I'm in the pencil-and-paper stage of building a radio receiver that has the following characteristics:

  • AM shortwave/HF
  • three fixed frequencies - 3.33, 7.85 and 14.67 MHz
  • single upper sideband with full carrier
  • BFSK modulation, varying between 2025 and 2225 Hz at 300 baud
  • The bands are allocated between 170 and 700kHz of BW, which I think includes guard band

Keeping the focus of this question narrow, to only the frontend - between the antenna and the preselectors -

If feasible, it would be nice to receive all three channels at once, though I understand that that would increase cost and complexity, and the bands might interfere with each other on the board. Failing that, it would be nice to have digital switching between the above frequencies, or mechanical if that's not at all possible.

If I have an antenna tuning varicap, perhaps I could have fixed capacitors across the first and second antenna selection circuits so that it's possible for all three stations to be tuned at the same time.

If I were to build simultaneous receive sections after one wideband antenna, each with their own preselector, what would a power splitter circuit look like? How would this affect impedance matching?

  • \$\begingroup\$ I'm not the most familiar with radio work, but wouldn't you need three antennas if you wanted to receive three stations simultaneously? \$\endgroup\$
    – Hearth
    Commented May 16, 2018 at 14:24
  • 3
    \$\begingroup\$ Quit the pencil and paper stage and use a simulator and ask a far simpler and more direct question than scatter gun the whole design. Nobody is going to produce a decent answer to all your questions and provide circuit diagrams and nobody is really going to provide answers given that you have pre-empted some techniques that might well be inappropriate. This requires too much work to answer. \$\endgroup\$
    – Andy aka
    Commented May 16, 2018 at 14:46
  • 1
    \$\begingroup\$ Many of the things you mention would be seperate questions with fairly long answers. You really need to ask direct questions rather than putting up a lot of stuff in one. Many folks will look at this and think "well, I don't have time (or knowledge) to answer all of this" but would (or could) have answered something shorter. \$\endgroup\$
    – JRE
    Commented May 16, 2018 at 14:57
  • 1
    \$\begingroup\$ You cannot design anything , and worse select a topology unless you state your design specs for signal level , CNR , known noise spectrum , BER of acceptable results and coherence rate of data collisions on each channel( 2 or more at one time). Do try harder to complete this task first before anything else. \$\endgroup\$ Commented May 16, 2018 at 15:00
  • 3
    \$\begingroup\$ Those are the carrier frequencies for CHU, the Canadian time code service. Been there, done that -- about 20 years ago! What exactly are you trying to accomplish? My experience was that interference, fading and variable path length were problems that no receiver could overcome. It might be workable if you're within ground-wave range of the transmitter. \$\endgroup\$
    – Dave Tweed
    Commented May 16, 2018 at 15:16

1 Answer 1


Try a vertical whip.

Have 3 high-Q PI matching networks; these frequencies are not harmonically related, thus high-Q networks will minimally interact even tho fed by the same (whip) antenna. Let the whip be quarterwave at 3.33 MHz, and craft the high-Q networks at 7.85 and 14.67 to handle long-wire vertical-wire antennas.

Feed the outputs of the 3 high-Q networks into your choice of LNAs.

  • \$\begingroup\$ Is there any way to do this without quarter wave? Wouldn't that be like 21 metres? \$\endgroup\$
    – Reinderien
    Commented May 17, 2018 at 9:08
  • \$\begingroup\$ Matching networks of adequate complexity can match anything to anything; the more "wire", the more energy extracted from the EM wave impinging on the "wire", up to point where skin-effect losses penalize the longer distances traveled by the energy heading to the 3 matching networks. \$\endgroup\$ Commented May 18, 2018 at 2:34
  • \$\begingroup\$ Ok. The matching networks - since their input would be in parallel, for an antenna impedance of 300ohm would they each need to be seen as 900ohm? \$\endgroup\$
    – Reinderien
    Commented May 18, 2018 at 7:34
  • \$\begingroup\$ No. The PI networks will be somewhat resonant, at each of your 3 primary frequencies. These PI networks WILL INTERACT via the antenna, so I'd be tempted to have a resistive splitter, or a 3-way resonant splitter (if such exists) feed each of the PI networks. The splitter may be a mess of reflected energy (that is how almost all "filters" work-----by reflecting the non-transmitted frequencies). \$\endgroup\$ Commented May 24, 2018 at 0:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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