SLAA577F 3-phase Watt-Hour Meter Report

Hello Ryan

Thanks for your email.

Let me briefly explain my motivation to develop a Watt-Hour Meter based on the SLAA577F report.

Recently, in Brazil, due to very bad government decisions about inflation control and extremely dry seasons (last two) which obliged the activation of expensive thermo electric power generators, the price of electricity is skyrocketing. Due to this, I think that brazilians will begin to use electricity with more parsimony and sense of economy. This means that it will be important to know, for example, the cost of a hot shower, or how much is the home air conditioning costing, or a week of normal washing machine usage, or even how much it costs to leave a TV on standby for a month. The meter that I am trying to develop, I hope, will help answer these questions. It will have the communication characteristics of the TIDU531 smart plug (with WI-FI Direct or Bluetooth) and be based on the SLAA577F project using split core Current Transformers.

Simplicity and flexibility will be important aspects of this meter. This is why I would like to provide an automatic and easy way for the (1 phase) meter to be calibrated. In this case, the user (a technician) should have access to a single phase voltage (110 or 220V) connected to a resistive load that should demand a current of approximately a 5 Amps, and set up the meter to measure this voltage and current. In parallel, with a good multimeter, the voltage and current should also be measured and the values passed to the meter software by means of a celphone with WI-FI Direct communication. This way, the scale factors for voltage, current and apparent power can be determined and stored in the proper non volatile NV segment of memory. In addition, since the load is purely resistive, the phase delay introduced by the Current Transformer can also be determined. Additional current values to determine a dynamic phase correction curve can easily be obtained.

About your first observation, concerning the moment I read the voltage channels, I would like to say that I am not reading the voltages directly. Instead I am calling SLAA577F routines for that and am assuming that if I load a SD24BPREx voltage register with some value Z < 256, that the respective voltage will be read by the SLAA577F software Z*1/255*1/60 seconds later, same way as the current SD24BPRE registers.

Note that the SLAA577F software admits the setting of the Voltage Preload Registers by means of the parameter DEFAULT_V_PRESCALE_FACTOR of the configuration data file, which initializes the 3 voltage preload registers SD24BPRE[0,1,2] with the same DEFAULT value.

Could you also please give an opinion about my question relative to the STEPs calculation.

Thanks for your advice.

Regards

Claudio R. Sonnenburg

Hi Ryan

Thanks very much for your comments.

For economy, please let me answer directly in your text.

• The preload register provides delays in modulation clock cycles. Therefore, you would expect

Z/(4096*256) seconds delay when applying a preload value of Z.

Sure. I mistakenly wrote the frequency instead of the sampling rate.

• A preload value of 128 is not applied as you mentioned. This part of the application note is meant to

just mention one method of providing phase shifts.

OK

• Please also keep in mind that the delay due to I_HISTORY_STEPS is seen not only in the current

samples used for calculating power but also the voltage samples.

OK

• The phase correction calibration value can be seen as the delay between voltage and current

pairings. Please keep in mind that a voltage sample is paired with a current sample that is ready

before and not after this voltage sample is ready. In addition, generally, a smaller and more negative

value of this calibration value corresponds to a voltage being more delayed with respect to the

current sample that it is grouped with.

I assumed that:

1. All voltages and currents start with 2 delay STEP’s (I_HISTORY_STEPS) which facilitates the anticipation of any individual current or voltage;

2. STEP=1  for a specific current is equivalent to 256 units of SD24BPRE correction (delay); In other words, starting with SD24BPRE4 = 0 and STEP = 2 is the same as starting with SD24BPRE4 = 256 and STEP = 1;

3. DEFAULT_BASE_PHASE_A_CORRECTION_LOW = +1 (or -1) means that the phase A current will be anticipated (or delayed) by 1/(256 * 4096) sec.  If this is correct, the values of STEP’s derived by SLAA577F are not OK.

Please comment.

Regards,

Claudio

Hello Ryan

With respect to item 2, I meant the following:

If I initialize the phase A current STEP variable with STEP=1 and the corresponding ADC SD24BPRE register with 256, the end result, as far as phase correction is concerned, is the same as if I initialize phase A with STEP=2 and the corresponding SD24BPRE register with 0.

Thanks for your time.

Regards

Claudio