On Sep 20, 8:18 am, fl wrote:
Hi,
I am learning some RF knowledge. The following is a post years ago on
this group. I don't understand how the last poster got -89.5 dBM.
Could you tell me that? Thanks in advance.
....
In the example using the Mini-Circuits amplifiers, the output 3rd
order
intercept point will remain at essentially +47 dBM. With an output
signal at
the +1.5 dBM level, the distortion products will be at a level of
-89.5 dBM.
This is only 91 dB down from the output level and not the 100 dB
desired.

OK, how far did you get in the calc? If the 3rd order intercept is
+47dBm, and the input is +1.5dBm, note that the TOI is 45.5dB higher
than the input. For that to be the case, since 3rd order products
increase in amplitude at three times the rate (in dB) as the input
levels, and the input levels of the two tones used in the third order
intermod test are assumed to be the same level, the distortion will be
three times as far down as the two tone levels, measured from the
intercept point. So for that +1.5dBm input, the third order intermod
products are 3*45.5dB below the intercept point, or 2*45.5dB below the
two tones.

Note that this is from an ideal model where the output y(t) is related
to the input x(t) as y(t) = a*x(t) + b*x(t)^2 + c*x(t)^3. Second
order products, like the second harmonic and the sum and differences
between frequencies, are caused by the x^2 term, and third order
products like the third harmonic and 2*f1-f2 (the usual third order
intermod products) are caused by the x^3 term. But in actuality, the
relationship also has higher order terms: x^4, x^5, and so forth.
The x^5 term will also contribute to third-order intermod, and that
causes the simple relationship of "increase the input by 1dB and the
third order products go up by 3dB" to not strictly hold true.
Amplifiers operated well below saturation generally conform very
closely to the 3-to-1 rule, but ADCs and DACs most certainly do not,
and neither do amplifiers that are driven into clipping.

Cheers,
Tom