In my application a down-conversion of the RF band (ranging from 6.0GHz to 9.0GHz) to the IF band (from 1.0 to 4.0 GHz) is operated via the LTC5549 by enabling the mixer's internal duplicator and feeding the LO Port with a local oscillator signal at fLO = 5.5GHz.
The problem concerns the presence of a second-order intermodulation products that are generated at the frequency (fIN – fLO) and at the frequency (fIN + fLO) which has an amplitude of -6 to -10dBc relative to (2*fLO – fIN) which is the useful signal at the IF Port.
If by hypothesis fIN = 7GHz, I get at the output of the mixer a down-converted signal at fIF = 4GHz but also two IMD2 products: the first at the frequency (fIN – fLO) = 1.5GHz and the second at (fIN + fLO) = 12.5GHz.
So, if at the IF Port output of the EVB module I have the useful signal at 4GHz with an amplitude of, for example, -40dBm, I also observe other signals at 1.5GHz and 12.5GHz with an amplitude of -46dBm to -50dBm, depending on the frequency.
Incidentally, I point out that if the input signal decreased e.g. by 10dB, the spurious would also decrease by the same amount since this comes from second-order linearity.
This behaviour does not show up at all if the built-in duplicator is disabled and the 11GHz LO signal feed directly the LO Port of the mixer.
These spurious patterns are generated only if the mixer's internal duplicator is used in Down Conversion with High Side LO scheme.
The strange thing is the fact that the Data Sheet neither specifies nor reports anything about that. Am I wrong?
I also wonder if you have any knowledge of this strange phenomenon which seriously prevent the application of the LTC5549 in Down Conversion, High Side Injection LO Schemes when the internal frequency duplicator is enabled.
Thank you for your attention.
Pier
