Prev: FDL / DASIP / S4D 2009 > Invitation to participate
Next: PLL and AFC
From: Jerry Avins on 22 Sep 2009 13:49
Tim Wescott wrote:
> On Tue, 22 Sep 2009 13:00:35 -0400, Jerry Avins wrote:
>
>> HardySpicer wrote:
>>> On Sep 21, 10:04 am, Jerry Avins <j...(a)ieee.org> wrote:
>>>> Randy Yates wrote:
>>>>> HardySpicer <gyansor...(a)gmail.com> writes:
>>>>>> On Sep 20, 8:01 pm, Jerry Avins <j...(a)ieee.org> wrote:
>>>>>>> HardySpicer wrote:
>>>>>>>> On Sep 21, 12:57 pm, Jerry Avins <j...(a)ieee.org> wrote:
>>>>>>>>> PalapaGuy wrote:
>>>>>>>>>> Single-sideband FM was a popular topic several years ago but I
>>>>>>>>>> haven't seen any articles on it since then. It looks attractive
>>>>>>>>>> as a way to transmit narrowband FM efficiently at VHF over short
>>>>>>>>>> distances. Does anyone know of current uses for this
>>>>>>>>>> technology?
>>>>>>>>> Narrow-band FM (NBFM) is the same as low-modulation-percentage AM
>>>>>>>>> with the carrier shifted 90 degrees. If the modulation index is
>>>>>>>>> low enough, what you mention is identical to the SSB-SC that is
>>>>>>>>> standard on ham and CB bands. I suppose people eventually
>>>>>>>>> realized that. Or is there some other meaning to what you write?
>>>>>>>>> Jerry
>>>>>>>>> --
>>>>>>>>> Engineering is the art of making what you want from things you
>>>>>>>>> can get.
>>>>>>>>>
> ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
>>>>>>>> Not sure it behaves the same. FM thresholds but AM doesn't. AM
>>>>>>>> (SSB) has problems demodulating it I suppose whereas with FM you
>>>>>>>> can easily use a PLL.
>>>>>>> NBFM doesn't threshold. You confuse the properties of FM with those
>>>>>>> of SSB NBFM. As far as I know,n SSB version of
>>>>>>> high-modulation-index FM doesn't exist.
>>>>>>> Jerry
>>>>>>> --
>>>>>>> Engineering is the art of making what you want from things you can
>>>>>>> get.
>>>>>>>
> ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
>>>>>> NBFM must threshold. In fact it thresholds worse than wide-band
>>>>>> since the beta is lower. That anoying sound on every police radio in
>>>>>> the UK (as it was of course).
>>>>>> Hardy
>>>>> Mischa Schwartz et al. discuss the FM threshold effect in quite a bit
>>>>> of analytical detail (a quad integration is performed at one point on
>>>>> four-dimensional probability distribution!) in section 3.6, p134.
>>>>> --Randy
>>>>> @book{schwartzcommtechniques,
>>>>> title = "Communication Systems and Techniques", author = "{Mischa
>>>>> Schwartz and William R. Bennett and Seymour Stein}", publisher =
>>>>> "IEEE Press (reissue, originally New York: McGraw-Hill 1966)", year
>>>>> = "1996"}
>>>> My Schwartz, "Information Transmission, Modulation and Noise", 1st
>>>> Edition, 1959; $10.34, says in italics* on p.304, section 6.3,
>>>> /"Narrowband FM thus provides no S/N improvement over AM."/
>>>>
>>> Yes but Fm thresholds. SSB loses lock of course.
>> Don't you get it? Thresholding happens when excess bandwidth is used to
>> improve SNR. When the signal is too low, the excess bandwidth allows
>> more noise into the receiver than the bandwidth-for-SNR trade rejects:
>> that's the threshold. With NBFM, there is no excess bandwidth, hence no
>> SNR improvement, hence no threshold. Unless, of course, your gain
>> control is calibrated up to 11.
>>
>> Jerry
>
> If you limit the IF signal in NBFM then you do see reduced noise at high
> signal levels (because you're clipping out the portion of the noise
> that's in phase with the carrier) and enhanced noise at low signal levels
> (because you're amplifying the snot out of the noise).
>
> So it does do something akin to thresholding -- but it's certainly much
> softer than the weak signal behavior of broadcast FM.

Distortion results if you clip NBFM in the IF. Take low-modulation-%age
AM, strip out the carrier, and reintroduce it with quadrature shift.
Voila! NBFM!* Clipping either one distorts. You may have
sort-of-narrowband FM in mind.

Jerry
_____________________________
* One of Armstrong's early FM stations had this kind of modulator. The
FCC requires crystal control all the way, and the options for
crystal-controlled FM modulators are limited. Armstrong used frequency
multipliers to increase both the carrier frequency and deviation, then
heterodyned that signal to a low carrier frequency, keeping the higher
deviation intact. A few passes of that (multiply up, heterodyne down)
and he got the standard (50 MHz band) deviation he needed. Clever guy!
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
From: Jerry Avins on 22 Sep 2009 14:00
Laloo wrote:

...

> The excess BW itself is not why threshold happens but it does
> exacerbate it. The threshold effect is explained in quite a bit of
> detail in books by Taub and Schilling and Misha Shwartz et al. as it
> was orignally studied by Rice. To understand it fully, one needs to
> look at the phasor diagram of the signal and noise. When noise is
> small (high SNR), it only causes the recieved signal phasor to have
> small pertubations around the signal phasor and hence only small
> variations in the received phase. However, as the noise becomes large
> (below thresold), it can result in the recieved phasor making full
> encirclements around the origin and each of those encirclements in a
> 2pi phase shift in the received phase (cycle slip in PM or clicks in
> FM). The noise contribution from these clicks results in the rapid
> decrease in the SNR for input CNRs below threshold.

Agreed. I soft-pedaled the details. As to whether the excess bandwidth
is the cause of quieting, I take no stand. It certainly makes quieting
possible, ans allows more noise into the receiver while doing that.

> The threshold effect is a consequence of these clicks. It is not true
> that clicks happen only for large modulation indices but it is true
> that the BW is larger for larger modulation indices and thus more
> noise enters the receiver increasing the rate of clicks. That is why
> the threshold effect occurs at lower CNR for higher modulation indices
> but it occurs for all modulation indices including small ones.
>
> Anyone interested in studying the mechanism of threshold effect and
> the accompanying analysis should check out books by Taub and Schilling
> and Misha Shwartz et al.
>
> As far as the statement "Narrowband FM thus provides no S/N
> improvement over AM", the gain FM provides over AM is a function of
> modulation index as one trades off BW, so if the modulation index is
> low, there will be no S/N improvement but it depends on the modulation
> index.

"Narrow" is merely qualitative. When the modulation index is high enough
to permit some noise reduction, "NBFM" doesn't accurately describe the
signal.

Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������
From: HardySpicer on 22 Sep 2009 21:41
On Sep 23, 5:23 am, Tim Wescott <t...(a)seemywebsite.com> wrote:
> On Mon, 21 Sep 2009 23:05:25 -0700, HardySpicer wrote:
> > On Sep 21, 10:04 am, Jerry Avins <j...(a)ieee.org> wrote:
> >> Randy Yates wrote:
> >> > HardySpicer <gyansor...(a)gmail.com> writes:
>
> >> >> On Sep 20, 8:01 pm, Jerry Avins <j...(a)ieee.org> wrote:
> >> >>> HardySpicer wrote:
> >> >>>> On Sep 21, 12:57 pm, Jerry Avins <j...(a)ieee.org> wrote:
> >> >>>>> PalapaGuy wrote:
> >> >>>>>> Single-sideband FM was a popular topic several years ago but I
> >> >>>>>> haven't seen any articles on it since then.  It looks attractive
> >> >>>>>> as a way to transmit narrowband FM efficiently at VHF over short
> >> >>>>>> distances.  Does anyone know of current uses for this
> >> >>>>>> technology?
> >> >>>>> Narrow-band FM (NBFM) is the same as low-modulation-percentage AM
> >> >>>>> with the carrier shifted 90 degrees. If the modulation index is
> >> >>>>> low enough, what you mention is identical to the SSB-SC that is
> >> >>>>> standard on ham and CB bands. I suppose people eventually
> >> >>>>> realized that. Or is there some other meaning to what you write?
> >> >>>>> Jerry
> >> >>>>> --
> >> >>>>> Engineering is the art of making what you want from things you
> >> >>>>> can get.
>
> ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯>> >>>> Not sure it behaves the same. FM thresholds but AM doesn't. AM
> >> >>>> (SSB) has problems demodulating it I suppose whereas with FM you
> >> >>>> can easily use a PLL.
> >> >>> NBFM doesn't threshold. You confuse the properties of FM with those
> >> >>> of SSB NBFM. As far as I know,n SSB version of
> >> >>> high-modulation-index FM doesn't exist.
>
> >> >>> Jerry
> >> >>> --
> >> >>> Engineering is the art of making what you want from things you can
> >> >>> get.
>
> ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
>
>
>
> >> >> NBFM must threshold. In fact it thresholds worse than wide-band
> >> >> since the beta is lower. That anoying sound on every police radio in
> >> >> the UK (as it was of course).
>
> >> >> Hardy
>
> >> > Mischa Schwartz et al. discuss the FM threshold effect in quite a bit
> >> > of analytical detail (a quad integration is performed at one point on
> >> > four-dimensional probability distribution!) in section 3.6, p134.
>
> >> > --Randy
>
> >> > @book{schwartzcommtechniques,
> >> >   title = "Communication Systems and Techniques", author = "{Mischa
> >> >   Schwartz and William R. Bennett and Seymour Stein}", publisher =
> >> >   "IEEE Press (reissue, originally New York: McGraw-Hill 1966)",
> >> >   year = "1996"}
>
> >> My Schwartz, "Information Transmission, Modulation and Noise", 1st
> >> Edition, 1959; $10.34, says in italics* on p.304, section 6.3,
> >> /"Narrowband FM thus provides no S/N improvement over AM."/
>
> > Yes but Fm thresholds. SSB loses lock of course.
>
> As Jerry has been saying, NBFM doesn't threshold, not really.  For that
> matter, SSB has no lock to lose.
>
> So you're wrong on two counts.
>
> You _do_ understand the difference between broadcast FM and narrow-band
> FM, yes?
>
> --www.wescottdesign.com

I have heard it thresholding many times on police radios. When there
is no signal at all you get thresholding.
Not so with AM. (though SSB is a little different).

Hardy
From: Jerry Avins on 22 Sep 2009 21:47
HardySpicer wrote:

...

> I have heard it thresholding many times on police radios. When there
> is no signal at all you get thresholding.

What kind of threshold exists in the absence of a signal?

...

Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������
First  |  Prev  | 
Pages: 1 2 3 4
Prev: FDL / DASIP / S4D 2009 > Invitation to participate
Next: PLL and AFC