How to decide the input power of the RF source and the frequency space for LNA?

[jameslee]

Re: IP3 measurement

sorry for the missing info,

im working on the LNA,gain at 18dB, i need to figure out the P1dB first.
the spectrum analyzer attenuation setting i set at very high value(50dB), i know that if my attenuation level is too low,the IM3 will appear to be very high,usually what i do is i inject two -10dBm tone thru a power splitter, then slowly reduce the SA attenuation bit by bit until i see the IM3 product stay fix,and i further increase the attenuation by 10dB so that the IM3 product will not likely to be cause by the SA.

to BigBoss,
what i mean is,lets say i give you an example,
1) at -10dBm input level, i getting delta_IM3 product at -80dB, so my IIP3 is +30dBm right?
2) however when at -20dBm input level, i still getting delta_IM3 at -80dB, that will give me a IIP3 of +20dBm

im always thinking the IIP3 should be very much constant until it reaches compression point, but in this case im seeing different senario, as if higher input power will give me higher IIP3, but this should not be the case.

i wondering whether there is something else that causing this inaccurate measurement.

James

 

[BigBoss]

Re: IP3 measurement

to BigBoss,
what i mean is,lets say i give you an example,
1) at -10dBm input level, i getting delta_IM3 product at -80dB, so my IIP3 is +30dBm right?
2) however when at -20dBm input level, i still getting delta_IM3 at -80dB, that will give me a IIP3 of +20dBm

im always thinking the IIP3 should be very much constant until it reaches compression point, but in this case im seeing different senario, as if higher input power will give me higher IIP3, but this should not be the case.

i wondering whether there is something else that causing this inaccurate measurement.

James

To be able to get the right OIP3, output level will much below from P1dB (at least 10dB) .This equation is valid only when the amplifier is in linear region.If amplifier has already entered into saturation or near saturation region, this equation is not valid anymore.
In your case, I think your amplifier has already entered into saturation region that's why when you increase your input level the gain drops off and consequently neither first harmonic nor f1+-2f1 or 2f1+-f2 is not increased, remains constant.Some nonlinearities may show such characteristics.

 

[RCinFLA]

Re: IP3 measurement

IP3 should follow the track of 3 db spur increase for every 1 db increase of fundamental tones. The projection of the slope of these two lines defines the imaginary intercept point.

When spur levels reach a point where they are about -25 db down from two tone levels the relationship will start to signficantly deviate.

You should measure several levels to verify the 1 db for 3 db relationship. When it departs significantly you have gone too high on the two tone level.

---------- Post added at 16:04 ---------- Previous post was at 15:47 ----------

sorry for the missing info,

im working on the LNA,gain at 18dB, i need to figure out the P1dB first.
the spectrum analyzer attenuation setting i set at very high value(50dB), i know that if my attenuation level is too low,the IM3 will appear to be very high,usually what i do is i inject two -10dBm tone thru a power splitter, then slowly reduce the SA attenuation bit by bit until i see the IM3 product stay fix,and i further increase the attenuation by 10dB so that the IM3 product will not likely to be cause by the SA.

to BigBoss,
what i mean is,lets say i give you an example,
1) at -10dBm input level, i getting delta_IM3 product at -80dB, so my IIP3 is +30dBm right?
2) however when at -20dBm input level, i still getting delta_IM3 at -80dB, that will give me a IIP3 of +20dBm

im always thinking the IIP3 should be very much constant until it reaches compression point, but in this case im seeing different senario, as if higher input power will give me higher IIP3, but this should not be the case.

i wondering whether there is something else that causing this inaccurate measurement.

James

IP3 of spectrum analyser, due to its overload, is a common mistake.

The spur levels should be at least ten db above spectrum analyser display noise floor. Narrow up the spectrum analyser RF bandwidth will decrease display noise floor.

Another common mistake is not having enough isolation/attenuation between input generators. One generator's signal will backup into other generators's ALC amp causing it to IM.

Avoid electronic PIN diode switched step attenuators unless you have measured their IM performance and know is much higher then device you are measuring.

 

[bluetooth]

Re: IP3 measurement

@jameslee: The IMD3 is 2*f1 -f2. Probably you apply the harmonic of f1 direct in the lna?. Did you use an lpf to cut off the harmonic off your signal source generator?