AD8436

Rev. 0 | Page 13 of 20

For simplicity, Figure 28 shows ripple vs. frequency for four

combinations of CAVG and CLPF

INPUT FREQUENCY (Hz)

1

0.0001

100

1k

10

0.001

0.01

0.1

CAVG = 1µF, CLPF = 0.33µF

CAVG = 10µF, CLPF = 3.3µF

CAVG = 10µF, CLPF = 0.33µF

CAVG = 1µF, CLPF = 3.3µF

AC INPUT = 300mV rms

Figure 28. Residual Ripple Voltage for Various Filter Configurations

Figure 29 shows the effects of averaging and post-rms filter

capacitors on transition and settling times using a 10-cycle,

50 Hz, 1 second period burst signal input to demonstrate time-

domain behavior. In this instance, the averaging capacitor value

was 10 μF, yielding a ripple value of 6 mV rms. A postconversion

capacitor (CLPF) of .068 μF reduced the ripple to 1 mV rms. An

averaging capacitor value of 82 μF reduced the ripple to 1 mV

but extended the transition time (and cost) significantly.

INPUT

50Hz 10 CYCLE BURST

400mV/DIV

TIME (100ms/DIV)

CAVG = 82µF

CAVG = 10µF FOR BOTH

PLOTS, BUT RED PLOT HAS

NO LOW-PASS FILTER, GREEN

PLOT HAS CLPF = 68nF 100mV/DIV

Figure 29. Effects of Various Filter Options on Transition Times

Capacitor Construction

Although tolerant of most capacitor styles, rms conversion

accuracy can be affected by the type of capacitor that is selected.

Capacitors with low dc leakage yield best all around performance,

and many sources are available. Metalized polyester or similar

film styles are best, as long as the temperature range is appropriate.

For practical applications such as the rms-to-dc function in

DMMs or power monitoring circuits, surface mount tantalums

are the best over-all choice.

Basic Core Connections

Many applications require only a single external capacitor for

averaging. A 10 μF capacitor is more than adequate for acceptable

rms errors at line frequencies and below.

The signal source sees the input 8 kΩ voltage-to-current conversion

resistor at Pin 2 (RMS); thus, the ideal source impedance is a

voltage source (0 Ω source impedance). If a non-zero signal source

impedance cannot be avoided, be sure to account for any series

connected voltage drop.

An input coupling capacitor must be used to realize the near-zero

output offset voltage feature of the AD8436. Select a coupling

capacitor value that is appropriate for the lowest expected

operating frequency of interest. As a rule of thumb, the input

coupling capacitor can be the same as or half the value of the

averaging capacitor because the time constants are similar. For

a 10 μF averaging capacitor, a 4.7 μF or 10 μF tantalum capacitor

is a good choice (see Figure 30).

VEE

CAVG

VCC

OUT

RMS

OGND

+5V

10µF

–5V

IGND

CAVG

AD8436

19

2

11

10

8

9

17

4.7µF

OR

10µF

Figure 30. Basic Applications Circuit

Using a Capacitor for High Crest Factor Applications

The AD8436 contains a unique crest factor feature. Crest factor

is often overlooked when considering the requirements of rms-

to-dc converters, but it is very important when working with

signals with spikes or high peaks. The crest factor is defined as

the ratio of peak voltage to rms. See Table 5 for crest factors for

some common waveforms.

VEE

CAVG CCF

VCC

OUT

RMS

OGND

+5V

10µF

–5V

IGND

CAVG

0.1µF

CCF

AD8436

19

18

2

11

10

8

9

17

4.7µF

OR

10µF

Figure 31. Connection for Additional Crest Factor Performance

Crest factor performance is mostly applicable for unexpected

waveforms such as switching transients in switchmode power

supplies. In such applications, most of the energy is in these

peaks and can be destructive to the circuitry involved, although

the average ac value can be quite low.

Figure 13 shows the effects of an additional crest factor

capacitor of 0.1 μF and an averaging capacitor of 10 μF. The

larger capacitor serves to average the energy over long spaces

between pulses, while the CCF capacitor charges and holds the

energy within the relatively narrow pulse.