Energy meter display graph

I updated tellmon about 1.5 hours ago. I must have introduced a bug then. Will look into it. I dont have an energy meter myself, so it's a bit tricky to test for it. 

I asume the data sent by any user's sensor is collected the same way, is just a number sent to Telldus Live. Something is going wrong afterwards. But thank you for looking into it.

If I am to guess, I'd say that the graphs are reversed between sensors, although only one sensor is declared Energy Meter in Settings. Yet, both have 1 Watt-hours per blink saved and I cannot erase that 1 Watt-hours per blink from neither of them, even if I uncheck Energy Meter from both of them.

I did some bug fixing. Can you please have another try now?

I have declared them both now as energy monitor, and they seem fine, although the logged data differs from one another. Which one should I have selected? Which is the "primary" sensor? Or can I monitor them both, just to check and compare them? It is not clear from the settings page how this should be configured.

 I still cannot erase 1 number in the blinks field.

the graph seems better now.

I can see though that the labels on the graph do not show the correct value. It displays 0.3 but it shows 0.01 kW

Should be fixed now

I don't know, the label on the graph is still showing 0.00 although the sensor indicates 0.004kW. That should stand for an actual consumption of 4Watts, right? But this is NOT ACCURATE, because I have hardware digital plug meters where all devices are powered from, and only one of them shows 97 Watts instant consumption. Another plug meter on the fridge shows 70 Watts. I have a few other small consumtion devices in hous which should add a few watts as well to the total... So, I don't know what Energy Meter actually sends to Tellmon or what Tellmon displays...



Here is what Telldus Live shows at the moment:

The formula used to convert the "temp" reading to kW is:

kW = temp * 10 * WhPrBlink / 1000

Your sensor is reporting a temp value of 0.5. And you have entered that your power meter blinks once pr. Wh.

0.5 * 10 * 1 / 1000 = 0,005 kW which corresponds to what your graph is showing. Either your value for Watt-hours pr. blink is wrong (see https://tellmon.userecho.com/topic/412401-how-to-s... or your reader is not aligned properly to catch all the "blinks".

Either way, please contact FooGadgets where you purchased your power reader as I don't have anything to do with them.

Yes, I have done all of these, yet.... How about the graph's label? Is it ok to indicate 0.00 at any other value? Best regards.

The graph values are rounded to two decimals. 0.004 rounded is 0.00 so the label is fine.

I noticed that they say: "The detected number of blinks will be presented as a temperature value on the receiving side divided by ten (i.e. 324 number of detected blinks will be shown as a temperature of 32.4 °C."

http://foogadgets.tictail.com/product/wireless-ene...

If my sensor indicates 0.5 it means it reads 50 blinks, which means 0.5 * 10 * 1 / 1000 = 0.05 kW; not 0.005 kW!

Which should stand for a 50 Watt instant consumption measured by Energy Meter where all my other digital meters sum up almost 200 W. So there might be other things that go wrong as well.

They also say that: "One Primary sensor that carries the last minutes number of detected blinks. And a Secondary sensor that carries the penultimate number of detected blinks."

Which means a correct display by Tellmon should take into account both sensor readings. If so, one sensor would show 0.1 and the other 0.5. If corectly interpreted, the penultimate number being 1 and the ultimate number being 5 maybe the formula should be different? Maybe the total displayed should be around 0.150kW ?

Even so... the digital meters show around 0.200 kW so either way, the FooGadgets might measure differently from the digital meters that I have.

I think you need to recalculate that. 0.5 * 10 * 1 / 1000 is actually 0.005.

But I agree that there is something wrong somewhere :)

My ERROR! I appologize.

OR maybe, if the other sensor is primary then my consumption is 0.210 kW, which sounds more ACCURATE to what my other meeters are reading. Which means Tellmon should take into account both sensors, and distinguish between the Primary and Secondary. 

I don't know, I am confused.

I dare to suggest that a different setting interface should e used for the sensors declared "ENERGY METER" and also the ENERGY should be displayed in a separate Energy Meter Graph interface...

From blog post: http://foogadgets.blogspot.se/2014/01/wireless-ene...

The following algorithm can be used to detect lost packages and also log to a database. See it as pseudo code,

// Callback function (If it looks similar to Telldus it is because it is based on an example from them)
void sensorEvent( int sensorId, const char *value, char *temp) {
if ( sensorId == 69 ) { //Store the secondary virtual sensor value
f_temp[sensorId] = atof( temp );
i_seqno[sensorId] = atoi( value );
return;
} else if ( sensorId == 68 ) { //Store the Primary virtual sensor value
f_temp[sensorId] = atof( temp );
i_seqno[sensorId] = atoi( value );
i_diff[sensorId] = i_seqno[sensorId] - i_prev[sensorId];
switch ( i_diff[sensorId] ) {
case 1:
case -99: // All OK
// Log f_temp in a MySql-db or send it to Xively for example.
case 2:
case -98: // One transmission lost. Check if we can recover.
if ( (i_seqno[sensorId]-i_seqno[sensorId+1]) == 1 ||
(i_seqno[sensorId]-i_seqno[sensorId+1]) == -99 ) {
// It seem like we can recover. Log the f_temp[sensorId+1] and the f_temp[sensorId] value.
} else {
// If we are here, we have lost one transmission and can not recover. Anyway, we log the current value f_temp[sensorId] and do a best guess of what the lost value was. We could use an average number or the same as current value f_temp[sensorId]. This would likely be less wrong than leaving it empty.
}
break;
default : // If we are here we have lost more than one complete transmission. If this occur frequently you have to look over your setup.
}
i_prev[sensorId] = i_seqno[sensorId];
}
return;
}

This is all from: http://foogadgets.tictail.com/product/wireless-ene...

The primary sensor and the secondary sensor both consists of a temp/humidity pair.
The temperature is the number of blinks the last minute (/10). The humidity is a reading counter (0-99).

The secondary sensor is showing the primary sensors previous value. So using the humidity counter one can detect that a package has gone missing and then use the secondary sensors value instead. 

Tellstick is not a reliable system and packages go missing all the time. That is why FooGadget has built in the secondary sensors as a backup. The primary and secondary values should follow each other. If not, there is something very wrong. Perhaps you need to update the firmware of the energy meter.

Tellmon does not support the secondary sensor yet since it's monitoring about 2100 sensors and 4900 devices the database is about 5 GB pr. month since I store all values. I'm working on resampling the data into hourly averages for data more than one month old. I'm also looking into ways to support the secondary sensor, but it needs to be in a way that is not CPU intensive because of the large volume of data coming in.

I understand. 

What I don't understand is why my plug meters show something so different than Energy Meter. I mean, is desirable to consume 200W and my counter have on record 5W, but I don't think this is the actual fact. I have rechecked the alignment of the sensor, and is ok.

I have counted the number of blinks myself and noticed that if there are 4 blinks over one minute for 240W indicated by my digital meters, the sensor also sends 0.4 value. That means the power consumed in this minute (for 1000 blinks/kW) it is given by formula:

4 blinks per passed time between 2 records / 1000 blinks per kW = 0.004 kW in the passing time

If I had a 324 blinks within one minute, the sensor would have sent 32.4 at one recorded time. and the consumption in that minute would be:

4 blinks per passed time between 2 records / 1000 blinks per kW = 0.324 kW in the passing time

All goods, but the power is displayed in kWh not kW. So the last number should be transformed to kWh. Say the interval for the above example is exactly 60 seconds; that means the numbers in kWh would be 0.24kWh and 19.44kWh, which is the nominal power.

In fact the precise formula is given by FooGadgets:

Pmom = temp*10/1000*60* 60/(Time_now - Time_previous)

That means that the way is displayed now, it is correct, but with confusion, because the power is usually expressed in kWhour not kW. That's all. I dare to suggest to implement this exact formula, to avoid missunderstandings. It is much more intuitively to read an energy meter that shows an instant consumption of say 0.240kWh instead of 0.004kW since the last reading. Otherwise, eeach user would have to figure out by multiplying 0.004 with 60 minutes.

The real problem with the measurement without implementing a formula involving TimeNow - TimePrevious is that each 0.1 given by the sensor as temperature translates into 60Wh. Nothing in betwin, if measurement is done every 60 seconds but on server side the logging time is not taken into account.

A B wrote: "All goods, but the power is displayed in kWh not kW."

That's not true. 

http://en.wikipedia.org/wiki/Power_(physics)#Units
"The SI unit of power is the watt (W)." 

http://en.wikipedia.org/wiki/Units_of_energy#Electricity
"The energy unit used for everyday electricity, particularly for utility bills, is the kilowatt-hour (kWh)."

Both your statements are true. It is my "mistake" to refer to "power" I should have said consumption. Anyway, it is nothing incorrect here, except is unusual to display consumption in watts per minute, every minute, and also not make corrections based on timeNow - timePrevious. The graph is a coarse guess and it will never display the real instant consumption numbers. just an aproximation and a trend variation.


Maybe it should take into consideration the way other energy monitors on the market display their numbers for instant consumption.

How does other energy monitors display instant consumption? That is what I was aiming for with the FooGadget support. I want the graph to show instant consumption varying over time. Since I don't support the timeNow-timePrev and correction stuff it is only an approximation. Thus I cannot show the total energy consumption in any what that compares to the actual physical energy reader at your home. 

Unfortunately I have neither a FooGadget energy reader of my own, now any other energy monitors that display instant consumption. I've been trying to simulate the FooGadget device in order to develop support for it. I have some ideas about how to support time variation and missing data packages. But since I'm running in Azure on a scaling infrastructure (max 2 servers) I need a way that is both computationally cheap and that can scale out over multiple servers.

As of now tellmon.net costs me about NOK 700 a month in Azure hosting and I really need to either move to another cheaper host, or add enough features so I can start offering a subscription for the more advanced stuff. The main cost is actually the SQL server for storing all the data. It's growing with about 5 GB a month. I've been looking into other storage means like https://tempo-db.com/ but that again costs from USD 100 a month and up, but would fit tellmon's use great.

Here is for instance how the newly introduced Energy Monitor from Coco Technologgy displays this in their mobile app http://www.klikaanklikuit.nl/shop/nl/producten-1/...



Here you can see the units used to display the numbers. I do not have that actual monitor.

All the monitors on the market display in these unit of measure. So it is just about that, display instant consumption in kWh. You display in kWmin.

And I really can't understand how adjusting the formula to take into account time interval would complicate things in hosting. The result of the formula would be just another number to store. and the calculus I doubt it would be so processor intensive, but yes, I do not know that for a fact, you may know better.