Question: How can I figure out the output flatness vs. frequency of an RF Detector such as the AD8363
Answer:
Using the Slope and Intercept specs it is possible to get a sense for and RF Detector's flatness vs frequency. Let’s take the example of the AD8363. Let’s consider the easy-to-calculate case where the input power is equal to 0 dBm. The general equation for the output voltage of a linear-in-dB RF Detector is:
Vout = Slope x (Pin – Intercept)
So for Pin = 0 dBm, this reduces to
Vout = - Slope x Intercept
Using the published slope and intercept specs of AD8363, we get
100 MHz Vout = 2.99 V
900 MHz Vout = 3.0044 V
1.9 GHz Vout = 2.86 V
2.14 GHz Vout = 2.8188 V
2.6 GHz Vout = 2.5921
Since the slope of the AD8363 is around 52 mV/dB, this corresponds to a variation vs frequency (over this freq range) of approximately 8 dB.
On a newer device ADL5511 whose output voltage is linear in V/V, we have measured the variation and found it to be much smaller. In this case, we see that the device is quite flat up to around 4 GHz. The trade off with this device (compared to AD8363 or AD8318) is that it has less range (47 dB). For more information on this, take a look at Figures 57 and 58 in the Rev. A Datasheet.
The plot below shows the relative errors of various RF Detectors vs. frequency after they have been calibrated at a single frequency. Unsurprisingly, the ADL5511 that was referenced above has the flattest overall performance (this plot has the brown tick marks in the plot). This device also has another advantage in applications where RF power needs to be measured over a wide frequency range. Many of ADI's RF Detectors have external temperature compensation nodes. These inputs allow for the temperature drift of the device (which can vary with frequeny) to be compensated. Since the drift can vary with frequency, the required compensation voltage will also vary. In the case of the ADL5511 RF Detector, no external temperature compensation is required.