Archived Forum

12.03 V, but what was your mains voltage at the time? You can't calculate it without knowing that.

The best way, if you can measure mains voltage, is to adjust the voltage calibration constant that you send to emonLib with

ct1.voltageTX(222.2, 1.7);

etc. In this case, the constant is the first parameter: 222.2, and you adjust it so that emoncms (or the serial output) reports the same as your voltmeter. (You change it in the positive sense - make the constant larger and the voltage gets larger.)

If you look at the statement like

typedef struct { int ct; double realPower, apparentPower, Vrms, Irms, powerFactor;} PayloadTX;

in your sketch, the entries in emoncms are the variables inside that structure in the order they appear.

Hi Robert, Thanks for your helpful reply.

I only possess one multi-meter, so I was only able to measure the mains voltage first and then the adapter voltage soon after (244.8 - 246.5v mains, 12.03v from the adapter)

Just for info: (to all)

I purposely laid out my question in such a way as to imply very little knowledge of this subject, as I have looked through the forum for a simple guide on this, and found nothing. I hope that for the non-technically minded like me, this will be helpful.

Phil. T

The calibration procedure is actually under Modules -> emonTx -> Firmware ->Calibration.

Comment removed by the author

Phil, spend some time and read through the guide which Robert directed you to.

If you have a specific question about the guide, then ask the question, but please don't label the guide as gibberish, because the majority of members have found it useful, and your comment is an insult to the author.

Paul

Apart from heartily despising the ubiquitous practice of "dumbing down" just about everything in the modern world, I'm having a big problem in seeing how you can 'dumb down' these two lines any further:

The new calibration coefficient can be calculated by:
  New calibration = (correct reading / emonTx reading) * current calibration

The only criticism I could have there is the use of the word "current" rather than "existing" because it could be interpreted as meaning electrical current (only) when in fact the expression will work for both calibration constants.

I have now changed that, but I'm not prepared to tear apart and rewrite Glyn's page any more than that without his knowledge and agreement.

If you want to send me your comments, feel free to do so. Bear in mind that any calibration procedure must be written so as to be usable with the sort of test equipment that you can reasonably expect to be available to the average constructor. In many cases this might be restricted to a budget multimeter with limited accuracy and ranges.

Apologies to all for my ranting, I allowed my frustration to get the better of me :(

hi Robert, your change of "current" to "existing" is good, because I'm sure I did interpret it as applying only to the electrical current calibration, so thanks for that.  I also mis-interpreted the phrase "another calibrated meter" to mean something like "another calibrated power meter, such as another emon".  

Perhaps we could add another sentence like "For voltage calibration, a reasonable quality multimeter will do, but only knowledgeable and experienced people should attempt to directly measure mains voltage.".  

For current calibration, is a clampmeter is the appropriate tool?  But, for the common case where the CT clamp and burden resistor are supplied by the OEM shop, the need for calibration is probably low?  In contrast, for voltage the sensor (9v AC plugpack) is often not supplied by the OEM shop, at least for foreign customers like myself.  So the need to calibrate is probably higher for voltage? 

If all the above is fair, then maybe another few sentences like "For current calibration, a multimeter is not of use. An appropriate clampmeter could be used to obtain an electrical current reading, again when used by a knowledgeable and experienced person. For current measurement, where a standard emonTx kit is used, the default calibration is likely to be quite accurate."

Um - sorry if thats dumbing it down too much!  

Noted all that, thank you, and some minor changes done. The reason is I've been contemplating multimeter accuracy and calibration for a while, and trying to write an article on the subject. I've failed thus far  - it just didn't come together, so I'm going to split it into two, one on multimeter accuracy and one on the sources of calibration errors in the emonTx (although that's touched on in various places, one article pulling the lot together would be good, I think). Then I noticed that the step-by-step emonTx calibration procedure that I thought we had has disappeared, so I reckon a replacement taking into account the user's meter's capabilities would be useful. That's a third article.

Unfortunately, some of the things you say aren't right (for example, the worst case initial error could be around 15%) - and there's no simple way to give a single set of instructions that will always give the best result. Hopefully, you'll see what I mean when I've finished writing those pieces.

Watch out for something in Building Blocks or Modules, but don't hold your breath. Meanwhile, constructive comments from everyone are welcome. If you could mention what meter(s) you have (make & model so that I can work out the accuracy for the actual values you measure), and your country (or supply system & voltage), all the better.