The graph shows power in kiloWatt while the legend suggests energy in Watthours

Thank you for your feedback. I'm in discussions with Niclas over at FooGadgets. There might be a bug in the formula I use. Since I don't have an actual wireless energy meter myself it's a bit difficult to get right.

Can you try again now? I have modified the formula to show kWh.

In my opinion the graph was perfectly correct. It was only the label Ws that annoyed me since Ws suggests energy while the graph shows power. Now I get "Error 400, bad request" when I try to update the graph. The temperature senors works ok as before. 

That's strange. Can't find any errors in the log files. Do you still get the 400 bad request? 

Yes, I still get it. The reading in columns "Last update" and "Data" are correctly updated, it's only when clicking on the graph symbol I get the error message. Once more, the graph as such was correct before I wrote the (unfortunate) ticket, sorry!

Is it only for the energy meter, or do other sensors also fail?

Can you please logon to http://tellmon.net then open the following URL in another tab in the same browser?
http://tellmon.net/services/telldus.svc/GetSensorDataForGraph?sensorId=1332439&graphRange=Day

Thank you

All sensors and graphs are working correctly. Even the "numerical" energy display (the one that show energy as degrees and the humidity counter) is updated correctly. It is only the energy graph that generates the fault code.

And this is result from the URL you gave me:

Can you please try again but using test.tellmon.net instead of tellmon.net.

Logon to http://test.tellmon.net
Then open a new tab and try this URL:
http://test.tellmon.net/services/telldus.svc/GetSensorDataForGraph?sensorId=1332439&graphRange=Day

I have enabled more logging and diagnostics as well as running test.tellmon.net in debug mode. Hopefully that will give me some pointers to why you are having this issue.

You should get some more text below the "Request error" this time.

Done, I got the database dumped on the screen, I think. No error (and no graph, just a screen full of data).

That's what is supposed to happen when it works. But we wanted it to crash like earlier.

That's strange. Does it work when you click on the sensor to see the graph in the normal way (but on test.tellmon.net) as well?

I works ok when I klick on the sensor in test.tellmon.net in the normal way. In tellmon.net I still get the error.

I've changed the lables from kWh to kW and redeployed on tellmon.net
Do you still get the error on tellmon.net?

But how can I calculate the consumed energy in kWh? I though that simply adding all the Wh-values would be correct. 

Now tellmon.net works ok as well. Thank you!
If we assume 1 "blink" per Wh, then adding the blinks will result in the consumed energy i Wh. To display in kWh you have to divide by 1000. (Adding the average Wh values will not give the correct result, you have to add the "blinks").

That is what I do. The user can specify Wh pr. "blink" on the settings page (usually 1).

The "sum" below the graph is the sum of all the 1-minute temp values * whPerBlink * 10 * 60 / 1000

The *10 is because the energy meter reports blinks/10 (so temp of 2.4 means 24 blinks last minute).

I'm aware of that the temperature value reported by telldus server should be multiplied by 10.
Strictly speaking the energy meter reports the correct number of blinks pr. one minute.
Telldus' server processes this value as a temperature value in "Fine Offset" format, hence dividing it by 10.
Why are you multiplying by 60?
One Wh pr. blink is one Wh pr. blink regardless of the sample interval.
24 blinks represents 24 Wh regardless of sample interval.
It's only when calculating the average power you have to take the sample interval into account.
24 Wh gives 24*3600 = 86400 Ws or 1440 Watt minutes if you wish.
86400 Ws during 60 seconds gives an average power of 86400/60 = 1440 W or 1,44 kW.
This is value I expect to see in the graph.

I was told by Niclas at FooGadgets that I need to multiply by 60 to convert from Watt-minute to Watt-hours:

------------------------- From Niclas ---------------
Pmomentan = Temp*10/1000*60*60/(Time_now - Time_previous)
Temp är inkomna temperaturen
60/(Time_now - Time_previous) är en faktor som behöver läggas till eftersom varje blinkdetektor har olika sändningsintervall. Det skiljer sig med kanske +-2 sekunder. Intervallet är också beroende på batterispänning och chipets temperatur. Resultat i kW
-------------------------
I've skipped adjusting for the +-2 seconds since I instead interpolate into 60 one-minute values pr. hour based on what ever was logged. Quite often several reading are missing. Either because of a sensor collision (two sensors sending via RF at the same time can cause Tellstick to miss the value) or just some other issue.

I am supposed to receive 60 values pr. hour by they vary from as low as 49 to 59 at best.

Hi,
Here is the (latest) formulas that I sent to Pål.
The first show how one calculate the Total consumed Energy by summing the incoming numbers and get the result in kWh.
The second show how one calculate the current power consumption during the last minute in kW.

---------------------start snippet----------------
TotalEnergi = Startvärde + Sum(temperaturN*10)/1000
Startvärde är något man kan tänka sig att kunden vill lägga till så att mätvärdet i graferna stämmer med var elmätaren visar. Ska anges i kWh. Det lägger man till vid ett tillfälle. Samma tillfälle som man anger att den här temperatursensorn egentligen är en energimätare.
---------------------end snippet----------------

So I think the confusion is in the last post from Pål above, where you Pål state, "...I need to multiply by 60 to convert from Watt-minute to Watt-hours..." and then refer to my formula that calculate Power (W).
It is the TotalEnergi-formula that should be used and there there is a factor 60 involved.

...and it might be me who is the source of it, since I answered an email from Pål where he wanted me to review a formula for Power. I confirmed it was OK. However I did not see until now that it stated result in kWh <-- did not note this.

I hope I got that right now so that I do not confuse this even more :)

Excellent work you have done Pål! The site looks great and I am sure a lot of people will appreciate your service!

Edit: What I have forgotten to include is the compensation factor for if the energy meter is blinking once per Wh or 10 times per Wh or whatever. This is however just a factor to include in the formulas.

... and I got the important statement wrong :/ 
"It is the TotalEnergi-formula that should be used and there there is a factor 60 involved."

This should of course be,
"It is the TotalEnergi-formula that should be used and there there is NOT a factor 60 involved."