I'm faced with the problem of explaining the harmonic production in a system where the second order intercept points of the cascaded components are unknown. As a way to bolster my explanation, I would like to provide some realistic sample calculations that might reasonably approximate what we observe the harmonic levels to be, but I don't have any IIP2 information except what I can get from the observations.

In other words, I would like to show that the observations agree (within some reasonable amount) with what we could expect from a cascade analysis based on the IIP2 and IIP3 values of the devices involved. (Or, alternatively, what we observe is inconsistent with what should be expected!)

I know from experience (IIRC) that for the devices we are using, a rule of thumb is that the IIP3 is generally about 15 to 16 dB above the 1 dB compression point.

However, I never needed such a rule of thumb for the IIP2 until now. Is there one? If someone can cite an authoritative reference in addition to providing the rule of thumb that would be even better. (I will have to be convincing with my analysis.)

The active devices involved are multi-octave class A RF amplifiers.

I understand how to do the cascade analysis and I am experienced in doing it for narrow bandwidth systems. The system at hand has a multi-octave bandwidth, so harmonic performance is an important issue. I am trying to characterize an individual stage. Generally, I know the 1 dB compression point and IIP3 for each stage.

(Could someone help with the tags on this questions?)

  • 2
    \$\begingroup\$ As far as I know for IIP2 there isn't such a rule of thumb as the actual value of the IIP2 depends on the internal architecture/circuit of an amplifier. If I design a balanced amplifier (maybe with differential in and outputs) I could achieve very low 2nd order distortion (it cancels out) making the IIP2 very high. \$\endgroup\$ – Bimpelrekkie Feb 16 '18 at 15:41
  • 1
    \$\begingroup\$ The same is true for IIP3 (3rd order distortion), in your case you know the amplifiers already so using that 15 dB above 1dB compression level is OK but realize that other amplifiers could behave differently. I can design an amplifier with feedback resulting in very low 3rd order distortion making IIP3 very high (and IIP2 as well). But RF amps. generally do not have feedback so in practice you would not come across this situation. \$\endgroup\$ – Bimpelrekkie Feb 16 '18 at 15:42
  • 2
    \$\begingroup\$ The mechanicsm for 2nd harmonic distortion is very different to the one for 1dB compression, so you would not expect to see a consistent relationship between the two, across different amplifiers or power classes. With balanced amplifiers, the degree of balance can vary wildly (from OK to accidentally excellent) in a normally operating amplifier, just the way that IP3 mechanisms don't. \$\endgroup\$ – Neil_UK Feb 16 '18 at 16:19
  • \$\begingroup\$ "I know from experience (IIRC) that for the devices we are using, a rule of thumb is that the IIP3 is generally about 15 to 16 dB above the 1 dB compression point." It is generally accepted that the rule of thumb is about 12 dB and this is the basic problem here. See this \$\endgroup\$ – Andy aka Feb 16 '18 at 18:14

If you know the other variables, you might be able to solve for IIP2.

Second order intercept point in cascading stages.

1/sqrt(OIP) = (1/sqrt(G2 x G3 x OIP1)) + (1/sqrt(G3 x OIP2)) + (1/sqrt(OIP3))

IIP = (OIP/(G1 x G2 x G3))


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.