1
$\begingroup$

I am having trouble understand what I think should be a pretty simple concept. Conceptually I can understand that a signal that has been down-converted, and then delayed in time is different than a signal that has been delayed in time and then down-converted.

I am trying to compensate for this difference on an FPGA (I am down-converting the signal, delaying in time, and then up-converting)

The way I see it, the delayed (and attenuated signal) looks like: $A e^{i(\omega_{\text{c}}-\omega_{\text{lo}})(t-\tau_1)} m(t-\tau_1)$

That is different from a down-converted delayed signal, so I need a phase correction constant: $\phi = e^{i\omega_{\text{lo}}\tau_1}$

Which gives: $A e^{i(\omega_{\text{c}}-\omega_{\text{lo}})t - i\omega_{\text{c}}\tau_1} m(t-\tau_1)$

So the correction is a function of the LO and the delay itself, right? I have my sampled signal in I and Q, but I need to multiply it by the correction before up-converting, and that is where I am lost.

| improve this question | |
$\endgroup$
2
$\begingroup$

Let your original signal be (leaving out any constant multiplicative factor):

$$s(t)=e^{i\omega_ct}m(t)$$

Then the down-converted signal is

$$\tilde{s}(t)=e^{i(\omega_c-\omega_{lo})t}m(t)$$

and its delayed version is

$$\tilde{s}(t-\tau)=e^{i(\omega_c-\omega_{lo})(t-\tau)}m(t-\tau)$$

whereas the delayed and down-converted signal is

$$s(t-\tau)e^{-i\omega_{lo}t}=e^{i\omega_c(t-\tau)}e^{-i\omega_{lo}t}m(t-\tau)= \tilde{s}(t-\tau)e^{-i\omega_{lo}\tau}$$

So the correction factor $e^{-i\omega_{lo}\tau}$ indeed depends on both $\omega_{lo}$ and the delay $\tau$.

| improve this answer | |
$\endgroup$
  • $\begingroup$ OK, thank you. So where do I got from there? If I have an LO of 2437MHz say, and a delay of 15ns, I am strugging to see what the I/Q adjustment would be. I = sin(2437e6*15e-9) and Q = cos(2437e6*15e-9)? $\endgroup$ – toozie21 Jul 1 '14 at 17:02
  • $\begingroup$ @toozie21: If I understand your problem correctly, you need to apply the complex correction factor to the complex (combined) IQ signal, i.e. $(I+iQ)\cdot C$, where $C$ is the correction factor. By taking the real and imaginary parts of this corrected signal you get the new I and Q components. $\endgroup$ – Matt L. Jul 1 '14 at 17:23
  • $\begingroup$ I think you are on track with what I am wanting. So in this case you are thinking that I would have a C of 36.55 and I would multiply that by I and iQ? Am I understanding your suggesting right? $\endgroup$ – toozie21 Jul 1 '14 at 17:29
  • $\begingroup$ @toozie21: $C=e^{-i\omega_{lo}\tau}$ is the complex correction factor, so it can't be 36.55, but it is of the form $x+iy$ with $x=\cos\omega_{lo}\tau$ and $y=-\sin\omega_{lo}\tau$. $\endgroup$ – Matt L. Jul 1 '14 at 18:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.