Suppose that I'm working on a system which needs to use a transceiver with an FM signal. If that system moves very fast (i.e. it is a satellite or whatever) in which the maximum doppler effect can be comparable with frequency spacing, do I need an higher spacing than the doppler? Or it is not related at all? Or the usefulness of simply changing the spacing is related only to achieve a better insulation between other frequencies, and/or obtain an higher SNR, and therefore nothing to do with the doppler?

Thank you, and sorry for the dummy question. :)


When moving at 1km per second (very fast for anything using a road) a radio wave of carrier frequency 100 MHz will appear to shift to 99.9997MHz if traveling away from the source of the transmission. This is a deviation of 300 Hz.

Satellites (such as ISS) travel at 7.66 km/sec so therefore the 100MHz appears to be 99.9974MHz i.e. a deviation from the nominal carrier centre of 2.6 kHz.

Typical FM communication channel separation might be as low as 12.5 kHz so there's good reason to suspect high speed being a potential cause of poor reception BUT, at an offset of 2.6KHz I believe an FM receiver should easily cope with aligning itself with the apparent off-centeredness of the transmission. They use PLLs for keeping aligned to transmissions so I don't think it will be a problem.

For a simple radio system like one running at 434MHz there will be a dc shift in the demodulation output but given that a lot of these transceivers are low precision (relying on a data slicer for extracting the data from the raw demod output), I don't see it as an issue. They have a receive bandwidth of over 1MHz so it should be easy to see it won't be an issue.

  • \$\begingroup\$ I've an CC10xx family datasheet but it do not talk about any capability to keep trace of the frequency doppler. Or at least do not mention it in words that I'm aware of. Some university cubesats are using frequencies around 435MHz with that transceiver chip, and doppler effect at those orbits can reach 10kHz with a channel spacing of 12.5kHz. How can I understand if this is a reliable condition? \$\endgroup\$ – thexeno Feb 26 '15 at 14:00
  • \$\begingroup\$ With a channel spacing of 12.5kHz and an offset of 10kHz, providing there isn't another transmission that can superimpose there will still be no problem providing the software in the receiver recognizes what is happening and chooses the next frequency allocation to receive on. It won't be a hardware problem but one of managing the doppler scenario in software I believe. \$\endgroup\$ – Andy aka Feb 26 '15 at 14:13
  • \$\begingroup\$ Thanks. If you can, may I have some suggestions on some readings about this subject? On internet I find stuff which is or too basic or not related to this. Thanks anyway \$\endgroup\$ – thexeno Feb 26 '15 at 14:20
  • \$\begingroup\$ I can't recommend a book because I use the internet for all my data and info that I learnt many years ago! \$\endgroup\$ – Andy aka Feb 26 '15 at 14:23
  • \$\begingroup\$ I've found, finally, that seems that if I don't want to cope with that in software and assuming that this is not handled by the transceiver, I need to extend the filter BW to cover the shift in reception. But it decrase the sensitivity. It is always a tradeoff. \$\endgroup\$ – thexeno Apr 15 '15 at 14:48

If all these channels are on the same object, then they will all be doppler shifted together, and they won't interfere with each other any more than they do without doppler shift.

The receiver does have to tolerate and be able to lock onto the signals despite the apparent carrier frequency drift caused by the doppler shift.

  • \$\begingroup\$ There are keywords to find in datasheet to understand if the transceiver's hardware can trace this shift? For example, I was studying the CC1020, and I don't see stuff related to frequency shift, at least not with words that I'm aware of. \$\endgroup\$ – thexeno Feb 26 '15 at 15:17
  • \$\begingroup\$ I suspect that there's no provisions in the CC1020 to account for doppler shift. To account for doppler shift, you need to be able to derive a frequency offset from the demodulator, and feed that back into the receiver to adjust the local oscillator. This could be done externally with tweaking the crystal oscillator, but that probably would have other effects. \$\endgroup\$ – rfdave Mar 21 '15 at 14:59

According to a CC1020 datasheet (but could be also any other similar device), the frequency deviation is the channel frequency separation divided by half.

From data-sheet:

"The signal bandwidth must be smaller than the available receiver channel filter bandwidth. The signal bandwidth (SBW) can be approximated by (Carson’s rule): SBW = 2 ∙ fm + 2 ∙ frequency deviation"

I'm interpreting this as considering in the bandwidth of received signal also the imperfections (tolerance, medium, DOPPLER etc).

They report also ChBW > SBW + 2 ∙ f_error, where f_error is the error bring from oscillator. It seems that my answer to my question is: yes, you need an higher bandwidth taking into account the doppler which extends the bandwidth, but as a consequence the frequency spacing must be higher than that.

If this is right, I may need a correction on what exactly is the difference between the frequency separation and frequency spacing.


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.