MCP3912 datasheet(56/82 Pages) MICROCHIP

56 / 82 page

MCP3912

DS20005348A-page 56

 2014 Microchip Technology Inc.

The MCP3912 is highly recommended in applications

using di/dt as current sensors because of the extremely

low noise floor at low frequencies. In such applications,

a low-pass filter (LPF) with a cut-off frequency much

lower than the signal frequency (50-60 Hz for metering)

is used to compensate for the 90 degree shift and for

the 20 db/decade attenuation induced by the di/dt sen-

sor. Because of this filter, the SNR will be decreased,

since the signal will attenuate by a few orders of mag-

nitude while the low-frequency noise will not be attenu-

ated. Usually, a high-order high-pass filter (HPF) is

used to attenuate the low-frequency noise in order to

prevent a dramatic degradation of the SNR, which can

be very important in other parts. A high-order filter will

also consume a significant portion of the computation

power of the MCU. When using the MCP3912, such a

high-order HPF is not required since this part has a low

noise floor at low frequencies. A first-order HPF is

enough to achieve very good accuracy.

7.6

Energy Measurement

Error Considerations

The measurement error is a typical representation of

the nonlinearity of a pair of ADCs (see Section 4.0

“Terminology And Formulas”

for the definition of

measurement error). The measurement error is

dependent on the THD and on the noise floor of the

ADCs.

Improving the measurement error specification on the

MCP3912 can be realized by increasing the OSR (to

get a better SINAD and THD performance) and, to

some extent, the BOOST settings (if the bandwidth of

the measurements is too limited by the bandwidth of

the amplifiers in the sigma-delta ADCs). In most of the

energy metering AC applications, high-pass filters are

used to cancel the offset on each ADC channel (current

and voltage channels), and therefore a single-point

calibration is necessary to calibrate the system for

active energy measurement. This calibration is a

system gain calibration, and the user can utilize the

EN_GAINCAL bit and the GAINCAL_CHn registers to

perform this digital calibration. After such calibration,

typical measurement error curves like Figure 2-7 can

be generated by sweeping the current channel

amplitude and measuring the energy at the outputs (the

energy calculations here are being realized off-chip).

The error is measured using a gain of 1x, as it is

commonly used in most CT-based applications.

At low signal amplitude values (typically 1000:1

dynamic range and higher), the crosstalk between

channels, mainly caused by the PCB, becomes a

significant part of the perturbation as the measurement

error increases. The 1-point measurement error curves

in Figure 2-5 have been performed with a full-scale

sine wave on all the inputs that are not measured,

which means that these channels induce a maximum

amount of crosstalk on the measurement error curve.

In order to avoid such behavior, a 2-point calibration

can be put in place in the calculation section.

This 2-point calibration can be a simple linear

interpolation between two calibration points (one at

high amplitudes, one at low amplitudes at each end of

the dynamic range) and helps to significantly lower the

effect of crosstalk between channels. A 2-point

calibration is very effective in maintaining the

measurement error close to zero on the whole dynamic

range, since the nonlinearity and distortion of the

MCP3912 is very low. Figure 2-6 shows the

measurement error curves obtained with the same

ADC data taken for Figure 2-5, but where a 2-point

calibration has been applied. The difference is

significant only at the low end of the dynamic range,

where all the perturbing factors are a bigger part of the

ADC output signals. These curves show extremely tight

measurement error across the full dynamic range

(here, typically 10,000:1), which is required in

high-accuracy class meters.

MCP3912 Datenblatt(PDF) 56 Page - Microchip Technology

MCP3912 Datenblatt(HTML) 56 Page - Microchip Technology

Html Pages

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82

Datenblatt Download

Link URL