I am trying to understand if it's feasible to transmit -1,0,+1 data using FM on 20-30m wavelength to be received 10,000km away. From Europe to US.

Ultimately I would like to create a solution which is sensitive to wave slope. I know it is possible as I read GPS time signal works this way. Hence I can transmit trinary signal not binary

  • What SNR should I expect?
  • What kind of power on TX side?

"Desired performance" here is (if the trinary data achievable as above):

  • latency is 4ms
  • data rate 2 tribit per second
  • as its not commercial i'd rather calibrate optimal missing reception etc per session (dependent on weather etc), so 20-80%..

Also note: I looked into known ham radio digital modes and signal processing algorithms - they seems to be too slow as they are designed to use minimal power for reasonable detection rates with pretty much infinite latency. My task is the other way around - to use maximal reasonable power to achieve low latency.

  • 1
    $\begingroup$ What kind of data rate do you need? How many hours a day would you need the connection? Would the data be for business purposes? Rather than answering in the comments, please edit your question. $\endgroup$
    – rclocher3
    Jun 16 '20 at 17:19
  • 1
    $\begingroup$ hey Boppity Bop, again, as I've told you on your first and on your (deleted) second question: you will need give your TX waveform a more complex structure if you want it to be detected so far away, unless you have a couple of nuclear power plants to power your transmitter. Again, you need to really describe your application's requirements (data rate, acceptable latency, acceptable probability of missing a transmission, acceptable probability of seeing a transmission where there's none) to even start answering this. And to little surprise, the recommendations stay the same: $\endgroup$ Jun 16 '20 at 18:00
  • 1
    $\begingroup$ use a highly robust digital mode. It's not as easy as you think it is, but once you actually sit down and edit your question to include the specifications I mentioned in parentheses () above, it's answerable. Without that, we can't tell you anything. $\endgroup$ Jun 16 '20 at 18:02
  • 1
    $\begingroup$ SNR is the result of transmit power, which you need to select, and the propagation properties, which you are subject to. You can ask "given a TX power of x, what's my SNR going to be?", but you can't ask "what's my SNR and my TX power going to be", because one defines the other. $\endgroup$ Jun 16 '20 at 18:07
  • 2
    $\begingroup$ This is a verbatim copy of the comment I gave on your deleted question; it still 100% applies: All in all, I think you'd be doing me a big big favor if you described what you intended to do with the knowledge you gain by this. I promise I wouldn't judge you by it! We tend to just technically advise, while still trying to answer your questions, as far as at all possible $\endgroup$ Jun 16 '20 at 18:09

What SNR should I expect?

Wrong question! You need to ask yourself: What's the SNR I need to still achieve my desired performance.

"Desired performance" here is

  1. How much data do I need to transmit in a definite time (data rate)
  2. How long can I wait for the transmission to complete (latency)
  3. Which probability of missing the reception can I accept?
  4. Which probability of detecting a reception where there is none can I accept?

You need to answer all these four questions. Then, a theoretical lower bound for the necessary SNR can be given.

Also, it will then be possible to actually give practical "how to" advice.

What kind of power on TX side?

That depends on how much the channel attenuates your signal (we can find that out), times how much power needs to reach RX - which is defined by the necessary SNR above.

Can it be done on amateur licence?

That depends mainly on the TX power you'll need, which depends on the SNR you need.

Can I use polarization to transmit two sets -1,0 and +1,0?

Probably not on 30m, not realistically. You can use MIMO techniques if you have multiple far-distributed transmitters, though. Maybe, that is. If the channels are different enough. I'd call this "very advanced design topics, however".

Would it improve SNR?

If you could use independent polarizations, yes, you'd have one more degree of independent transmissions which you could use for diversity combining schemes. However, over that distance, I'd be very surprised if two orthogonal polarizations would see1 wildly different channels. So, no.

How difficult is to build it?

Depends on the techniques and power you use, which both depend on the SNR you need.

Start with asking a question where you describe your use case and answer the four performance requirements above.

  • $\begingroup$ Might want to include TX and RX bandwidth, Marcus. IMHO, the OQ is pretty broad. EDIT: see my comment here. $\endgroup$
    – Mike Waters
    Jun 16 '20 at 23:18

If you have a Ham license, and your application is for ham use, not encrypted, non commercial, and follows proper amateur radio etiquette, bandwidth, and power limits, you can. You will be also required to sit at the radio (no automatic operation for this band in the US IIRC).

As for transmitting FM, I'm not sure what you mean. There are plenty of digital modes you can base your application on. FT8 is the hottest right now. PSK31 may be worth investigating into.

Whatever you do, you're in for a LOT of error correction, like Reed-Solomon coding or other FEC algorithms.


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