A Tech Note About AM at the Band Edge

[radio8z]

Hello all,

I have been reading posts here and elsewhere where part 15 AM stations are operating in the high end of the broadcast band. For part 15, as is true for the amateur bands, the entire signal must be contained within the allowed band.

The upper band limit for AM is 1705 kHz. The AM transmitted bandwidth is twice the highest modulating audio frequency, and is centered around the carrier frequency. If the highest audio frequency is 15 kHz. for a 1700 kHz. transmitter, that extends the upper sideband to 1700 kHz. + 15 kHz. = 1715 kHz. which is out of band.

This is something to consider when operating at any frequency since your sidebands can interfere with a licensed broadcaster even though you are 20 or 30 kHz. away from their frequency. You have to stay out of their sidebands also.

My advice is to limit your transmitted audio to a maximum of 5 kHz. with software or hardware filtering.

Hope this helps.

Neil

 

[pronut]

There are licensed stations that are on 1700Khz. Don't they need to observe the 15 khz also? If so, woulden't that mean that they are illegal? The spacing of the AM band frequencys are at 10 khz, so, 1600, 1610, 1620 etc would be the correct legal spacing. So then, if your right that would make all the licensed stations illegal.





> Hello all,
>
> I have been reading posts here and elsewhere where part 15
> AM stations are operating in the high end of the broadcast
> band. For part 15, as is true for the amateur bands, the
> entire signal must be contained within the allowed band.
>
> The upper band limit for AM is 1705 kHz. The AM transmitted
> bandwidth is twice the highest modulating audio frequency,
> and is centered around the carrier frequency. If the
> highest audio frequency is 15 kHz. for a 1700 kHz.
> transmitter, that extends the upper sideband to 1700 kHz. +
> 15 kHz. = 1715 kHz. which is out of band.
>
> This is something to consider when operating at any
> frequency since your sidebands can interfere with a licensed
> broadcaster even though you are 20 or 30 kHz. away from
> their frequency. You have to stay out of their sidebands
> also.
>
> My advice is to limit your transmitted audio to a maximum of
> 5 kHz. with software or hardware filtering.
>
> Hope this helps.
>
> Neil
>

 

[radio8z]

> There are licensed stations that are on 1700Khz. Don't they
> need to observe the 15 khz also? If so, woulden't that mean
> that they are illegal? The spacing of the AM band frequencys
> are at 10 khz, so, 1600, 1610, 1620 etc would be the correct
> legal spacing. So then, if your right that would make all
> the licensed stations illegal.
>
My comment applies to part 15 operation and I am not qualified to speak to the issue of commercial AM rules.

The FCC accounts for adjacent stations and distance when granting broadcast licenses and spaces assigned frequencies accordingly for a given region. You will find that the frequencies are spaced greater than 10 kHz. apart.


Neil

 

[westlife]

> There are licensed stations that are on 1700Khz. Don't they
> need to observe the 15 khz also? If so, woulden't that mean
> that they are illegal? The spacing of the AM band frequencys
> are at 10 khz, so, 1600, 1610, 1620 etc would be the correct
> legal spacing. So then, if your right that would make all
> the licensed stations illegal.

Part 15 is bound by different standards than licensed broadcast stations. For example, Part 15 FM stations are forbidden to transmit on 87.9 MHz, yet there are several legally licensed stations on 87.9 FM. The same difference in band edges applies to AM. I believe the lower band edge for Part 15 AM is 510 kHz, yet there are no legally licensed broadcast stations on anything lower than 530 kHz in North America. Similarly on the upper end there are licensed stations on 1700 kHz which are allowed to use the same full 10 kHz NRSC audio bandwidth as licensed stations on any other AM frequency.
<P ID="signature">______________
Honorary Member of <a target="_blank" href=http://www.apostrophe.fsnet.co.uk/>The Apostrophe Protection Society</a></P>

 

[rfry]

> My advice is to limit your transmitted audio to a maximum
> of 5 kHz. with software or hardware filtering.
_________________

This also makes good sense because almost no commercial AM receiver reproduces modulating frequencies above 3 or 4 kHz in the first place. This is done to minimize the mutual sideband interference of stations on adjacent carrier frequencies -- which is a problem when skywave signals are present at night. Many AM broadcast stations limit their audio response to about 4.5 kHz for this reason.

The narrow RF bandwidth of most Part 15 AM tx antenna systems can reduce the relative radiated power of the higher modulating frequencies. But if the antenna system isn't tuned exactly right, it may reduce them more on one sideband than the other. The safest approach would be as Neil suggests above.
//

 

[meisrc]

Come on now, Anyone to suggest limiting audio bandwith to 5khz has to be smoking
something...We have several commercial stations here that have done that and it sounds totally horrible...

> > My advice is to limit your transmitted audio to a maximum
> > of 5 kHz. with software or hardware filtering.
> _________________
>
> This also makes good sense because almost no commercial AM
> receiver reproduces modulating frequencies above 3 or 4 kHz
> in the first place. This is done to minimize the mutual
> sideband interference of stations on adjacent carrier
> frequencies -- which is a problem when skywave signals are
> present at night. Many AM broadcast stations limit their
> audio response to about 4.5 kHz for this reason.
>
> The narrow RF bandwidth of most Part 15 AM tx antenna
> systems can reduce the relative radiated power of the higher
> modulating frequencies. But if the antenna system isn't
> tuned exactly right, it may reduce them more on one sideband
> than the other. The safest approach would be as Neil
> suggests above.
> //
>

 

[rfry]

> Come on now, Anyone to suggest limiting audio bandwith to
> 5khz has to be smoking something...We have several commercial
> stations here that have done that and it sounds totally horrible.
_____________

Are you sure that what you hear is related only to their audio bandwidth? A lot of stations use aggressive audio processing and level control to keep modulation near -100/+125% at all times, and to give them the "sound" that the program director wants. They might still sound bad even with more audio bandwidth.
//

 

[wkbam1690]

> Come on now, Anyone to suggest limiting audio bandwith to
> 5khz has to be smoking
> something...We have several commercial stations here that
> have done that and it sounds totally horrible...
>
>

Case in point: WCCO 830 Minneapolis...used to have great full 10 kHz NRSC audio, but implemented IBOC a ways back...now sound like SH*T. The sharp 5 kHz filters generate a harsh sound on all of my radios, which is very fatiguing to listen to. They used to sound great in wideband mode on my GE Superadio.

My part 15 operation uses CRL processing including NRSC filtering out to 10 kHz and sounds excellent in wideband mode.

 

[pianoplayer88key]

> Come on now, Anyone to suggest limiting audio bandwith to
> 5khz has to be smoking
> something...We have several commercial stations here that
> have done that and it sounds totally horrible...

I'd be ok with limiting audio bandwidth to 5kHz.
.
.
.
.
.
.
.
.
.
.
.
+/- 0.5dB, with a 1dB/octave rolloff to 22.05kHz, then a brick-wall lowpass filter.

 

[al_atl]

Here in Atlanta, WGST (Clear Channel) cut their bandwidth some time ago. They haven't launched IBAC yet as far as I know. Anyway, they lost me as a listener when they chopped off the audio. I find them unlistenable now. Fortunately, there is little worth listening to on WGST anyway. I now listen to an AM station about 50 miles away (WDUN, Gainesville) for the syndication that WGST has. Yeah there is a little more static and noise, but the audio is so much better. It's a night and day difference. It makes WGST sound like they have run-down, beat-up, poorly maintained equipment. Maybe they do...but it seems most of the Clear Channel stations are pretty well equipped. Switching between WGST and the other Atlanta area AMs is like switching from AM to FM. I think the idea of restricting bandwidth to 5khz is insane. Want to reduce interference? Why not reduce bandwidth to 3, maybe even 2khz? And while we're at it, let's lower modulation to 50%. And reduce power too. Change the rules so that maximum power for any AM is 1-kw. That should do a lot to reduce interference. Probably make room for plenty of LPAM stations, too! Of course, no one will be there to listen because we will have run them all off to satellite where the audio goes well beyond 5khz. As I suggested in a published letter...anyone who wants to "improve" AM by reducing the bandwidth, I have a request of you: Please, get out of radio *now*! Let those who still give a damn about quality try to salvage what little is left of AM.


> > Come on now, Anyone to suggest limiting audio bandwith to
> > 5khz has to be smoking something...We have several
> commercial
> > stations here that have done that and it sounds totally
> horrible.
> _____________
>
> Are you sure that what you hear is related only to their
> audio bandwidth? A lot of stations use aggressive audio
> processing and level control to keep modulation near
> -100/+125% at all times, and to give them the "sound" that
> the program director wants. They might still sound bad even
> with more audio bandwidth.
> //
>
<P ID="signature">______________
Consolidated Corporate Radio S U C K S !!!!!!</P>

 

[radio8z]

Question for Big Al

Hello,

You wrote "And while we're at it, let's lower modulation to 50%".

Please explain the difference in bandwidth between 50% and 100% modulation on AM.

I am looking forward to your reply.

Neil

 

[al_atl]

Re: Question for Big Al

There is a difference in bandwidth at 50% modulation vs 100% modulation? I'm not aware of this. Perhaps you are the one who should explain to me, I am always eager to learn. I suggested 50% modulation to reduce PEP. Combine reduced bandwith with lower modulation and lower carrier power and interference to adjacent channels and even co-channel stations will be greatly reduced. Of course, these suggestions are absurd...as is, in my opinion, the suggestion that AM stations roll back their bandwidth to 5khz. If we really want to reduce interference to adjacent frequencies, first of all IBAC has to be scrapped. And maybe we should re-examine the NRSC standard and consider going back to flat eq which would greatly reduce sideband splatter.


> Hello,
>
> You wrote "And while we're at it, let's lower modulation to
> 50%".
>
> Please explain the difference in bandwidth between 50% and
> 100% modulation on AM.
>
> I am looking forward to your reply.
>
> Neil
>
<P ID="signature">______________
Consolidated Corporate Radio S U C K S !!!!!!</P>

 

[meisrc]

Just read the responses...Damn AM is dieing lets dont go pulling the banana peel
from under its foot...

> > Come on now, Anyone to suggest limiting audio bandwith to
> > 5khz has to be smoking something...We have several
> commercial
> > stations here that have done that and it sounds totally
> horrible.
> _____________
>
> Are you sure that what you hear is related only to their
> audio bandwidth? A lot of stations use aggressive audio
> processing and level control to keep modulation near
> -100/+125% at all times, and to give them the "sound" that
> the program director wants. They might still sound bad even
> with more audio bandwidth.
> //
>

 

[rfry]

> I think the idea of restricting bandwidth to 5khz is insane.
> Want to reduce interference? Why not reduce bandwidth to 3,
> maybe even 2khz? And while we're at it, let's lower modulation
> to 50%. And reduce power too.
__________________

If all AM stations reduced power output and modulation depth by the same amount without changing audio bandwidth, interference ratios at the receiver would remain the same as they are now.

The 10 kHz channel spacing in the US dictates that an RF bandwidth of 10 kHz (~5 kHz of audio bandwidth) is the maximum that can be transmitted using conventional AM without creating interference to the sidebands of adjacent channels. Here is the reason why.

Receivers can limit their RF bandwidth to 10 kHz, yielding an audio response to 5 kHz. But that will not remove modulation from adjacent channels that exceeded 10 kHz of RF bandwidth on the adjacent channel. Those modulation components will lie in the RF passband of the receiver, and cause interference to the desired signal. The higher the audio frequency response on the adjacent channel, the more it will interfere with a receiver tuned one channel away.

For most locations, the only time that 10 or 15 kHz audio response would be usable for AM broadcast would be in the daytime (not that many of today's receivers could deal with it). At night, all stations with that audio response would have to cut it back to ~5 kHz if they wanted to avoid creating skywave interference to stations on adjacent channels. This is just a reality of physics.

Probably most people would choose listening to clear, intelligible program audio with less audio bandwidth than listening to "wideband" audio with a lot of interference in it.
//

 

[jdiamantis]

> > Come on now, Anyone to suggest limiting audio bandwith to
> > 5khz has to be smoking something...We have several
> commercial
> > stations here that have done that and it sounds totally
> horrible.
> _____________
>
> Are you sure that what you hear is related only to their
> audio bandwidth? A lot of stations use aggressive audio
> processing and level control to keep modulation near
> -100/+125% at all times, and to give them the "sound" that
> the program director wants. They might still sound bad even
> with more audio bandwidth.
> //
>
As an enginner for commercial stations, as well as a Part 15 experimenter, I've found that despite the narrow bandwidth of most receivers, pre-emphasis provided by NRSC-2 does increase perceived high frequency response. Also, because the 5kHz filters currently employed in commercial processors are close to the 3-4khz filters in radios, the combination of both create some interesting (read bad) artifacts. I have experimented with various filters, and have found a good compromise with limiting bandwidth to 7.5kHz. For some reason, even limiting badwidth to 6-6.5khz sounds much better than 5khz.

The difference between 10khz and 7.5khz is negligble on most receivers, and distortion due to bad antenna matching and transmitter non-linearities is reduced. It will also minimize, though not eliminate, adjacent channel interference. I have had success using steep multi-pole filters removed from decommissioned satellite receivers, but a 3 pole butterworth filter at 7.5khz should work just fine.

John Diamantis

 

[radio8z]

Re: Question for Big Al

> There is a difference in bandwidth at 50% modulation vs 100%
> modulation? I'm not aware of this. Perhaps you are the one
> who should explain to me, I am always eager to learn. I
> suggested 50% modulation to reduce PEP. Combine reduced
> bandwith with lower modulation and lower carrier power and
> interference to adjacent channels and even co-channel
> stations will be greatly reduced. Of course, these
> suggestions are absurd...as is, in my opinion, the
> suggestion that AM stations roll back their bandwidth to
> 5khz. If we really want to reduce interference to adjacent
> frequencies, first of all IBAC has to be scrapped. And
> maybe we should re-examine the NRSC standard and consider
> going back to flat eq which would greatly reduce sideband
> splatter.
>
>
> > Hello,
> >
> > You wrote "And while we're at it, let's lower modulation
> to
> > 50%".
> >
> > Please explain the difference in bandwidth between 50% and
>
> > 100% modulation on AM.
> >
> > I am looking forward to your reply.
> >
> > Neil
> >
>

Hi Big Al,

Of course it doesn't change the bandwidth. I just wanted to make sure you knew that.

You are correct about lowering PEP but listeners toward the fringe of a local AM station will hear a reduction in signal to noise ratio.

My original post referred to part15 AM stations since they are not frequency coordinated and have a greater chance of causing first or second adjacent channel interference than do licensed AM stations, at least for their neighbors. You have extended the discussion to commercial stations and that is fine, but my advice was for part 15'ers and not licensed stations.

Thanks for your comments,

Neil

 

[radio8z]

> As an enginner for commercial stations, as well as a Part 15
> experimenter, I've found that despite the narrow bandwidth
> of most receivers, pre-emphasis provided by NRSC-2 does
> increase perceived high frequency response. Also, because
> the 5kHz filters currently employed in commercial processors
> are close to the 3-4khz filters in radios, the combination
> of both create some interesting (read bad) artifacts. I have
> experimented with various filters, and have found a good
> compromise with limiting bandwidth to 7.5kHz. For some
> reason, even limiting badwidth to 6-6.5khz sounds much
> better than 5khz.
>
> The difference between 10khz and 7.5khz is negligble on most
> receivers, and distortion due to bad antenna matching and
> transmitter non-linearities is reduced. It will also
> minimize, though not eliminate, adjacent channel
> interference. I have had success using steep multi-pole
> filters removed from decommissioned satellite receivers, but
> a 3 pole butterworth filter at 7.5khz should work just fine.
>
>
> John Diamantis
>
Hello John,

Interesting comments. I have used multipole butterworth filters for other applications in the audio spectrum (doppler radar speed sensing). The analog filters work fine, but digital filters can produce aliasing unless the analog signal is filtered to remove the frequencies above the folding frequency before the digital filter or the clock for the digital filter is high enough in frequency to move the folding frequency well out of the audio range. Could this be the cause of the artifacts you mentioned? I don't see any other reason that using a 5 kHz. filter at the tx. and a 3-4 kHz. filter in the receiver would cause artifacts unless they were both digital.

As a side note, elliptical digital filters produce a sharp cut but allow frequencies above cutoff to pass.

Neil

 

[billybee]

Fact 1 - try this at home

Most AM radio receivers DO reproduce frequencies over 3-4 kHz. Easy to prove; just hook up an audio oscillator to your transmitter and you'll hear it on the receiver. Of course there is a rolloff, and the rolloff curve will vary widely from receiver to receiver. The NRSC standard helps compensate for the rolloff.

Fact 2 - try this at the hearing aid place

The frequency response of the human ear varies widely, perhaps even more so than the radios.

Fact 3 -

Phase shift colors the sound. Given that most ears will hear what's going on in the 5 kHz area, it's probably not a good idea to install filters with a 5kHz cutoff.

Fact 4 -

Adjacent AM channels are not granted in close proximity. Closely-located stations on frequencies 40 kHz apart (which is not uncommon) will not have problems with the 10kHz transmitted frequency.


Speculation 1 -

Many of the "5kHz" proponents have another motive... IBOC. In some cases, it really has nothing to do with whether the higher frequencies are necessary, but that the narrow analog bandwidth is needed to make AM IBOC work. Unfortunatley, it is IBOC on AM that will destroy the band, spewing digital hash, making it impossible to receive signals at night that happen to be in the way. The interim solution alone, which is daytime-only use of IBOC, will most certainly add nails to the AM coffin.

 

[rfry]

> Fact 4 -
>
> Adjacent AM channels are not granted in close proximity.
> Closely-located stations on frequencies 40 kHz apart (which
> is not uncommon) will not have problems with the 10kHz
> transmitted frequency.

Not true at night when skywave propagation is present. A good radio can receive stations every 10 kHz over most of the AM broadcast band at night. Skywave propagation of 10 kHz audio response through the AM transmitter can create large amounts of interference to the sidebands of stations on adjacent frequencies. And it can't be removed by using a narrower RF/IF bandwidth in the receiver, because the desired and interfering sidebands all occupy the same r-f spectrum.

> Speculation 1 -
>
> Many of the "5kHz" proponents have another motive... IBOC.
> In some cases, it really has nothing to do with whether the
> higher frequencies are necessary, but that the narrow analog
> bandwidth is needed to make AM IBOC work.

Even without nighttime IBOC, 10 kHz transmitted AM audio bandwidth would produce considerable nighttime interference to adjacent channels, as per my comment above.
//