What is the EmonTx actually recording ? is it Wh?

My EmonTx v3 gives me data every 5 seconds which I think is an Average of the last 5 seconds. I have been blithely recording this for years and converting it into kWh. Or so I thought but someone asked me today what I’m actually measuring and I realised that if I did know once I have now totally forgotten.
could someone let me know please

It’s not really a 5 second average, it’s a few of the 50Hz cycles every 5 seconds. It gives you the Watts being used at that point in time.
E.g. If everything in your home is switched off it would be zero watts. This is unlikely in most modern homes, as there’s stuff like internet routers, central heating, smoke alarms, burglar alarms, TV on standby, etc that use a few watts each, so your away from home load may be something like 100W to 500W (or even higher).
Switch on the electric immersion and the power use will jump to 3000W+.
If you have solar panels, the reading may go negative as power is exported (leaves the building).

The “power to kWh” functions turn all the watts into kWh, much like a electricity meter counts up power usage.
The “power to kWh per day”, is the same, it just resets every midnight, so shows the power used today.

That’s not necessarily true - it depends on when it was purchased and whether it has the “CM” software or not. If it was shipped after 18/10/2019, it should have the CM software, in which case “it’s a few of the 50Hz cycles every 5 seconds” is quite wrong. Before that date and unless it’s been updated, then you’re correct.

Mine was bought in 2014 approx
So the numbers I get out are Watts is that correct?
so if the number is say 300 does that mean a total of 300w in 5 seconds or an average of 300w in 5 seconds and I can’t even work out if that is the same thing :frowning:

You’ve almost certainly got the ‘old’ Discrete Sample software.

It is watts, but it’s not even the average over the last 5 seconds, it’s the spot value just before it sent the reading (in reality, it’s the average over 200 ms just before it sent the reading). emonCMS assumes this is typical of the power in watts over that 5 s interval, and computes the energy in watt-hours as 500 W × 5 s = 0.69444 Wh.

No, they’re not the same thing. Power in watts is the rate at which energy in Wh (or kWh) is used.
Or the other way round, energy in Wh is the power your appliance consumes (in watts) multiplied by the time it’s using it (in hours).

And if you want an hydraulic analogy, power is the rate at which water fills your bucket, and energy is the amount of water in the bucket when it has.

I’m not surprised you’re confused, a lot of people who really ought to know better - particularly journalists and publicity seekers - seem not to know the difference and get it wrong.


If you want, you can update the firmware to the newer Continuous Monitoring (CM) firmware. Need to check that version does support it, though and you would need a programmer to plug into the UART on the back (the row of pins through the slot).

EmonCM works for all emonTx’s so far (not counting the yet to be released V4, which it won’t without extensive modification),

Including V2 or just any V3?

It was largely developed on a V2, on account of all the I/O pins, which I could hang a 'scope on to be able to check all the timings.

1 Like

Thanks for all the replies.
Yes I think I understand Wh means now.
I have a programmer So I could install the new firmware.
Is it much more accurate?
I get quite decent accuracy, Comparing it to my meter readings over 8 years It underestimated by 500kWh so 160 watts per day

The principal advantage of Continuous Monitoring is when you have rapidly switched loads. In this case, it will definitely be more accurate in the short term, though the effects of averaging over time will mean that long-term, you’re unlikely to see a substantial difference. If you don’t have anything that switches rapidly, and you’re comparing over many hours or days, you’re unlikely to see any difference. It’s the same hardware, so all the inaccuracies due to that remain the same.

Ok Thanks
I have a Car charger and a PvRouter Mk2 butI don’t think they would switch rapidy?
I make daily totals By averaging the readings do you think the errors would more likely to be there?

Hmm, if I’ve understood it right, the switching for this is quite rapid. There’s a bit about how this works on LearnLearn→PV Diversion→Introduction→→Choosing an Energy Diverter.

Actually, I think the work of Robin Emley, the designer of the Mk2, contributed a lot to the continuous monitoring sketch we use.

Still, the results you have gotten still look pretty decent to me regardless.

If it’s Robin Emley’s Mk2PVRouter, it most certainly does switch rapidly - down to a few cycles depending on the software version and the mode setting.

And Gwil is right - emonLibCM is largely based on Robin’s software.

The CT method of recording, gives better instant use value.

If you want better overall accuracy, the Pulse Sensor is the best bet, but that gives poor instant use output (as you only get a flash once you have used x amount of energy).

Thanks for replies :slightly_smiling_face:
I realised that Over the 8 yrs that I have had the EmonTX There were a lot of missing days data due to various problems. So I tried to test the accuracy against my solar power and mains meters from 1st April this year as there have been no outages during this period.
Unfortunately over the 190 days since 1st of April I have recorded an extra 81kWh for imports and an extra 110kWh production.
So over the 190days since 1st of April, that averages 0.426Wh/day for imports and .578kWh/day for exports which is a lot more than I’d expected.
So taking into account what I’ve been told By Robert Wall, I should probably look into installing the new software.

And then calibrating it.

Despite our best efforts, some major components have quite large tolerances. If you’re lucky, they cancel each other. If you’re not, they add up in the same direction and give a large cumulative error. Calibration should remove most of the long-term error, but I’d expect to still see some error when light loads and heavy loads are compared - this is down to the way c.t’s work and in the fullness of time, I hope we can produce some software to alleviate the effect. Until then, I’m afraid we must live with it. You should be able to get well within 1% of your meter with careful tweaking of the calibration values.

I’d suggest you put all the c.t’s on one cable with a big resistive load and follow the calibration procedure, then transfer to the meter tail and make fine adjustments over an extended period to get the energy to tally with what your meter shows. When you’re happy, you can deploy the c.t’s back to their designated locations.

I will do that but.
I’m re-reading all of this thread and realise that you say “emonCMS assumes this is typical of the power in watts over that 5 s interval, and computes the energy in watt-hours as 500 W × 5 s = 0.69444 Wh.”
Apart from the fact that I’m not sure how I’d implement that.
I am not doing that obviously.
I’m just collecting the raw data every 5 seconds and then averaging it over the day, then mulitplying by 24 and dividing by 1000 to give me kWh’s/day.
I can’t remember where I got that from but I’m thinking that might be the wrong way of doing it?

Sorry, but something you know but don’t write and tell us is by no means obvious to us even if it is to you.

If you want us to be able to help beyond the original question “What is the EmonTx actually recording?” then you need to tell us what you’re doing and what you really want to know.

What I’m trying to say is that all the raw Numbers I get from the EmonTx are being added to a raw feeds table a database and then every day I Average these inputs and then multiply by 24 and divide by 1000 and save them in a new Daily results Table.
So these are the numbers I am using.
If I am making a mistake in the way I’m doing this, it’s more than possible that it’s my mistake rather than than the EmonTx and If I change the software and Calibrate the data, it will still be wrong.
What I’m asking is, is the way I am calculating the days totals the right way to do it Or should I be doing it as in the example you gave from the emoncms?
Hope that’s clearer :slightly_smiling_face: