AD8338 compression?

I need to figure out how much power I can input into the AD8338, but the 1dB compression point nor the IP3 is listed on the datasheet. Is there anyway to figure this out without actually having the chip in front of me and increasing input power til it goes nonlinear? I know the 1dB compression point is dependent on the gain, which is variable in this case, so honestly I'm not even sure if it's possible to actually have one. Any clarification on this topic would be greatly appreciated.

Also, I need the noise figure for the chip, but they give the noise density on the datasheet and I don't really understand how to interpret it.

Parents
  • 0
    •  Analog Employees 
    on Aug 17, 2016 6:22 PM

    Hi John,

    Thank you for considering AD8338 in your design.

    Yes, you are right that the P1dB and OIP3 of AD8338 greatly depends on the Gain. A very good representation of those parameters is shown in Figures 23 and 24.

    Please be careful on the output. AD8338 is linear up to 1V (0.707 Vrms) only, beyond this level the output gets really distorted and is not very useful for systems that require high linearity. However, for systems where linearity is irrelevant and the only concern is the dominant frequency content, then it can go up to 1.4V (1 Vrms).

    And for the Noise figure, you can't find it anywhere in the datasheet since it is as well variable. This parameter greatly depends on the Gain, frequency and input configuration.

    The gain equation for AD8338 is:

    Gain (dB) = 80*VGAIN +20log(9500/Rin) – 34

    For a 500ohm input resistors, and a maximum VGain=1.1V, we will get 80 dB gain.

    Now, at VGain=1.1V the Voltage noise is 4.5 nV/√Hz. This means that on the output we will have 4.5 nV/√Hz *80 dB (4.5 nV/√Hz * 10,000) = 45 uV/√Hz of noise. Suppose we have a 1kHz bandwidth filter, that would be 1.423 mVrms noise. The maximum linear output is 1V (0.707 Vrms). This wil give us 54 dB SNR.

     

    Regards,

    Jino

Reply
  • 0
    •  Analog Employees 
    on Aug 17, 2016 6:22 PM

    Hi John,

    Thank you for considering AD8338 in your design.

    Yes, you are right that the P1dB and OIP3 of AD8338 greatly depends on the Gain. A very good representation of those parameters is shown in Figures 23 and 24.

    Please be careful on the output. AD8338 is linear up to 1V (0.707 Vrms) only, beyond this level the output gets really distorted and is not very useful for systems that require high linearity. However, for systems where linearity is irrelevant and the only concern is the dominant frequency content, then it can go up to 1.4V (1 Vrms).

    And for the Noise figure, you can't find it anywhere in the datasheet since it is as well variable. This parameter greatly depends on the Gain, frequency and input configuration.

    The gain equation for AD8338 is:

    Gain (dB) = 80*VGAIN +20log(9500/Rin) – 34

    For a 500ohm input resistors, and a maximum VGain=1.1V, we will get 80 dB gain.

    Now, at VGain=1.1V the Voltage noise is 4.5 nV/√Hz. This means that on the output we will have 4.5 nV/√Hz *80 dB (4.5 nV/√Hz * 10,000) = 45 uV/√Hz of noise. Suppose we have a 1kHz bandwidth filter, that would be 1.423 mVrms noise. The maximum linear output is 1V (0.707 Vrms). This wil give us 54 dB SNR.

     

    Regards,

    Jino

Children
No Data