same. I'm primarily using the analog side of the hardware design. I also expect that I'll use a good portion of the emon code once I get my board design finished and move on to the coding. The digital and communication portions will be my own. I plan to post my design files, code, and documentation to GitHub or similar once I get a little further along. I expect to license the work under the GPLv2 (just like emon).
My original worries about the idea of not taking a Vref locally at each of the sensors and instead using the value taken at the breaker panel revolved around any inherent capacitance or inductance in the wiring system itself. That could further alter the voltage vs. current phase shift that CTs naturally introduce. I hadn't even considered the voltage drop from the wire runs. I'm quite embarrassed that I had failed to take that into account.
I'll have to take a look at the local code from the era my building was constructed to see if there's a max voltage drop specified. I'll also open up a few more boxes to verify what type of wire was used.
I really appreciate your help.
--adam
Re: Centralized voltage monitoring with distributed current monitors in a 110 V singe phase system
I doubt that the additional capacitance or inductance of the house wiring at 60 Hz will make a measurable difference - I think you can safely forget that.
"There is no primary breaker, switch, or disconnect between the panel's primary bus and the power company's meter outside" I find that quite astonishing - unless there is a main isolating switch or a breaker outside, next to the meter. It's certainly not UK nor European practice to not have a main switch somewhere - usually it is part of the main distribution panel. (I'd get someone who knows the local rules to check that. What you have sounds wrong to me, unless of course you've misidentified the main switch).
The problem still remains if you want to calculate real power for each appliance. The sums are:
Apparent power = rms voltage x rms current. You can measure and average voltage and current independently.
Real power = average instantaneous power = average of (instantaneous voltage x instantaneous current). You need to measure voltage and current at the same instant.
EmonTx measures voltage and current samples consecutively, and applies an extrapolation to correct for the timing difference and the phase errors introduced by the transformers. If you're measuring voltage centrally you have the problem of distributing the voltage measurement at the rate of about 2000 per second (emonTx rate - you could run somewhat slower) and with a timing accuracy of around 100 µs (360° being 16.6 ms - 100 µs is a little over 2°, and that's a reasonable value to aim for). I'll leave you to think about that. (Or your problem is getting the current measurements to a central point to do the maths - in essence it's the same problem.)
You could of course go round measuring the voltage drops in the circuits under load - it should be the same in both line and neutral, so the safest would be to measure the neutral drop and double it; or if neutral and earth are at the same potential at the main distribution board, measure the neutral-earth voltage at the socket outlet and double it. In the UK, the limit is 4% of the declared supply voltage, which if neglected would give you around 8% error in your power measurement.