A side discussion around an answer to the question How many stations could one hear with an AM/FM radio in front of the ISS' cupola window? has arisen with regard to the possibility that a change in frequency of order 2.6 kHz would need someone to turn the knob of an FM broadcast receiver carefully to track it, or whether a rudimentary PLL or other circuit might have been able to allow reception.

The radio in question (shown below) had only 9 transistors, and covered both the AM and FM bands.

So there are really two parts to this question:

  1. Could a 9 transistor AM/FM radio implement something that could be called a phase-locked loop?
  2. Could a 9 transistor AM/FM radio provide good reception if the broadcast drifted by 2.6 kHz due to a doppler shift? (The receiver would be in low earth orbit, presumably intercepting a lobe or side-lobe of the transmitter.)

enter image description here

  • 1
    \$\begingroup\$ 2.6kHz doppler won't affect FM tuning enough to be observable on such a basic radio ... but it'll be quite significant on the AM band. \$\endgroup\$
    – user16324
    Feb 20, 2018 at 12:52
  • \$\begingroup\$ What do the two questions have to do with each other? If there is no microcontroller (which is obvious for this radio) using a PLL would even make it more difficult to implement frequency tracking because frequency had to be controlled digitally. \$\endgroup\$
    – Curd
    Feb 20, 2018 at 15:36

2 Answers 2


Maximum doppler of 2.5 kHz is only about 3% Modulation. As others have said, the radio's 110 MHz. local oscillator would drift at least that much - it would likely be a mechanically rotated variable capacitor resonating with a coil inductor.
Such a radio may include Automatic Frequency Control (AFC). A switch is often included to allow AFC to be disabled. The station is manually tuned by ear with AFC off, then AFC is enabled to help the local oscillator to track any drift. Doppler shift (very slow) would be regarded by the AFC circuitry as drift.

AFC is a proportional servo loop whose input comes from the radio's FM detector discriminator. Audio is also derived from this discriminator. A low-pass filter ensures that audio is eliminated from the AFC servo voltage that is sent back to the local oscillator. A 9-transistor radio could have an AFC circuit, since the low-pass filter might include only passive components.
The local oscillator accepts this near-DC voltage to control its frequency over a narrow range - not enough to swing its frequency to an adjacent channel.

Even without AFC, doppler shift would very slightly increase audio distortion, probably not even noticed, since discriminator bandwidth is larger than the selective intermediate-frequency amplifier bandwidth.

  • \$\begingroup\$ In 2018 I find out I'd been using the AFC switch wrong all these decades! ;-) Thanks for the helpful post! \$\endgroup\$
    – uhoh
    Feb 20, 2018 at 15:29

Could a 9 transistor AM/FM radio have something that could be called a phase-locked loop?

No, this kind of radio is far too simple in construction to contain a PLL. For a PLL-based synthesizer in a radio receiver a frequency divider is needed which can be programmed to different division ratios so that we can receive many channels. Such a frequency divider requires in the order of a hundred transistors or much more. So in any radio receiver containing a PLL you would find at least one IC containing many transistors.

I do not think the doppler shift will be any issue as 2.6 KHz is much less than the reception frequency shift you can expect from such a simple (non PLL) receiver. My guess is that temperature changes will cause more frequency offset change (in the receiver itself) than that 2.6 kHz.

  • 4
    \$\begingroup\$ A nine-transistor radio could have an AFC circuit (auto-frequency-control). Not quite a frequency-locked-loop, but it does tend to track the carrier. \$\endgroup\$
    – glen_geek
    Feb 20, 2018 at 12:02
  • 2
    \$\begingroup\$ @EdwinvanMierlo I am afraid you do not understand FM modulation. What you say would be correct for an AM transmitter, but the spectrum occupancy of FM is way more complicated, being a set of sidebands of amplitude defined by the bessel function of the modulation index for simple tones and something much nastier for anything more interesting. The 200KHz channel spacing on FM predates stereo, and is a compromise even then.... \$\endgroup\$
    – Dan Mills
    Feb 20, 2018 at 12:05
  • 2
    \$\begingroup\$ A PLL does not necessarily require a frequency divider. Besides which, all we're really talking about here is automatic frequency control, a simple form of frequency locked loop. \$\endgroup\$
    – Dave Tweed
    Feb 20, 2018 at 14:23
  • 1
    \$\begingroup\$ @DaveTweed A PLL does not necessarily require a frequency divider. True, but I have never seen a design for an FM receiver with a PLL-based synthesizer without a frequency divider. The point of the frequency divider is that a reference frequency (for example from a crystal oscillator) is multiplied such that it can be the LO signal for the receiver. But I'll update my answer with "for a receiver". \$\endgroup\$ Feb 20, 2018 at 17:42
  • 2
    \$\begingroup\$ PLLs were used for FM demodulation long before they were used for frequency synthesis. There are at least two ways to do it: 1) make the loop bandwidth relatively wide, and take the audio output from the VCO control voltage, or 2) make the loop bandwidth relatively narrow and take the output directly from the phase detector. The latter only works with narrowband FM/PM (or digital FSK/PSK), in which the carrier phase can't slip by more than one cycle as a result of the modulation. \$\endgroup\$
    – Dave Tweed
    Feb 20, 2018 at 17:58

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.