Hi everyone,

I am very new to the ade 7753 metering chip and have been trying to calibrate it correctly (previous user tried some interesting techniques to get outputs to look right...)

I began by stepping through the ade7753 pdf ( www.analog.com/static/imported-files/data_sheets/**ADE7753**.pdf )

And have got a bit stuck with the watt gain outputs. I have been able to stick to the settings laid out in the example for the most part, but my device is outputting an LAEnergy of nearer ~60,000 (instead of expected ~19,000 with CFDEN of 489, or ~11,000 with CFDEN of 293) when running on 10A @ 220V with power factor of 0.999 (verified) ((accurate power source)).

The confusing bit is that if I change the CFDEN value it doesn't appear to change the LAEnergy output. Yet the example suggested that it should.

Could it be that the CT input is giving much larger values than the ade will ever expect and I should be looking at the hardware?

Or is there simply a magnitude of error that I should just attenuate before comparison to the Expected LAEnergy?

Thanks for any help, and sorry if I miss anything useful.

I'll be keeping a close eye so will provide anything that will help.

-Chris

Hi Chris,

Actually that isn't a problem at all and could even make things easier. This way you will just be reading the energy registers, which means that anything that has to do with CFNUM or CFDEN can be ignored in the calibration steps. If you don't need the CF pulse output at all then I would suggest leaving it disabled with the DISCF bit in the MODE register.

I'll start with the CT, you mention that it has a turns ratio of 1:2500, now a burden resistance has to be chosen such that it corresponds to the +/-500mV peak input of the ADE7753. A helpful equation for this is shown in the FAQ application note, under "How do I calculate the burden resistor to use with my current transformer?" on Page 4. This equation will give the correct burden value in order to scale the input so that the maximum current you expect on the line will correspond to a +/-250mV peak at the inputs to give some headroom for overcurrent situations.

I would suggest doing phase calibration first, if the CT introduces phase delay, to cancel this out and make the rest of calibration as accurate as possible. "Calibrating Phase with an Accurate Source Example" will help with that.

Reading the LAENERGY register with line cycle accumulation mode, find the Wh/LSB constant with equation 36. This Wh/LSB is your constant to translate a register reading to a real world energy reading.

If you want to adjust the Wh/LSB constant to a certain value or to match multiple meters up then use the above equation and plug in Wh/LSB to find your LAENERGYexpected. This way all meters will have the same Wh/LSB constant. Then use equation 40 to find the WGAIN needed, with AENERGYnominal being the actual register value found with you nominal input. It can be 220 V and 10 A or any other value as long as you take that into account as your Load(W).

Then follow "Calibrating Watt Offset with an Accurate Source Example" with Imin being the minimum current you expect to measure on the line. For the rest, if needed, follow the VRMS, IRMS and apparent energy calibration as shown in the data sheet. This way the CF is completely independent of your calibration procedure.

Let me know if you run into any trouble.

Regards,

David