Data Sheet

AD8318

USING THE AD8318

analog.com

Rev. E | 14 of 24

RTADJ, nominally 499 Ω for optimal temperature compensation at

2.2 GHz input frequency, is connected between the TADJ pin

and ground (see Figure 23). The value of this resistor partially

determines the magnitude of an analog correction coefficient that is

employed to reduce intercept drift.

Table 5 lists recommended resistors for various frequencies. These

resistors provide the best overall temperature drift based on meas-

urements of a diverse population of devices.

The relationship between output temperature drift and frequency is

nonlinear and is not easily modeled. Experimentation is required to

choose the correct RTADJ resistor at frequencies not listed in Table

5.

Table 5. Recommended RTADJ Resistors

Frequency

Recommended RTADJ

900 MHz

500 Ω

1.9 MHz

500 Ω

2.2 GHz

500 Ω

3.6 GHz

51 Ω

5.8 GHz

1 kΩ

8 GHz

500 Ω

TEMPERATURE SENSOR

The AD8318 internally generates a voltage that is proportional to

absolute-temperature (VPTAT). The VPTAT voltage is multiplied by a

factor of 5, resulting in a 2 mV/°C output at the TEMP pin. The

output voltage at 27°C is typically 600 mV. An emitter follower

drives the TEMP pin, as shown in Figure 30.

Figure 30. TEMP Sensor Interface

The internal pull-down resistance is 5 kΩ. The temperature sensor

has a slope of 2 mV/°C.

The temperature sensor output varies with output current due to in-

creased die temperature. Output loads less than 1 kΩ draw enough

current from the output stage causing this increase to occur. An

output current of 10 mA results in the voltage on the temperature

sensor to increase by 1.5°C, or ~3 mV.

Best precision from the temperature sensor is obtained when the

supply current to AD8318 remains fairly constant, that is, no heavy

load drive.

MEASUREMENT MODE

When the VOUT voltage, or a portion of the VOUT voltage, is

fed back to VSET, the device operates in measurement mode. As

shown in Figure 31, the AD8318 has an offset voltage, a negative

slope, and a VOUT measurement intercept greater than its input

signal range.

Figure 31. Typical Output Voltage vs. Input Signal

The output voltage vs. input signal voltage of the AD8318 is linear-

in-dB over a multidecade range. The equation for this function is

VOUT = X × VSLOPE/DEC × log10(VIN/VINTERCEPT)

(3)

= X × VSLOPE/dB × 20 × log10(VIN/VINTERCEPT)

(4)

where:

X is the feedback factor in VSET = VOUT/X.

VINTERCEPT is expressed in Vrms.

VSLOPE/DEC is nominally −500 mV/decade and VSLOPE/dB is nominal-

ly −25 mV/dB.

VINTERCEPT, expressed in dBV, is the x-axis intercept of the linear-

in-dB transfer function shown in Figure 31. VINTERCEPT is 7 dBV (20

dBm, referenced to 50 Ω or 2.239 Vrms) for a sinusoidal input signal.

The slope of the transfer function can be increased to accommo-

date various converter mV per dB (LSB per dB) requirements.

However, increasing the slope can reduce the dynamic range.

This is due to the limitation of the minimum and maximum output

voltages, determined by the chosen scaling factor X.

The minimum value for VOUT is X × VOFFSET. The offset voltage,

VOFFSET, is equal to 0.5 V and is internally added to the detector

output signal.

VOUT(MIN) = (X × VOFFSET)

(5)

The maximum output voltage is 2.1 V × X, and cannot exceed 400

mV below the positive supply.

VOUT(MAX) = (2.1 V × X) when X < (VPOS − 400 mV)/(2.1 V)

(6)

VOUT(MAX) = (VPOS − 400 mV) when X ≥ (VPOS − 400 mV)/

(2.1 V)

(7)