Building an FM transceiver

Question

As an electronics hobbyist, I've already built a thing or two so this didn't seem like a complicated thing to do, but I was terribly mistaken. I wanted to build an FM modulated radio transceiver controlled by an Arduino board that would work anywhere between 86 and 520 MHz so that it'd include normal FM radio, VHF and UHF amateur bands and PMR and CB channels.

I expected there to be a miracle IC that would just require an audio and carrier wave input, rf amp and antenna, or that there would be plenty of similar projects already done that I could bounce off of, but hours of research gave me no answers, just more questions.

I came here to ask why are the radios built always in specific bands like 136-148/200-260/400-430 MHz instead of working continuously - is there a legislative or physical limitation? And my second question is whether is there a way to approach this problem that would be friendly for someone who usually works with digital stuff (like an IC or module) instead of analog/radio electronics.

Thanks.

EDIT: Thank you all for your time, you were very helpful.

Answer 1

I expected there to be a miracle IC that would just require an audio and carrier wave input

Ah, but FM actively modulates a carrier; you can of course use the external oscillation input as carrier for a superhet design, but then you'll still need to generate the FM-modulated IF and suppress the leakage of the oscillator at the output (wíthout suppressing the frequency-varying intended carrier). That's a rather complex thing to do in a single IC.

I came here to ask why are the radios built always in specific bands like 136-148/200-260/400-430 MHz instead of working continuously - is there a legislative or physical limitation?

Yes :D

both, mostly!

Also: If you only offer specific bands as device manufacturer, you don't have to guarantee performance in between. So, since it's not a big market you'd target with anything that's not commercially legal to do (the couple thousand ham rigs you could sell at most ... pffft).

When you do a superhet FM transmitter, you produce RF at \$f_{\text{LO}} \pm f_{IF}\$ (and of course other harmonics/intermodulation products), where your message signal is actually a frequency-changing \$f_{IF}\$, but you only want the sum, not the difference (or vice versa); to isolate the sum frequency for a clean signal, you will need to filter everything below \$f_{LO}+f_{IF}\$. That only works with a fixed filter bank if you can't pick from more than an octave of \$f_{LO}\$.

And my second question is whether is there a way to approach this problem that would be friendly for someone who usually works with digital stuff (like an IC or module) instead of analog/radio electronics.

Sure; you could generate an IF signal with e.g. a microcontroller (FM modulation of a carrier between say 100 Hz and 75.1 kHz is not that mathematically hard to do); then, mix that up with about any LO (you can buy digitally controllable oscillators, Silabs has such) using about any mixer (SA612 is certainly a classic). Then, you get all the intermodulation products, and your filtering needs to select the one you actually want to transmit.

A very digital way to do that is the rpi_tx software, which uses the PWM units on a raspberry Pi SoC as generator for an RF signal; you'll have to add a solid amount of filtering to get rid of these harmonics you don't want (you only want exactly one of them).

Answer 2

Most of the answers are treating the transmit side. Basically however, the fundamental problem is that building such wideband tranceivers, fm, am or whatever is just very difficult.

I suspect that you are not going to find any transceivers that cover a continuous range up to 500 MHz. If you did, they would probably include FM along with everything else cuz once you have the basic circuitry in place for a transceiver, it is easy to add different modes.

On the legislative side, there is a sharp difference between broadcast FM and FM for voice communications. Basically, FM designed for voice is narrow band, it does not occupy a much greater bandwidth then a similar AM signal. It is a little bit better from a noise perspective than AM., but it is quite spectrally efficient.

Broadcast FM, on the other hand, occupies a much greater bandwidth because it uses a much higher deviation ratio. This provides much better noise suppression when you have a good strong signal. However a broadcast FM signal is somewhere around 100 or 120 kilohertz wide (I cannot remember exactly).

Answer 3

Legislative:
It is required to transmit "clean" signals that interfere with others very little. Frequency-selective filters are almost always required between radio signal generators and antenna in order to comply.

These filters at minimum must attenuate harmonics: at twice carrier frequency, three-times carrier frequency, four-times........

To span such a wide frequency range from 136 -to- 430 MHz, inductor and capacitor components in this filter would have to be tune-able, and would have to track one another.
This is not generally done...inductors and capacitors in these filters are fixed so that they pass the carrier frequency, and attenuate harmonics. These fixed-tuned filters only allow a small span of frequencies to be transmitted.

Answer 4

Wide ranging FM signal generators are expensive so for cost reasons Engineers use a limited range for VCO’s or VCXO’s or TCVCXO’s and design a 50 Ohm power amp using microwave principles with a PLL to synthesize lower bands as required.

Making a sine wave VCO with low THD over many decades is not simple so the economical way for narrow bands is to use RLC filters to suppress the harmonics. Phase noise and f stability are complex specs. So often a VCXO with a varactor diode or “varicap” or direct synthesis using a SoC (System on a chip) or other method is used to design within specs.

No design can start unless you have good understanding of all the design specs first in your design spec.