CN0178

The RF signal being measured is applied to the input of the ADL5902, a linear-in-dB rms-responding rms detector. The external 60.4 Ω resistor, R3, combined with the relatively high input impedance of the ADL5902 ensures a broadband 50 Ω match to the RF input. The ADL5902 is configured in its so-called “measurement mode,” with the VSET and VOUT pins connected together. In this mode the output voltage is proportional to the logarithm of the rms value of the input. In other words, the reading is presented directly in decibels and is scaled to 1.06 V per decade, or 53 mV/dB.

The power supply voltage and reference voltage for the AD7466 12-bit ADC are provided by the ADL5902 internal 2.3 V reference. Because the AD7466 consumes so little current (16 μA when sampling at 10 kSPS), the ADL5902’s reference voltage output can supply the ADC, as well as the temperature compensating and rms accuracy-scaling network consisting of R9, R10, R11, and R12.

The ADC full-scale voltage is equal to 2.3 V. The maximum detector output voltage (when operating in its linear input range) is approximately 3.5 V (see ADL5902 data sheet figures 6, 7, 8, 12, 13, and 14) and must, therefore, be scaled down by a factor of 0.657 before driving the AD7466. This scaling is implemented using a simple resistor divider R10 and R11 (1.21 kΩ and 2.0 kΩ). These values provide an actual scaling factor of 0.623, which ensures that the ADL5902 RF detector does not overdrive the ADC by building in some room for resistor tolerance.

A typical plot of detector output voltage vs. input power is shown in Figure 2 (without output scaling).

The transfer function of the detector can be approximated by the equation:

VOUT = SLOPE_DETECTOR × (PIN − INTERCEPT)

where SLOPE_DETECTOR is in mV/dB; INTERCEPT is the x-axis intercept with a unit of dBm; PIN is the input power in dBm.

At the output of the ADC, VOUT is replaced by the ADC’s output code, and the equation can be rewritten as:

CODE = SLOPE × (PIN − INTERCEPT)

where SLOPE is the combined slope of the detector, the scaling resistors, and the ADC, and has the unit of counts/dB; PIN and INTERCEPT still have the unit of dBm.

Figure 3 shows a typical detector power sweep in terms of input power and observed ADC output codes for a 700 MHz input signal.

Overall SLOPE and INTERCEPT will vary from system to system. This variation is caused by part to part variations in the transfer function of the RF detector, the scaling resistors, and the ADC. As a result, a system level calibration is required to determine the complete system SLOPE and INTERCEPT. In this application, a 4-point calibration is used to correct for some nonlinearity in the RF detector’s transfer function, particularly at the low end. This 4-point calibration scheme yields three SLOPE and three INTERCEPT calibration coefficients, which should be stored in nonvolatile RAM (NVM) after calibration.

The calibration is performed by applying four known signal levels to the ADL5902 and measuring the corresponding output codes from the ADC. The calibration points chosen should be within the linear operating range of the device. In this example, calibration points at 0 dBm, −20 dBm, −45 dBm, and −58 dBm were used.

The SLOPE and INTERCEPT calibration coefficients are calculated using the equations:

SLOPE1 = (CODE _1 – CODE_2)/(PIN_1 − PIN_2)
INTERCEPT1= CODE_1/(SLOPE_ADC × PIN_1)

This calculation is then repeated using CODE_2/CODE_3 and CODE_3/CODE_4 to calculate SLOPE2/INTERCEPT2 and SLOPE3/INTERCEPT3, respectively. The six calibration coefficients should then be stored in NVM along with CODE_1, CODE_2, CODE_3, and CODE_4.

When the circuit is in operation in the field, these calibration coefficients are used to calculate an unknown input power level, PIN, using the equation:

PIN = (CODE/SLOPE) + INTERCEPT

In order to retrieve the appropriate SLOPE and INTERCEPT calibration coefficients during circuit operation, the observed CODE from the ADC must be compared to CODE_1, CODE_2, CODE_3, and CODE_4. For example if the CODE from the ADC is between CODE_1 and CODE_2, then the SLOPE1 and INTERCEPT1 should be used. This step can also be used to provide an underrange or overrange warning. For example, if the CODE from the ADC is greater than CODE_1 or less than CODE_4, it indicates that the measured power is outside of the calibration range.

Figure 3 also shows the transfer function variation of the circuit vs. the above straight line equations. This error function is caused by bending at the edges of the transfer function, small ripple in the linear operating range, and drift over temperature. The error is expressed in dB using the equation:

Error (dB) = Calculated RF Power − True Input Power
= (CODE/SLOPE) + INTERCEPT – PIN_TRUE

Figure 3 also includes plots of error vs. temperature. In this case the measured ADC codes at +85°C and −40°C are compared to the straight line equations at ambient. This is consistent with a real world system where system calibration is generally only practical at ambient temperature.

Figure 4 and Figure 5 show the performance of the circuit at 1 GHz and 2.2 GHz, respectively.

The performance of this or any high speed circuit is highly dependent on proper PCB layout. This includes, but is not limited to, power supply bypassing, controlled impedance lines (where required), component placement, signal routing, and power and ground planes. (See MT-031 Tutorial, MT-101 Tutorial, and article, A Practical Guide to High-Speed Printed-Circuit-Board Layout, for more detailed information regarding PCB layout.)