Direct conversion problem.

[wccheng]

Dear all,

I have a question about the direct conversion. I hope anyone could answer my question. Thanks first.


Q.1 For example, I have a RF signal band 900MHz +/- 10MHz. For the direct conversion, the LO should be 900MHz or not? If LO=900 MHz, there are 10MHz is in the negative frequency side after down conversion. How this problem could be solved? Or, LO should be 890MHz instead?

Q.2 For the mixer simulation (direct conversion), if RF=900MHz and LO=900MHz, there are a large DC at 0Hz. So, how could I simulate the down converted frequency result?

Thanks

chi

 

[noiseless]

Q1: people usually use I and Q paths to eliminate the negative frequency problem. you can use sine and cos to calculate it.

Q2: people use somekind of HP filter to eliminate the DC offset.

 

[wccheng]

Dear noiseless,

Could you give me some materieals to me? I want to learn it more. I have tried to see "RF Microelectronics" book. It does not give me any information about my questions.

thanks
wccheng

 

[artem]

Search forum .

**broken link removed**

 

[wccheng]

Dear all,

I have read the paper suggested by the previous replied mail. However, I am still confusing it.

The paper stated that using I/Q method in order to solve the negative-frequency half-channel problem. I am not understanding that after I/Q path and doing the addition, the direct down-conversion negative-frequency half-channel cancel out and positive-frequency half-channel is double in amplitude. Then the half signal band's data will loss? How could this negative-frequency half-channel signal recover after down conversion?

Actually, we could get the negative frequency signal in the mathematical calculation. However, it could not see in the real measurement. Am I right?

thanks

wccheng

 

[artem]

https://www.highfrequencyelectronics.com/Archives/Jun04/HFE0604_DirectConv.pdf
**broken link removed**

 

[stober]

The nagative half is cancelled because it is useless. but if it's useful, then you would rather adopt lower LO Frequence like 868MHz than 900MHz.

 

[wccheng]

Dear all,

Why will you say the negative half is useless? Could you explain it more?

For the direct conversion mixer design, how could I simulate it? If I give RF=900MHz and LO=900MHz, the IF=0Hz and there is a stronge DC component and I could not get the real down converted signal strength.

Moreover, "noiseless" stated that someone will use High-pass filter to eliminate the DC offset. However, I have seen someone using ADC to cancel out the DC component. Which one is correct?

thanks

wccheng

 

[wccheng]

Dear all,

I have seen several books and paper. They just mention to use I/Q method in the zero-IF conversion in order to recover the negative half-side frequency band signal. However, it does not explain in detailed. I also don't know how could it work. Could anyone give me some helpful materials to explain this?

thanks

wccheng

 

[mesfet]

Hi 智者,

For simplicity, consider cos and sin. Do FFT on cos and sin, take the absolute value. What do you get? They are the same! All real signal are symmetrical on the frequency domain. Now, take into consideration of the phase information of the FFTed cos and sin. You can now see the difference between cos and sin. Basically, the idea of IQ is to distinguish cos and sin. From there, the what so called "negative frequency" information can be detected.


Mesfet+