Post Go back to editing

# ADL5535 1PdB - 1dB Compression Point Unterstanding

Hi Folks,

please could you help me to explain the fact, that the ADL5535 has a higher 1dB compression point than the power supply allows?

Example:

1dB P1 = 18.6dB as lowest in statistic, according to Fig.11. of the datasheet.

18.6dB is 72.44mW into a 50 Ohm load.

U = SQRT (P*R) = SQRT (72.44mW * 50) = 1.9Vrms

Upeak = 1.9Vrms * sqrt(2) = 2.69V

To run Upeak-peak the Voltage swing is: 2 * 2.69V = 5.38V.

This should be the voltage on the 50Ohm termination.

If we consider the internal driver has a 50Ohm series termination, the driving voltage must have 10.77V. Considering no extra headrom for the transistors to drive the load.

How can the outputpowerlevel of at least 18.6dB be explained?

Is the datasheet correct?

Thanks.

• I'm also facing the same problem, and the datasheet is definitely incorrect... I'm still waiting for an official answer from Analog.

According to my measurements, P1dB is around 9dBm, and Psat is 13.5dBm (@5V, @160MHz). I hope it helps...

• This is just a guess, but if the P1dB is specified for a conjugate match to the device output and the real part of the output impedance is less than 50 ohms, then the power output would be correspondingly higher. Looking at the values for S22 provided in the datasheet, it looks like that could be the case. If thomasww is taking his measurements in a 50 ohm system and not matching the output, this might also explain his results.

• The datasheet states "Input and output internally matched to 50 Ω", so I've naively connected the input to a matched source and the output to an Agilent spectrum analyzer directly, and the measured results are:

However, I will run some simulations to check if the "not so 50 ohm" matching could cause the difference...

• Based on S22 from the datasheet at 1000Mhz, it looks to me like the real part of the output impedance for the ADL5535 is 31.5 ohms, with a reactance of -4.9 ohms. I have to admit to being a little confused here since the S-parameters are apparently in dB format and the the tabular values for S22 magnitude do not seem to correspond to the plotted values for return loss in the datasheet, particularly at the lower frequencies.

• My measurements were done az 160MHz, as I wrote before. According to an AWR simulation, the S parameters do not explain the missing dB-s in my case... So I'm still saying, something in this datasheet is definitely wrong.

• RF Amplifiers such as the ADL5355 are biased using a choke inductor at the output. The choke inductor behaves as a DC current source, setting the DC voltage at the output to Vdd, the supply voltage.This permits the device to swing between ground and 2Vdd. Hope that clear things up.

• Have you modeled the ADL5355 configuration shown in Fig. 25 of the datasheet as a baseline?

• For clarification, I should point out that a conjugate match based on the S-parameters will minimize reflected power, but it will generally not maximize power delivered to the load. Steve Cripps' "RF Power Amplifiers for Wireless Communications, 2nd Ed." has an excellent discussion of real world matching of power amplifiers (see Preface and  P. 19). If one assumes that "internally matched to 50 ohms" implies matching for minimum reflected power, then it also seems to imply that maximum power is not delivered into a 50 ohm load and that another load may be used for maximum power measurements. The basic question here is: What is the load used for P1dB measurements?

• The ADL5535 is indeed internally matched to 50ohms on the input and output ports. Keep in mind, it's a broadband match so there will likely be some variance in the impedance profile. Table 3 provides the de-embedded s-parmaters, while the Figure 5 provides the embedded s parameter profile which includes board losses.

I verified that the amplifier exhibits 18.5dBm at 190MHz of P1dB as specified in the datasheet using the evaluation board, signal generator and a spectrum analyzer. The source and load for the measurement was 50ohms. Please refer to my post dated Feb 6 for an explanation on how these amplifiers can exhibit higher compression levels.

If you aren't to duplicate this result, then it's possibly a bad part or a test setup issue. Have you duplicated this result on another device?

• Thomas,

The part will reach the datasheet specified p1 dB. Either the part is damaged or there an issue with your test setup.