Ref voltages on emonpi and emontx calibration and more accuracy

andrewhf01 · 2018-01-03T11:37:17.060Z

For discussions related to emonPi, emonTx, emonBase and emonGLCD hardware and Arduino firmware

I just purchased an emonpi and two emontx V3 I think the system is excellent and I’m still very new to this DIY monitoring etc. I have noticed a significant difference in the ref vrms input/feeds for each of these (AC-AC Power Supply Adapter - AC voltage sensor purchased at the same time); the emontx4 and emonpi AC voltage sensors are plugged into the same socket. I also have two pulse counters for the grid (emonpi) and solar meters (emontx4).

Therefore, I’ve been monitoring the voltages on each of these an example daily:
EmonPi max 237.1, mean 233.0 and min 221.7
EmonTx3 max 247.4 mean 243.5 and min 232.5
EmonTx4 max 251.4 mean 246.8 and min 235.9

Although these differences are within (thereabout) the plus and minus differences published on the calibration page (Learn | OpenEnergyMonitor) this will have a significant effect on the results over a year when calculating the kWh’s for each for example for one day:

grid CT monitor (EmonPi) = 48.6 kWh
grid pulse sensor (EmonPi) = 51.5 kWh
a difference of 2.9 kWh (5.8%) - over a year 2.9*365.25 = 1059.2 kWh (or £117.25)

PV generation CT monitor (EmonTX4) = 0.9 kWh
grid pulse sensor (EmonTX4) = 0.7 kWh
a difference of -0.2 kWh (25%).

Is there a way that I can get this system more accurate?
Many thanks Andrew

stuart · 2018-01-03T12:44:48.584Z

Have you read through the emonTX calibration articles?

Emontx voltage Calibration Getting Started

Hi All, just wanting to get my thoughts correct as im new to modifying the programming for Audrino based items. i know nothing about “sketches”, are these another name for firmware? i have 3 Emontx V3’s ( got these before i knew about he IoTaWatt) all with EmonESP modules, all talking to my own EmonCMS server. so i have about a 10% discrepancy in my voltage readings on my emonTX’s compared to what is actually there. i have read the guide about calibration and comprehend what is needed to be c…

Calibration of emonTX3.4 and attached CT inputs (USA) emonTx

Thanks to Glyn, Robert and Bill (whose sigh of relief was heard across the Atlantic) my system seems to be up and running. However, my feeling is that the power calibration is not correct for power1-4 which I am using. I have adjusted the scale factor for Vrms (emonhub.conf) so that my emon input reads the same as my mains (Fluke 87 III). I have a 9v AC/AC adapter purchased in the US. On power1 &2 I have two SCT023Rs connected to the mains so I can read each plus a sum of the two feeds On pow…

andrewhf01 · 2018-01-03T14:38:11.313Z

Hi Stuart,
Yes, I’ve read through those; I bought these as ready built units and I assumed that in the manufacture of these they would have done these calibrations; is this correct or not? I’m also just getting familiarised with this kind of monitoring so expecting a a steep learning curve.
As Bob said “Is there a way to calibrate the system without having to go into the emonTX where things have to be edited and recompiled? Can it be done by software configuration in the RPI3 + RFM69Pi_433? I’m not looking for extreme precision but something fairly accurate and easy for a layman to modify.”
Thanks, Andrew

Robert.Wall · 2018-01-03T15:00:08.014Z

andrewhf01:

I assumed that in the manufacture of these they would have done these calibrations;

Not a correct assumption I’m afraid. That would mean that a “set” of components (emonTx, c.t, v.t.) would need calibration together and to be kept together, because the calibration process corrects for manufacturing variations in both the input sensor itself and in the components of the input to which it is attached.

If your emonBase is running emonHub, then the simplest (possibly not the most accurate because you cannot calibrate for phase & timing errors) is to adjust the ``scale = " constants in emonHub, as mentioned in Stuart’s quote.

andrewhf01 · 2018-01-03T17:51:16.165Z

Hi Robert,
I assume you are the Robert Wall referenced at the bottom of the page ‘summarising the rather convoluted standards surrounding UK grid voltages’ etc.?

I was under the impression that if I purchased the AC-AC Power Supply Adapter - AC voltage sensor it was used to monitor the voltage owithin the house circuit and this was then used real-time by the emonpi and the emontx V3 to make regular adjustments to the power calculations inline with the fluctuations in the mains voltage. If I use and adjust the scale values for the then this would be a work round for the vrms, would this impact on the power inputs and the feeds etc.?

I have monitored these:
for example of a daily reading over a 24hr period:
EmonPi max 237.1, mean 233.0 and min 221.7
EmonTx3 V3 max 247.4 mean 243.5 and min 232.5
EmonTx4 V3 max 251.4 mean 246.8 and min 235.9

The emonpi and the emontx4 V3 the AC-AC monitors are in the same socket and give different results; the emontx3 V3 and emontx4 V3 are in different sockets but give very similar voltages. I have also switched the AC-AC voltage sensors around between all the emonpi and the two month’s and the results do not change I have checked my mains voltage with a Fluke 189 multimeter (admittedly not recently calibrated) and the mains voltage is consistently around 242 to 246.

Why am I getting a consistently lower vrms reading from the emonpi (base station) than the two emontx V3 (sensor nodes)? It does not say anywhere in the online guide about setup and installing, that I need to do any calibration with the AC-AC voltage sensors (and this would be accounted for in manufacture).

I would be happy to rectify this problem first - how to correct the emompi vrms AC-AC voltage sensor reading (I feel the emontx’s seem reasonable accurate and consistent with my multimeter readings) the other elements of calibration i.e. CT, which I is perhaps diminishing returns for the effort at this point in time.

Many thanks
Andrew

Robert.Wall · 2018-01-03T20:01:03.065Z

andrewhf01:

If I use and adjust the scale values for the then this would be a work round for the vrms, would this impact on the power inputs and the feeds etc.?

No. The processing of voltage and current into power takes place inside the emonTx (or the front end of the emonPi, which has identical input circuitry for current and voltage), the voltage and powers are sent separately to emonHub and thence into emonCMS. So you must adjust the voltage coefficient for the correct voltage, and each power coefficient independently to give the correct power. Trimming the voltage on its own in emonHub does not affect the powers.

As I wrote above, calibration is particular to the combination of input sensor and input. As you’ve read in the analysis of the error sources, each sensor and each component makes an individual contribution to the calibration, and it’s the result of all those individual variations from the nominal value that you’re correcting. Switching the a.c. adapters around just means that they all have the same (as close as you can tell) voltage ratio - it does not mean that the ratio is exactly the nominal value, as the manufacturer only guarantees that to within ± 3%.

The emonTx V3.4.4 has some closer tolerance components fitted (“some”, because unfortunately a slip-up occurred writing the bill of materials) so that, when the change works its way through the supply chain, should be closer to the nominal specification. I don’t know when that change will happen to the emonPi. If you’ve got a V3.4.4, that may well explain the consistency between the two emonTx’s and at the same time the discrepancy between those and the emonPi.

andrewhf01 · 2018-01-05T16:17:14.943Z

Thanks Robert,
I’m not an electronics or electrical engineer - I just enjoy dabbling in/ around with electronics/ electrical projects for fun and enjoyment!
I have not played at all with either ardrino’s or raspberry pi’s so, clearly I’ve got some significant research and understanding (learning) to do with using and accessing Ardrino and raspberry pi devices and I hasten to add, using this community based forum for acquiring information and feedback etc (not used these much to get technical help on devices).
I will then (hopefully) have a better understanding how to access the serial connections and the config files etc, which will then allow me to change the values (mentioned in the calibration information) in these config files appropriately.