Regarding Current channel sensing through CT in ADE7953

You can use a single burden but you will be limited to singled ended amplitudes and worse common mode rejection. Differential allows for a larger swing into the amplifier so better snr and common mode rejection. 

Single ended burden may be fine for your testing. I recommend center taped burden for best performance. 

Dave

Hi Dave, 

Sorry if this is not the right place. I am trying to improve the performance of a commercial product that integrates an ADE7953 (Shelly EM) thanks to a new firmware (TASMOTA), we have access to all calibration registers.

This Shelly EM works with a pair of CT 50A, after adjusting PHCAL I have greatly improved the precision with low FP loads.

I had the idea of replacing a CT with a SHUNT (with a lower range 0-30A would be enough), according to what I have been able to read, SHUNTs do not suffer from phase changes like CTs and should be more precise, therefore PHCAL would remain at 0 , at this moment I have PHCAL at 200. Is this true? would you try the change?

Thanks for your help!

The ADE7953 has best performance with a shunt on channel A. A shunt will work on channel B with pga gain. This works if the voltage is the same for both current A and B(branch measurement of 2 legs on the same phase). If this is split phase shunts will not work you will need a shunt and CT or 2 CT. 

CT will require phase compensation and this may be a constant if all the CTs are from the same manufacture and part number. Really depends on the required accuracy. 

you care correct about shunts they do have a small phase shift and can still be improved with phase compensation for the absolute best performance but may not be required for the accuracy you are looking for.

Dave

HI Dave, let's see what you can think of that could be happening, with the calculation of reactive power.

In the images you see on one side the theoretical calculation, according to input values Vrms, Irms, PF. and next to it the TASMOTA firmware with the Shelly EM and the 50A CTs. (you are watching both channels on the same line)

Currently the Vrms, Irms, PF, P and S, are calibrated and correspond throughout their range, both with a low and high PF. They exactly match the values of my company's electric meter.

On the other hand, the Q, although I have tried to adjust it, I did not get it to work precisely, in fact you will be able to see that the error is not linear and changes from negative to positive according to the PF.

Thanks in advance!

What load are you testing? In your theoretical calculation you are dealing with fundamental reactive. Does your load have harmonics? 

How many bits do you get in the VAR register? is +/- and lsb = to +12 and -11 in your calculation? 

Can you give me the raw hex from the power registers you are testing above for watt var and va? 

Dave

What load are you testing? In your theoretical calculation you are dealing with fundamental reactive

In these tests the load is a MIX of consumers in my house. TV, led lighting, charges, refrigerator. (the habitual thing of a house)

In the second test where I=6.37A connect a resistive load (toaster).

 Does your load have harmonics? 

upsssss I had not counted on this! The problem is that I have no way of comparing the reactive with any device. (Only the theoretical calculation) My company's electric meter does not measure this value to take it as a reference.

How many bits do you get in the VAR register? is +/- and lsb = to +12 and -11 in your calculation? 

Sorry, I just didn't understand the question, I may lack knowledge!

Can you give me the raw hex from the power registers you are testing above for watt var and va? 

Let's see if this helps....

Here are the calibration values of both channels. a and b

CalibRegs aV 1596500, bV 1596500, aI 4800000, bI 4869356, aW 3100000, bW 3100000, aVA 3100000, bVA 3100000, aVAr 3285364, bVAr 3285364, aP 200, bP 200

Here you can see the values that are registered

HI dave.smith You could see the data that I provide below. Let's see what opinion you deserve. Thank you!

HI dave.smith You could see the data that I provide below. Let's see what opinion you deserve. Thank you!