Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
simply inject 2 signals with frequcy spacing very close to ur basband , via a power combiner and check the output , and then plot the curves and extrapolate the results to get the point
IP2 is simply 2nd harmonic test. You drive the DUT at various power levels, and measure the second harmonic. If you are looking for critical measurements, you should lowpass filter the frequency source, and highpass the input to the spectrum analyzer--so you get only the DUT's 2nd harmonic.
IP2 is simply 2nd harmonic test. You drive the DUT at various power levels, and measure the second harmonic. If you are looking for critical measurements, you should lowpass filter the frequency source, and highpass the input to the spectrum analyzer--so you get only the DUT's 2nd harmonic.
IP2 is also a cross modulation, an important parameter in direct conversion as two strong outband RF signals can result in inteference in baseband passband. First reply method is correct but usually done with two signals on either side of intended passband to avoid IP3 product showing up in baseband passband.
Instinctly, IP2 means when 2 tones are fed into the dut, some distortion with 2nd order can then be detected on the output.
In formula, f1+/-f2+/-FLO=+/-Fif
In NZIF and ZIF system, we can find some 2nd order intercept points.
Suppose RF frequency is Frf, RF span is Fspan, Intermediate frequency is Fif and signal band width is from 0 to Frbw. We have,
1. the most popular IP2
F1= Frf-Fdist, F2= Frf-Fdist+Fif, where Fdist <Fspan/2.
2. The 2nd popular IP2
single tone F1= Frf/2+Fif/2.
(For TX ,F1=Frf+fif, and test at 2*Frf+2*Fif.)
3. F1=Frf-Fdist, F2=Frf-Fdist+Fif, where Fdist < Frbw/2.
4. F1=Frf*3/2+Fif*3/2+Fdist, F2=Frf/2*3+Fif*3/2+Fdist.
IP2 is also a cross modulation, an important parameter in direct conversion as two strong outband RF signals can result in inteference in baseband passband. First reply method is correct but usually done with two signals on either side of intended passband to avoid IP3 product showing up in baseband passband.
"In fact, IP2 stays the same if f1 and f2 are identical in frequency. Hence, we can use a single signal generator, not two."
Not exactly. Two signals will create a modulated nulling pattern and single generator does not (unless it is a DSB suppressed carrier modulated signal). It is the high amplitude modulation that creates the IP2 product.
Measuring 2nd harmonic is pretty useless for finding out actual semiconductor amp IP2 at RF frequencies since there is almost alway some matching circuitry that has a frequency response. You will be measuring the input /output filter response attenuation of second harmonic not just the device IP2.
Most of the interest in IP2 measurements is for direct conversion mixers which are very vunerable to IP2 and must have very good IP2 performance. In cellular, interference from EDGE modulation is more damaging then GMSK modulation when there is poor IP2 mixer performance because EDGE has higher peak to average power ratio (AM modulation).
For cellular, say WCDMA direct conversion mixer, two generators on same side of mixer center frequency with 100 kHz separation, centered one MHz away from mixer input center frequency, will create 100 KHz spur in baseband output. It is important that the two generators' amplitudes are accurately matched which creates deep amplitude null two tone response. Mismatch in two generator amplitude greatly reduces nulling depth, making IP2 look better then it is.
I believe you. I think the answer is, "what are you using the IP2 measurement for?". In my work, which is often frequency synthesizer design, I am most interested in what harmonics, especially 2nd and 3rd, that a component might create. If I pass a high level 10 Ghz signal thru an amplifier, I really hate it when an almost equally strong 20 GHz unwanted spur shows up.
But I can see that if you were receiving complex modulation, you might have a different concern.
[RCinFLA] Not exactly. Two signals will create a modulated nulling pattern and single generator does not (unless it is a DSB suppressed carrier modulated signal). It is the high amplitude modulation that creates the IP2 product.
[cnm] Why two signals created a modulated nulling pattern? Are you referring to the beat tone pattern from two signals that are close in frequency?
[RCinFLA] Measuring 2nd harmonic is pretty useless for finding out actual semiconductor amp IP2 at RF frequencies since there is almost alway some matching circuitry that has a frequency response. You will be measuring the input /output filter response attenuation of second harmonic not just the device IP2.
[cnm] Not true. Lots of multi-octave systems do care 2nd harmonic IP2 more.
BTW, which amp does not have input/output matching? Measuring device IP2 is great. But eventually the complete amplifier (with matching networks)'s IP2 is what matters, right? Who is going to use a bare device by itself?
I seriously suspect that the matching network affect IP2 much. Most passive matching networks are practically consider as linear components. Normally they don't contribute to second order distortions. Yes, they do have attenuation. However, considering that IP2 is an mathematically extraplated figure of merit. Measuring IP2 at a lower power level (considering matching network's loss) vs. at a high power level (removing matching network's loss), does it really make much difference in figuring out the IP2 intercept point?
[RCinFLA] Not exactly. Two signals will create a modulated nulling pattern and single generator does not (unless it is a DSB suppressed carrier modulated signal). It is the high amplitude modulation that creates the IP2 product.
[cnm] Why two signals created a modulated nulling pattern? Are you referring to the beat tone pattern from two signals that are close in frequency?
[RCinFLA] Measuring 2nd harmonic is pretty useless for finding out actual semiconductor amp IP2 at RF frequencies since there is almost alway some matching circuitry that has a frequency response. You will be measuring the input /output filter response attenuation of second harmonic not just the device IP2.
[cnm] Not true. Lots of multi-octave systems do care 2nd harmonic IP2 more.
BTW, which amp does not have input/output matching? Measuring device IP2 is great. But eventually the complete amplifier (with matching networks)'s IP2 is what matters, right? Who is going to use a bare device by itself?
I seriously suspect that the matching network affect IP2 much. Most passive matching networks are practically consider as linear components. Normally they don't contribute to second order distortions. Yes, they do have attenuation. However, considering that IP2 is an mathematically extraplated figure of merit. Measuring IP2 at a lower power level (considering matching network's loss) vs. at a high power level (removing matching network's loss), does it really make much difference in figuring out the IP2 intercept point?
IP2 means Intermodulation Product of second order. Not exclusively second harmonic. Yes, multi-octive bandwidth amps are concerned with second harmonic distortion.
If your interest is 2nd harmonic output then measure second harmonic.
If you interest is 2nd order intermodulation then measure 2nd order intermod product.
The first is not equal the second and the second is not equal the first.
You can't use a second harmonic test to measure a mixer's IP2 performance.
You can't use an IP2 two tone test to measure second harmonic performance of a saturated mode, excitation biased, RF PA.
I am often concerned with 2nd order distortion as I design a lot of wideband receiver products. When trying to receive very small signals it is IP2 terms that often pop up first to cause interference (compared to IP3).
So I often have to measure 2*F1=RF or F2+-F1 = RF or 2*F1=IF or F1+-F2=IF etc
In reality I end up making lots of 2nd order measurements across the range whilst looking for the weakest link.
I am often concerned with 2nd order distortion as I design a lot of wideband receiver products. When trying to receive very small signals it is IP2 terms that often pop up first to cause interference (compared to IP3).
So I often have to measure 2*F1=RF or F2+-F1 = RF or 2*F1=IF or F1+-F2=IF etc
In reality I end up making lots of 2nd order measurements across the range whilst looking for the weakest link.
IP2 means Intermodulation Product of second order. Not exclusively second harmonic. Yes, multi-octive bandwidth amps are concerned with second harmonic distortion.
If your interest is 2nd harmonic output then measure second harmonic.
If you interest is 2nd order intermodulation then measure 2nd order intermod product.
The first is not equal the second and the second is not equal the first.
You can't use a second harmonic test to measure a mixer's IP2 performance.
You can't use an IP2 two tone test to measure second harmonic performance of a saturated mode, excitation biased, RF PA.
Not sure why you say "You can't use a second harmonic test to measure a mixer's IP2 performance."?
For multi-octave mixers, we have to measure second harmonic IP2. And there is way to do it, similarly to the amp 2nd harmonic IP2 measurement.
Search Agilent's website, there is plenty information on how to set up the harmonic IP2 measurment for mixers.
I need to measure IIP2 in 500 MHz. I can use two signals: 252 MHz and 248 MHz or I can use 750 and 250 MHz, there are many options. My system is wideband.
What defines the correct frequency of IIP2 measurment?
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
