Pulse count not correctly set up, need advice

I have some problems understanding how to set up my pulse counter. The energy meter is saying 2000imp/kWh. The pulse counter shows a green blink for every red blink from the meter, it does not seem to miss out an any. And it is always blinking (main household meter). From reading the guides here I would say that this is how the input config should be to convert to pulsecount n to energy in kWh:

energy[kWh]=n_meas[imp]/2000[imp/kWh]=n_imp*0.0005 [kWh]

This is how it is implemented in emoncms:

But the resulting energy per day number is way too low by about - but not exactly - two decades when compared with CT readings (which usually is about 1-2% from the energy company numbers). I do not have too much stats to show yet due to other problems.

Is there something I have missed completely?

Hi, my setup is a bit different in step 1. I don’t calculate the pulse delta between reads.

Have you seen the Optical Pulse Sensor setup guide https://guide.openenergymonitor.org/setup/optical-pulse-sensor/?

I tried something similar, and it does not work in my case:

Since the meter is defined as 2000 imp/kWh I multiply the pulses by 0.5 to get Wh. Then I used the kWh accumulator, which I guess will give realtime accumulated Wh (not kWh as I have put into the feed name). The funny thing is that this value becomes a constant with the value 3.00.

Finally, my thought was to multiply the Wh increments with 1000 to get kWh before running the “Wh incremements to kWh/d”. This was obviously wrong, it should have been multiplied by 0.001. Anyway, both the scale and kWh/d number value are wrong (not just the 10^6 factor error). So I’m at loss how to configure this.

I did try and read the setup guide, but so far I do not get any sensical outputs…

I just now tried to change the first multiplier from 0.5 to 0.0005 to actually get Wh. For some reason, the number 3.00000 now started to get increments. I’ll wait a couple of days to see what this results in.

Update: This is the input config now:

The kWh accumulator seems to work perfectly. Here is a 900s average plot with CT measurements (blue line) and pulse counts overlay (yellow circles). They match more or less 100%.

The Wh increments to kWh/d is not working as expected, though:

Typical daily values would be 30 to 60kWh/d, while the delta values here are < 1kWh/d. I experience this on the daily increments feed from the CT’s as well, where values are == 0. They are configured like this:


Great to hear!

Rather than use the wh increment to kwh/d process the recommended approach is to use the ‘delta’ mode when creating a graph to produce kwh/d on the fly, see the guide on how to do this here: Calculating Daily kWh - Guide | OpenEnergyMonitor

Yes, thanks, I’ve been using that method, but thought that a pre-calculated feed would be nice…

Does this mean that the wh increment function does not work? If you look at my previous post, I can not get this to work for the “kHw to kWh/d” from the CT feeds either.

Im not sure about the 'Wh increments to kWh/d" its been a while since I used it to be honest. If there’s anyone else who is using this process that would like to confirm if this issue can be replicated that would be great.

I have added these processes on my system here to test in the mean time.

I’ve created a github issue to keep track of this here: Review kwh/d input processors, are they working correctly? · Issue #1504 · emoncms/emoncms · GitHub

I am trying the same thing but am getting a constant 4.38 - my meter pulse says 1600/1000. Have been playing around with no luck so far

Brian - any ideas on my feed above from my post? Have also tried 0.000625 but think my maths is off

Trying to set a similar real-time feed from the pulse count to that of the CT clip for comparison. Similar to Olav

What is wrong with 0.0005 fed into a kWh accumulator process?

My pulse meter goes at rate 1600

Trying to get a real time (or as near as) feed similar to the CT feed to compare. And have updated the accumulator to be Wh and tried that too

Ah different person/meter sorry.

1/1600 =

So that is the right figure for a kWh accumulator.

Set Pulse Feed to 10s and CT Feed to 10s. Delete all the Feeds and start afresh.

Pulse Feed is 100% accurate (assuming it detects every pulse - which it should), CT Feed isn’t - it is an approximation. Currently I’m multiplying a CT feed by 0.9695 to make it agree with the pulse meter.

All my CT inputs apparently over read (I have not calibrated them).

Thanks will try that again and monitor for a bit

Cheers again Brian -

Two quick questions which hoping you might have a steer on:

A) the above seemed to work fine - but when graphing the axis values seem to be off? Not sure if i need to define them but it is set to auto (see graph screenshot). The reading between the ct clip and pulse seem to be broadly similar which i hope is a good thing…
B) was playing with getting the pulse count feed in a wh feed and kwh feed and tried the following process logic - think it is right - is that the way you would recommend getting both feeds?

That is what I do.

On graphing, looking at short periods of time can be odd. Also, unless both Feeds started at exactly the same time put one on left and one on right but check the scales are the same else you get some odd views. This is 2 feeds (pulse and CT) started about the same time.

This is 2 feeds started at different times - note the scale is different.

If I force the scales to be the same…

You can see the CT over reading.

If you look at the stats…

You can see the ‘Diff’ that is a the measured kWh in the frame. Missing data on the pulse will be when the rate of change means I do not get a pulse within the 10s window.

Mmmm, that doesn’t look right. The Accumulator should never go down!

I know

The whAccumulator process outputs an unchanged value so you need to multiply by 0.000625 for kWh.