Why isn't phase modulation commonly used in radio stations?

Question

I've tried searching about this and see answers saying that "frequency modulation can be considered as phase modulation" and I do understand that frequency and phase affect each other but why don't stations use direct phase modulation? As in the stations are actually called PM and phase is the variable directly being modulated. Probably there are ways that phase modulation is integrated but I'm kind of looking for a straight up answer that someone not into electronics who is curious why FM and AM stations are the only common radio stations.

Answer 1

Both are constant carrier angle modulation schemes with exponential side bands that must be attenuated to fit in the required BW to prevent adjacent channel interference in expected capture range of signals.

Variables include: the ratio FM modulation index which has a log proportional SNR improvement factor for \$\dfrac{\Delta f_c}{f_{in}}\$ for a carrier fc to noise ratio (CNR) above 10dB which requires a squelch below to block the abrupt change in noise. It also requires some pre-emphasis to maintain that SNR improvement somewhat like what vinyl music recordings had. This is because FM is the integral of PM so the deviation ratio improvement is needed more for high input frequencies but not simply differentiated over 3 decades but with plateaus.

Using PM for audio would challenge the dynamic range of the receiver demodulator as the carrier low frequency FM noise then would be approached by low frequency audio.

So the PM is better suited for data where the frequency span is more limited by design , so you would never see. A thousand zeros and ones followed by 010101., yet in audio the SNR of bass must be perceived as good as treble.

The metric for compressing BW in PM results in a higher ratio of Bits per Baud at the expense of threshold for SNR in PM for data is determine by the number of quantum phases out of 360 deg. For GOES uplinks, 8-PSK is now common and is QPSK. But when GOES 1 was first launched.

anecdotal

I recall in ‘78, GOES uplink was a low bit binary bit rate of 150 bps in a 750 Hz channel on a 420 MHz or so carrier with a 10W amplifier. The frequency error was very critical at power on during preamble and even tighter on the start of data. My mentor Bill Whitehead, designed the UHF Tx and he asked me to measure the phase modulation in degrees to very tight specs using a fancy HP “Time Interval counter” on the carrier to measure phase change on the UHF carrier. I had a HP9825 HP basic computer at the time, so it took me one afternoon to make a program to measure these and test more specifications and printout with the tiny strip printer. 90 degrees (?) at 420 MHz is a really small shift, yet this rack mount handled it with ease at the low bit rate. The internal clock was ultra-stable but used a quartz PLL to reduce the phase noise.

The benefit of PM here is controlled suppression of side-bands with an equal-weighting of SNR on input frequencies over a narrow 8 bit range.

Answer 2

There is actually little difference between FM and PM. The reason is that when you change the frequency you necessarily change the phase to get there. Likewise, when you change the phase you end up changing the frequency to get there.

Some FM schemes actually use PM and simply adjust the input signal to compensate for the slight difference so that the output ends up being exactly the same.

Why don't we use PM vs. FM? Well for the commercial FM broadcast band, FM was chosen long ago. If someone started broadcasting PM modulated music, it would sound funny on radios all over the place because they are expected FM, not PM.

Can PM be used for broadcast? Yes it can and there is no reason why a band could not be created that used it. But currently there is not such a band. PM is often used by Radio Amateurs (HAMs) along with just about all other forms of modulation. PM is used in your WiFi router, your mobile phone, and DBS satellite TV systems.

Answer 3

Actually, it [PM] kinda is used if pre-emphasis is used in FM broadcasting. Quoting from Leon Couch's textbook (Digital & Analog Communication Systems):

In angle-modulated systems, the signal-to-noise ratio at the output of the receiver can be improved if the level of the modulation (at the transmitter) is boosted at the top end of the (e.g., audio) spectrum—this is called preemphasis—and attenuated at high frequencies on the receiver output—called deemphasis. This gives an overall baseband frequency response that is flat, while improving the signal-to-noise ratio at the receiver output. (See Fig. 5–16.) In the preemphasis characteristic, the second corner frequency f2 occurs much above the baseband spectrum of the modulating signal (say, 25 kHz for audio modulation). In FM broadcasting, the time constant 1 is usually 75 µs, so that f1 occurs at 2.12 kHz. The resulting overall system frequency response obtained using preemphasis at the transmitter and deemphasis at the receiver is flat over the band of the modulating signal. In FM broadcasting, with 75-µs preemphasis, the signal that is transmitted is an FM signal for modulating frequencies up to 2.1 kHz, but a phase-modulated signal for audio frequencies above 2.1 kHz, because the preemphasis network acts as a differentiator for frequencies between f1 and f2. Hence, preemphasized FM is actually a combination of FM and PM and combines the advantages of both with respect to noise performance.

And such filters can a pretty simple/cheap affair, at least at textbook level:

I think an alternative viewpoint (if you want to see the system as PM) is that a non-linear compression scheme [FM] is used for the lower (audio) frequencies.

Slightly aside, but for this kind of [basic] Q, it's perhaps worth adding why pre-emphasis is done. These Stanford slides/diagrams (by John M Pauly) are kind-of self-explanatory once you consider the profile of the noise:

Anyhow, I think most countries FM broadcast standards mandate some use of pre-emphasis. Wikipedia says

The amount of pre-emphasis and de-emphasis used is defined by the time constant of a simple RC filter circuit. In most of the world a 50 μs time constant is used. In the Americas and South Korea, 75 μs is used.

A quick search on the net finds that even e.g. countries like Gambia seem to mandate it. (Interestingly, they use the US standard of 75 μs.)

More on that on the 'radio museum' site. There are e.g. some (theoretical) plots there on the [in]compatibility of US and European receivers--it seems to translate into a diff about about 4dB at the top end of the audio spectrum. One user there even posted a photo from inside a vintage set that had a switch for the two different pre-emphasis standards.

---

The ARRL Handbook also has something to say about this:

An FM transmitter with 6 dB/octave pre-emphasis of the modulating signal is indistinguishable from a PM transmitter. A PM transmitter with 6 dB/octave de-emphasis is indistinguishable from an FM transmitter. The reverse happens at the receiver. A frequency detector followed by a 6 dB/octave de-emphasis network acts like a phase detector. It is interesting to note that most VHF and UHF amateur “FM” transceivers should really be called “PM” transceivers due to the pre-emphasis and de- emphasis networks used in the transmitters and receivers respectively.

[...] Pre-emphasis of speech frequencies (a 6 dB per octave high-pass response from 300 to 3000 Hz) is commonly used to improve the signal-to-noise ratio at the receive end. Analysis shows that this is especially effective in FM systems when the corresponding de-emphasis (complementary low-pass response) is used at the receiver. (See reference for Schwartz.) By increasing the amplitude of the higher audio frequencies before transmission and then reducing them in the receiver, high-frequency audio noise from the demodulation process is also reduced, resulting in a “flat” audio response with lower hiss and high-frequency noise.

Note however that standards for those amateur radio transmitters/receivers differ from those for FM broadcast, discussed earlier.