I have an SDR receiver I'm interested in making into a FMCW radar. The bandwidth is somewhat narrow as it only samples at 64 MSPS.

There is a problem in that without a bunch of embedded programming the RX and TX chains won't have their timing synchronized. Basically the up-conversion oscillator is synchronized but the DSP portion on the TX and RX lines won't be.

What I'm wondering is, is there a way to use the initial signal sent from the transmitter and received by the receiver through direct LOS to establish timing offset for the reflected signal? Given that the RX and TX antenna will be close together you could use this to synchronize the system.

Also I'm not sure what the processing would consist of though(or if it's even viable), most FMCW radar mix the transmitting signal with the reflected signal and then downconvert to produce to a local beat. I would be attempting to do this through DSP and have a much narrower bandwidth useable.


We did the same a couple of years back with what we called "echotimer", but not form FMCW, but for OFDM radar: basically, if you have a very large target that's very close and at 0 doppler, that's your crosstalk, and you can just subtract the distance of that.

Same applies here: with TX and RX hopefully sharing the LO, you're looking for a zero-velocity target that's close. You determine your range based on that, it represents your RX-TX time offset, and you just subtract.

  • $\begingroup$ Thanks, could you explain this a little more? I should have been more precise in that the up/down conversion oscillator is synced between the transmit/receive lines. It's just that the transmit and receive lines aren't digitally synced so when you start receiving could be before/after you start transmission. $\endgroup$ – FourierFlux May 7 '20 at 19:26
  • 1
    $\begingroup$ exactly. So, you do a radar measurement. That's unsynchronized, but that doesn't matter: it shows the TX/RX cross talk as radar target. The distance of that that "fake" target is your TX / RX timing offset. Exactly as I explained! $\endgroup$ – Marcus Müller May 7 '20 at 19:44
  • $\begingroup$ Ok good, yes that makes sense. But in terms of processing, have suggestions on the best way to go about this? I'm also slightly worried about the crosstalk saturating the receiver but I think with directional antenna it should not be as big of a problem. Do you have suggestions? $\endgroup$ – FourierFlux May 7 '20 at 20:00
  • $\begingroup$ I don't think I can be any more explicit than I've been: estimate closest target, subtract distance. done. $\endgroup$ – Marcus Müller May 7 '20 at 20:01
  • 1
    $\begingroup$ @FourierFlux if this is of interest to you: Here's the relevant dissertation. $\endgroup$ – Marcus Müller May 7 '20 at 20:13

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.