I have a set of cheap computer speakers, which happens to use a TDA2822M. Its circuit is almost identical to the example circuit from TDA2822M's datasheet. It's USB powered and has no shielding.

I noticed a very faint audio coming from it, so I turned off the main breaker to reduce capacitively coupled mains hum (it's an apartment complex so this can't be fully eliminated) and plugged it to a portable USB battery. Then, walking around with the speakers, not touching the P2 connector to avoid more capacitively coupled noise, I could hear a faint, albeit very clear FM radio at regularly spaced spots, a few feet apart. I recognize the radio, it has the strongest signal in my town and operates at 105 MHz.

Outside, the radio couldn't be heard anymore. So my guess is a neighbor's radio feeding noise into mains and the speakers' wires acting like antennas.

What I can't understand is how could the speakers demodulate the FM signal so that it could be heard, since it obviously has no FM demodulation circuitry.

This answer mentions that "an FM signal will amplitude-modulate a tuned circuit", but it seems unlikely that the parasitic inductance and capacitance of the wires would resonate at that specific frequency. What am I getting wrong about that?

  • \$\begingroup\$ What sort of structure is you building made from? \$\endgroup\$ – Andy aka Jun 5 '20 at 14:30
  • 2
    \$\begingroup\$ Well, they're going to resonate at some frequency. If it's close enough to a radio station's frequency, it could well be picking up radio. \$\endgroup\$ – Hearth Jun 5 '20 at 14:30
  • \$\begingroup\$ @Andyaka It's built from CMUs - concrete masonry units, not reinforced concrete \$\endgroup\$ – comiy Jun 5 '20 at 14:43
  • 1
    \$\begingroup\$ Is "a few feet apart" roughly 5 feet? That would correspond to 1/2 wavelength at 105 MHz, the expected spacing for nodes in standing waves at that frequency. Are the building's floors roughly 10 feet apart? Then its steel structure could be fairly sharply resonant at the carrier frequency. \$\endgroup\$ – Dave Tweed Jun 5 '20 at 15:13
  • \$\begingroup\$ You've made a good start, noting that outside, the signal is weaker. Trying to determine where the signal is stronger is worthwhile - try more outside locations...you'd like to decide if your USB-power-source + amplifier/speaker is detecting the RF signal itself ,or picking up re-radiated signals from some other source. An interesting sleuthing problem. \$\endgroup\$ – glen_geek Jun 5 '20 at 15:28

Any amplifier will be amplifying tiny voltages or currents to something stronger, that's their job, if the frequencies of these fluctuations happen to end up in the frequency range the amplifier is sensitive to.

105 MHz is totally outside of the bandwidth of the TDA2822M (which really is a terrible amplifier IC and has been obsolete for the better part of 40 years, if you ever build an amplifier yourself, get something that's not from the 1970s).

However, intermodulation occurs on every nonlinear component. That's the principle that makes the crystal radio work, and the principle that allows your smart phone to communicate with wifi and 2G/3G/4G/5G base stations, no matter which of the bands they're using.

A diode is a nonlinear element. A transistor is one. The input stage of a totally terribly designed audio amplifier especially is one.

So, the FM carrier self-demodulates to baseband.

The strength of how much it does so depends on the sensitivity of your circuitry at picking up the frequency it has at any certain moment.

Your amplifier has something that inadvertedly acts as an antenna (ugly truth: everything is an antenna, and a lot of engineering time goes into making things bad antennas, so that they can go and also be something else, usefully).

That something seems to have a frequency selective behavior. For example, it is more susceptible to 104 MHz signals than to 106 MHz signals, with a nice falling slope of selectivity between these frequencies.

Tadah, a simple FM demodulator is born!

Therefore, this has nothing to do with your neighbor's electronics.

You just happen to have the cheapest thinkable speakers: No engineer will buy a TDA2822M these days, because they are simply worse in every thinkable aspect than other ICs, so the only reason was low price. That fits with the fact that they have been so cheaply designed that they pick up stuff at 105 MHz (literally in the band with the strongest interferers, where every self-respecting amplifier designer would have at least run a test to make sure this doesn't happen).

So, get a better amplifier. At this point, even the really cheap PAM8404 amplifiers will do better, and I've yet to see a single person mess up a PCB design so badly that they won't work.

The fact that it happens every couple of feet (sorry, you'll have to translate the freedom units to something metric, I don't know what that is; I think a feet is roughly 0.31 m?) is because indoors, you get such patterns in receive strength through constructively and destructively overlaying reflections, diffraction, refraction and through simple shadowing: Things like the steel bars in reinforced concrete, water pipes, electrical lines are very much spaced in the regions of wavelengths or below, so you get all kinds of nice wave phenomena that would make your physics teachers' eyes water.


Russian Radio Amateur Vladimir Polyakov (RA3AAE) has documented his reception of local FM radio stations on his portable crystal radio, intended for receiving local AM radio stations, without a tuned circuit.

'The very possibility of receiving VHF FM detector was discovered accidentally. One day I was walking in the Terletskiy park in Moscow, Novogireevo, I decided to listen to the broadcast - I had a simple crystal set without resonant tank (this circuit is described in the "Radio", 2001, N 1, Fig. 3). The receiver had a telescopic antenna with length of about 1.4 m. Wonder whether it is possible to receive radio broadcast with this short antenna? It was possible to hear, but weakly, simultaneous operation of two stations. But what is surprised me is the volume of receiving was rise and fall periodically almost to zero after every 5...7 m, and it was different for each radio station!

It is known that in the LW and MW bands, where the wavelengths are hundreds of meters, it is impossible. I had to stop at the point of receive with maximum volume of one of the stations and listen attentively. It turned out - this is "Radio Nostalgie", 100.5 MHz, broadcasting from the near city Balashikha. There were no line of sight between antennas. How does the FM transmission could be received by using the AM detector? Further calculations and experiments shows that it is quite possible and is not depends on the receiver.'

More information at: http://zpostbox.ru/fm_crystal_radios.html

  • \$\begingroup\$ haha, Vladimir is right, but if he discovered that in 2001, then he's at least 68 years late to the show: the first practical FM receivers where simply AM receivers with a bit of frequency selectivity up front. \$\endgroup\$ – Marcus Müller Jun 6 '20 at 10:51
  • \$\begingroup\$ Yes Marcus, slope detection is as old as the hills! But Vladimir received FM on a crystal radio with just a telescopic antenna and a couple of diodes and no tuned circuit. \$\endgroup\$ – vu2nan Jun 6 '20 at 11:02
  • \$\begingroup\$ a telescope antenna is something you tune to a specific wavelength \$\endgroup\$ – Marcus Müller Jun 6 '20 at 11:02
  • \$\begingroup\$ Marcus, I receive local FM stations on either side of the passband with my simple FM receiver. nandustips.blogspot.com/2019/03/my-pseudo-fm-crystal-radio.html Here's my version of Vladimir's receiver. The telescopic antenna is made to resonate on the local AM station frequency using a loading coil. nandustips.blogspot.com/2017/01/portable-am-crystal-radio.html \$\endgroup\$ – vu2nan Jun 6 '20 at 11:10
  • \$\begingroup\$ exactly! (nice clean circuit builds, by the way) You've got a frequency selective circuit in there (the gate capacitance of the FET, the adjustable capacitor, the inductor, and the parasitic capacitance and inductance of your cabling), and it's even designed to work at roughly audio bandwidth! \$\endgroup\$ – Marcus Müller Jun 6 '20 at 11:16

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.