The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl Simulators >> RF Simulators >> Direct upconversion mixer gain definition using spectre https://designers-guide.org/forum/YaBB.pl?num=1245846554 Message started by amr alaa on Jun 24th, 2009, 5:29am |
Title: Direct upconversion mixer gain definition using spectre Post by amr alaa on Jun 24th, 2009, 5:29am Hi We all know that the mixer is just a multiplier that can be modeled mathematically as V_out = m(t) * cos(wc*t) where: m(t) is the input modulating signal wc is the carrier frequency if we are upconverting from an IF-frequency then V_out = cos(w_IF*t) * cos(wc*t) = 0.5*(cos((wc+w_IF)t)+cos((wc-w_IF)t))) And we are interested in only one sideband. So, after filtering the unwanted side band V_out_filtered = 0.5*cos((wc+w_IF)t) However, if we are upconverting directly from baseband V_out = m(t)*cos(wc*t) From which we can conclude that the conversion gain of direct upconversion mixer = 2*conversion gain of IF upconversion mixer (after filtering) (Am I right at this point??) Anyway, I am using spectre to simulate the direct up-conversion gain of a gilbert cell. I am also using PSS+PAC method in which the PSS fundamental = flo and in the PAC, I am adjusting the input source's AC magnitude =1 so the output voltage is the conversion gain directly. So, does the resulting conversion gain is that corresponding to DIRECT CASE or the IF CASE???? |
Title: Re: Direct upconversion mixer gain definition using spectre Post by pancho_hideboo on Jun 24th, 2009, 6:11am amr alaa wrote on Jun 24th, 2009, 5:29am:
Consider spectrum components of m(t). This is no more than IF UpMixer except for DC component. m(t)*cos(wc*t) -> 0.5*M(ω+ωc)+0.5*M(ω-ωc) Here M(ω) is spectrum of m(t). amr alaa wrote on Jun 24th, 2009, 5:29am:
Your Mixer is single I-UpMixer not I/Q UpMixer. People often have confusions about gain definition of I/Q UpMixer. However there is no confusion for I-UpMixer. |
Title: Re: Direct upconversion mixer gain definition using spectre Post by amr alaa on Jun 24th, 2009, 8:44am Thank you pancho_hideboo I got your point. I think in both cases of direct up-conversion and IF up-conversion the mixer operation is the same. However, in the direct up-conversion mixer we are interested in the whole up-converted signal. However, in the IF up-conversion case, we are interested in only one side band of the signal (either the upper side band or the lower side band). This can be shown from the following mathematical relations: The IF up-conversion IF signal = m(t)*cos(w_IF*t) The transmitted signal = IF signal *cos(wc*t) = m(t)*cos(w_IF*t)*cos(wc*t) = 0.5*m(t)*(cos((wc+w_IF)t)+cos((wc-w_IF)t))) which has a spectrum of --> 0.25*(M(w-(wc+w_IF))+M(w+(wc+w_IF))+M(w-(wc-w_IF))+M(w-(wc-w_IF))) we are interested in the signal with spectrum (for example): 0.25*((M(w-(wc+w_IF))+M(w+(wc+w_IF)) or in time domain = 0.5*m(t)*cos((wc+w_IF)) The direct up-conversion Baseband signal = m(t) The transmitted signal = m(t)*cos(wc*t) which has a spectrum of --> 0.5*M(w+wc)+0.5*M(w-wc) However, we are interested in the hole signal which is =m(t)*cos(wc*t) So,apparently there is a factor of 2. Please, pancho_hideboo, I am confused could you explain it more to me. Also, If the spectre is calculating the IF conversion case. How can I calculate the direct up-conversion case?? Thanks for your time inadvance :) |
Title: Re: Direct upconversion mixer gain definition using spectre Post by pancho_hideboo on Jun 24th, 2009, 9:22am amr alaa wrote on Jun 24th, 2009, 8:44am:
amr alaa wrote on Jun 24th, 2009, 8:44am:
In double conversion, IF spectrum is same as direct conversion. Here two mixers and two local signals are needed. if(t)=m(t)*cos(ωIF*t) -> IF(ω)=0.5*M(ω+ωIF) + 0.5*M(ω-ωIF) if(t)*cos(ωc*t)=m(t)*cos(ωIF*t)*cos(ωc*t) -> 0.5*IF(ω+ωc) + 0.5*IF(ω-ωc) =0.5*0.5*M(ω+ωIF+ωc)+0.5*0.5*M(ω-ωIF+ωc) + 0.5*0.5*M(ω+ωIF-ωc)+0.5*0.5*M(ω-ωIF-ωc) amr alaa wrote on Jun 24th, 2009, 8:44am:
amr alaa wrote on Jun 24th, 2009, 8:44am:
If you understand correctly, there is no confusion. In double conversion, you need two mixers and two local signals. Again I phrase the following. Your Mixer is single I-UpMixer not I/Q UpMixer. People often have confusions about gain definition of I/Q UpMixer. However there is no confusion for I-UpMixer. |
Title: Re: Direct upconversion mixer gain definition using spectre Post by amr alaa on Jun 24th, 2009, 1:09pm Thaks for the reply pancho_hideboo In the double conversion architecture you mentioned. There is basically two mixers, one is to up-convert from baseband to the IF frequency (which behaves identically to the direct up-conversion case) while the other is to up-convert from the IF frequency to the RF frequency. If we rephrased my previous questionas follow. If I used identical circuit topology for both mixers, will they have the same conversion gain given that the second mixer will generate the output @ ωc+ωIF & ωc+ωIF (and their negative frequency counter parts) and we will take only one sideband and filter the other. Can you please also refer me to a good book or tutorial that elaborate on this issue. Thanks for your time |
Title: Re: Direct upconversion mixer gain definition using spectre Post by pancho_hideboo on Jun 24th, 2009, 3:03pm amr alaa wrote on Jun 24th, 2009, 5:29am:
Your "IF CASE" means "Double Conversion". Your "DIRECT CASE" means "Single Conversion". Your simulation setup in the above is "Single Conversion". amr alaa wrote on Jun 24th, 2009, 1:09pm:
amr alaa wrote on Jun 24th, 2009, 1:09pm:
Do you mean difference of conversion gain between LSB and USB ? If ωIF is large, they are quite different. amr alaa wrote on Jun 24th, 2009, 1:09pm:
Any RF textbook for very beginner is helpful. What procedure do you take in actual measurement using actual instruments ? If you understand actual measurement correctly, there is no confusion. You have to learn measurements using actual instruments. Not "EDA Tool Play". |
The Designer's Guide Community Forum » Powered by YaBB 2.2.2! YaBB © 2000-2008. All Rights Reserved. |