Hi, all !

I'm rather curious about electronics and "how it works", and a friend
of mine asked me how his FM stereo player works. After explaining him
FM modulation, FM spectrum shape and content, I've told him piece of
theory for demodulation. But there's a little problem I'd really like
you explain me, because I'm not sure of my answer:
- for separating Left from Right, you need to recover the 19kHz tone,
that's ok (doubling to 38kHz...), after the FM demodulation.
- the 19kHz tone is important because it contain also phase
information, so you don't "build" your own 19kHz, but you extract it
from the source. So you need a PLL to lock this pilot.
Now, that's where the problem is...
- If you have tones higher than the 19 kHz tone, but rather close to
this 19kHz, does the PLL will lock on the desired pilot frequency, or
the one with the higher amplitude ?
Did I missed anything in the demodulation theory, or is there any
technical strategies for extracting this pilot tone? Band-pass
filtering round 19 kHz could be possible, with enougth selectivity, but
I can't imagine which order choosing !??

I hope you could give me advices and maybe technical trick !
Thanks in advance for your help,

Regards,
Mike

On 24 Apr 2006 02:44:28 -0700, "Chromatik"
wrote:

Hi, all !

I'm rather curious about electronics and "how it works", and a friend
of mine asked me how his FM stereo player works. After explaining him
FM modulation, FM spectrum shape and content, I've told him piece of
theory for demodulation. But there's a little problem I'd really like
you explain me, because I'm not sure of my answer:
- for separating Left from Right, you need to recover the 19kHz tone,
that's ok (doubling to 38kHz...), after the FM demodulation.
- the 19kHz tone is important because it contain also phase
information, so you don't "build" your own 19kHz, but you extract it
from the source. So you need a PLL to lock this pilot.
Now, that's where the problem is...
- If you have tones higher than the 19 kHz tone, but rather close to
this 19kHz, does the PLL will lock on the desired pilot frequency, or
the one with the higher amplitude ?
Did I missed anything in the demodulation theory, or is there any
technical strategies for extracting this pilot tone? Band-pass
filtering round 19 kHz could be possible, with enougth selectivity, but
I can't imagine which order choosing !??

I hope you could give me advices and maybe technical trick !
Thanks in advance for your help,

Regards,
Mike

Reliable locking to the 19kHz pilot tone is simple - you just make the
loop filter narrow enough to exclude any possible sidebands from the
mono signal from getting in. This is not hard, as the baseband audio
is limited to 16kHz with quite a sharp cutoff filter.

The loop filter is not a bandpass filter working at 19kHz, but a
lowpass filter operating at DC - which is what the phase detector puts
out when the loop is in lock. This is actually all technically trivial
with phase-locked loops. When FM stereo started, it wasn't done this
way, but by actually filtering the pilot tone with a bandpass filter
and doubling the frequency. You had to adjust coils to centre the
frequency and guarantee zero phase shift - the only way to get decent
stereo separation. It is so much easier now.

d

--
Pearce Consulting
http://www.pearce.uk.com

"Chromatik" wrote in message
ups.com
Hi, all !

I'm rather curious about electronics and "how it works",
and a friend of mine asked me how his FM stereo player
works. After explaining him FM modulation, FM spectrum
shape and content, I've told him piece of theory for
demodulation. But there's a little problem I'd really
like you explain me, because I'm not sure of my answer: -
for separating Left from Right, you need to recover the
19kHz tone, that's ok (doubling to 38kHz...), after the
FM demodulation. - the 19kHz tone is important because it
contain also phase information, so you don't "build" your
own 19kHz, but you extract it from the source. So you
need a PLL to lock this pilot.

As Don correctly points out, you don't need a PLL to lock onto the 19 KHz
pilot tone. It was initially done with narrow bandpass filters back in the
days when all we had was tubes or discrete transistors and building a PLL
involved more circuitry than few if any wanted to afford.

Now, that's where the problem is...
- If you have tones higher than the 19 kHz tone, but
rather close to this 19kHz, does the PLL will lock on the
desired pilot frequency, or the one with the higher
amplitude ?

Theproblem of interferring signals near 19 KHz is largely solved by low-pass
filtering the music as part of the encoding process. Therefore there was is
music anywhere near 19 KHz in the signal that is encoded. Remember that FM
stereo has a defined operational audio range that only goes up to 15 KHz.

FM stereo transmitters initially used fairly complex analog filters to
eliminate passing any signals much above 15 KHz. I presume this is done with
digital filters today.

Did I missed anything in the demodulation theory, or is
there any technical strategies for extracting this pilot
tone? Band-pass filtering round 19 kHz could be possible,
with enougth selectivity, but I can't imagine which order
choosing !??

In the origional analog implementation, the 19 KHz signal was extracted
using something as simple as a second-order tuned resonant circuit.

Here's the schematic of a fairly common implementatation of a stereo decoder
from the days of tubes:

http://oldtech.net/Fisher/500B/schemmpx.gif

Note that the left half triode of V100 is simply an amplifier fed by a high
pass filter composed of R200 and C200. The right half triode of V100 drives
Z100 which is a dual-tuned resonant transformer, no doubt tuned to 19 KHz.
Z101 is a resonant filter tuned no doubt tuned to 38 KHz.

Compare this to the circuitry wrapped around Q401-420 from the Heath AR15,
which was AFAIK one of the most complex FM decoders of all time.
http://www.mcmlv.org/Archive/HiFi/Heath_AR15.pdf . It was also finickey and
tended to be unstable. Some had most of their active and passive components
replaced by a single MC1310 chip, which performed better and was rock-solid.

Thanks a lot for your help, very appreciated !
My idea of the PLL's work wasn't so clear... now it's better, not
perfect !
Something new, now: the current used by the VCO of the PLL might suffer
from a lack of precision, that may cause mismatch for the 19kHz tone.
Is there a way to lock the current to the desired value, to eliminate
this mismatch ?

"Chromatik" wrote in message
ups.com

Thanks a lot for your help, very appreciated !
My idea of the PLL's work wasn't so clear... now it's
better, not perfect !

Something new, now: the current used by the VCO of the
PLL might suffer from a lack of precision, that may cause
mismatch for the 19kHz tone. Is there a way to lock the
current to the desired value, to eliminate this mismatch ?

The PLL is self-adjusting - its purpose in life is to lock into the nominal
19 KHz tone as long as it is within the lock-in range of the PLL.

"Chromatik" wrote in message
ups.com...
Thanks a lot for your help, very appreciated !
My idea of the PLL's work wasn't so clear... now it's better, not
perfect !
Something new, now: the current used by the VCO of the PLL might suffer
from a lack of precision, that may cause mismatch for the 19kHz tone.
Is there a way to lock the current to the desired value, to eliminate
this mismatch ?

The PLL is constantly comparing its own frequency with that of the incoming
signal. Even if the 19 kHz component in the incoming signal is very weak,
the PLL will stay perfectly locked on it. The PLL is designed so that it
won't shift phase or frequency too rapidly, and as a result it does a very
good job of staying in sync with the incoming signal even when the incoming
signal fades.

mc wrote:

"Chromatik" wrote in message
ups.com...
Thanks a lot for your help, very appreciated !
My idea of the PLL's work wasn't so clear... now it's better, not
perfect !
Something new, now: the current used by the VCO of the PLL might suffer
from a lack of precision, that may cause mismatch for the 19kHz tone.
Is there a way to lock the current to the desired value, to eliminate
this mismatch ?

The PLL is constantly comparing its own frequency with that of the incoming
signal. Even if the 19 kHz component in the incoming signal is very weak,
the PLL will stay perfectly locked on it. The PLL is designed so that it
won't shift phase or frequency too rapidly, and as a result it does a very
good job of staying in sync with the incoming signal even when the incoming
signal fades.


You mean: The PLL is constantly comparing the frequency AND PHASE of
the prescaler with that of the incoming signal. As long as the pilot
signal is fairly clean it will be locked and in phase with the original
38 KHz signal.


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

In article . com,
"Chromatik" wrote:

Thanks a lot for your help, very appreciated !
My idea of the PLL's work wasn't so clear... now it's better, not
perfect !
Something new, now: the current used by the VCO of the PLL might suffer
from a lack of precision, that may cause mismatch for the 19kHz tone.
Is there a way to lock the current to the desired value, to eliminate
this mismatch ?

There are two low pass filters on the feedback signal. The feedback is
actually a massive mess of noise but there's a good signal lurking down
near DC.

You can significantly clean up stereo on weak stations by lowering the
frequency of the two low pass filters. The downside is that you won't
get a phase lock unless you adjust your 19KHz pilot VERY carefully.
Some stereo decoders aren't stable enough for this modification.

The low pass circuit will look something like this:

+----|(---/\/\/\/--+
| |
---+----|(------------+---

Keep the ratio of capacitors the same or the circuit will ring rather
than settle down to a lock.


There may be another low pass filter controlling the mono/stereo switch.
Some radio are helped by slowing that too. Don't make it too slow. The
sound is very disorienting when the PLL drifts off its lock and the
decoder doesn't switch to mono.

In , on 04/25/06
at 08:09 AM, "Arny Krueger" said:



"Chromatik" wrote in message
oups.com

Thanks a lot for your help, very appreciated !
My idea of the PLL's work wasn't so clear... now it's
better, not perfect !

[ ... ]

Here's a link to the LM565, an early and popular PLL integrated
circuit: http://www.ee.washington.edu/stores/...near/lm565.pdf

As I recall, Signetics originated this part as the NE565 (in the early
1970's).

Old timers might find this article interesting:
www.complextoreal.com/chapters/pll.pdf

And the Wikipedia article is useful:
http://en.wikipedia.org/wiki/Phase-locked_loops

-----------------------------------------------------------
spam:
wordgame:123(abc):14 9 20 5 2 9 18 4 at 22 15 9 3 5 14 5 20 dot 3 15
13 (Barry Mann)
[sorry about the puzzle, spammers are ruining my mailbox]
-----------------------------------------------------------