Anonymous QST Author suggests that Part 97 requires 6 kHz AM Bandwidth Limit

[k4kyv]

On p. 64 of the May issue, the anonymous writer of the monthly Q-A "Workbench" column in QST entitled "The Doctor is In", responds to a report by the coordinator of the ARRL OO program regarding "wide AM signals with bandwidths of up to 30 kHz on 75 m."

He briefly explains the relationship between signal bandwidth and the frequency response of the audio that modulates SSB and AM transmitters, then states that Bell Labs concluded many years ago that high quality voice transmission (toll quality in telco jargon) can be carried over a 300-3300 Hz frequency response, but that "This worked better for the grey-haired Bell scientists who were likely to have lost some of their high frequency hearing".  He goes on to allege that an AM signal occupies "a bit more spectrum" than two SSB signals, based on the notion that there is no need to transmit audio frequency components of the voice that fall below 300 Hz, and that the "usual approach" with SSB is to transmit frequencies from 300 to 2700 Hz.

He then cites §97.307 of the FCC rules that states "No amateur station transmission shall occupy more bandwidth than necessary for the information rate and emission type being transmitted, in accordance with good amateur practice."  Since voice is what is being transmitted, he suggests that the rules require AM to occupy a bandwidth of "no more than about 6 kHz".

[WD8BIL]

He goes on to allege that an AM signal occupies "a bit more spectrum" than two SSB signals, based on the notion that there is no need to transmit audio frequency components of the voice that fall below 300 Hz,

".... there is no need..."
What "need" is there to tranmit 99% of what we transmit regardless of bandwidth ??

"..... below 300 Hz,...."
Someone should inform him that tranmitting down to 250 Hz isn't adding anything to the total bandwidth of the signal in typical "voice" communication.

His 30Khz statement is also dubious. How many of us have voice components up to 15Khz. Now if he's talking about unwanted artifacts, they'll come from a 300 to 2700 Hz signal just as quickly as any other.

Just another AM hating league faithful.

[KA1ZGC]

He goes on to allege that an AM signal occupies "a bit more spectrum" than two SSB signals, based on the notion that there is no need to transmit audio frequency components of the voice that fall below 300 Hz,

".... there is no need..."
What "need" is there to tranmit 99% of what we transmit regardless of bandwidth ??

Exactly. Someone should remind him that all you really "need" to send information from point "A" to point "B" is a telegraph key controlling a carrier. I wonder how he'd take to SSB being outlawed (or forced to use 250 Hz filters) because there's "no need" to be any wider than a CW station.

"..... below 300 Hz,...."
Someone should inform him that tranmitting down to 250 Hz isn't adding anything to the total bandwidth of the signal in typical "voice" communication.

Bingo. He's showing his ignorance by applying SSB metrics to AM signals.

An AM signal passing audio from 1 Hz to 3000 Hz is exactly the same bandwidth as an AM signal passing audio from 750 Hz to 3000 Hz. Either way, you've got a 6 kHz wide AM signal.

The same is not true of SSB, but too many hams (just like the current manufacturers) simply think of AM as two SSB signals and a carrier. A lack of education by the organizations who claim to be the bastion of ham radio knowledge is partially to blame. Since these very uneducated statements now have the endorsement of one of those organizations, it would be hard to conclude otherwise.

--Thom
Killer Agony One Zipper Got Caught

[k4kyv]

Plus, a lot of the intelligibility of a voice signal depends on frequencies above 3000~  When those highs are chopped off, the brain subconsciously fills in the missing data.  Listening to restricted audio for prolonged periods generates increased listener fatigue.

It doesn't matter whether the low end is cut off at 250~ or at 300~; the human voice has audio components that go down to about 80~.  Chopping everything  off at 250 or 300 Hz robs the voice of much of its natural quality and further reduces intelligibility.

But that article is just one person's interpretation of "good amateur practice".  During periods when the band is uncongested, there is no reason not to transmit an AM signal that is 30 kHz wide, as long as those sideband products are the result of the frequency range of the audio being used to modulate, and not spurious distortion products.

So what, if some dead-air group is forced to actually use their VFOs and move their operation 5 or 10 kHz from their "usual" frequency, to get away from the sidebands of an AM signal in an ongoing QSO, if there is 25 kHz of unused empty space on either side, which is not unusual during the day or after about 11PM on 75, particularly "down below".

[W3DBB]

.

[ka3zlr]

It is an assigned Mode along with all the other modes given for Privilege of use..

No actions need to be taken, this continual drag and redrag of pointless Width arrangements is only more proof of the unqualified first person lobbying that plagues our representative voice.

Leave it for the Future as it was found in the past...No one person Today has the Right to disturb that which generates Inspiration and Experimentation.

[KL7OF]

I would be anonymous too if I wrote that!!!

[KF1Z]

Time to go step on the cat's tail..... 

[Opcom]

not the poor cat!!      -this anonymous guy is probably friends with the one that called me "indolent" once. Never mind a silly goose like this anon-a-mouse fellow, his words have no meaning and we do not hear them. Yeah I called him a silly goose. Just trying to keep it clean.

[wd8das]

Editor, QST -
 
The May 2008, QST "Doctor is In" column opened with a question about so-called "30 kHz wide" AM signals. The answer provided in your column accepted these reports of wide AM signals at face value, but I suspect that those making such reports (Official Observers?) need to think again about how bandwidth is measured.
 
Occupied bandwidth is NOT measured by tuning a communications receiver back and forth across a signal to see "how far you can hear it". The differing selectivity characteristics of various receivers, their dynamic range capabilities, and the strength of the incoming signal will produce HUGE differences in the results, and the results will almost always be greater than the actual bandwidth of the signal. A spectrum analyzer, a special type of test instrument with a scanning receiver and standardized bandwidth and display characteristics, would be a much more appropriate device for bandwidth measurement.
 
But if you do accept the flawed premise that an amateur receiver would a good way to judge the bandwidth of a signal, you should perform the same experiment on SSB and CW signals to make an appropriate comparison. You'll find that the 2700 Hz-wide SSB signal described in the column will be audible for a much wider tuning range than just 2700 Hz - typically about 6 to 7 kHz depending on filtering in use in the receiver and signal strength. And a CW signal, often claimed to have almost no bandwidth, will be audible for several hundreds or thousands of Hertz depending on the filtering and signal strength. Try it yourself!
 
Does that mean that the CW signal is 2 kHz wide? Or that the SSB signal is 7 kHz wide? Of course not! But those are the results when you apply the same method as is being used to judge the supposedly "30 kHz wide" AM signals. Turn on a noise blanker or use a receiver that is getting overloaded and the results are even worse!
 
Please make this clear in the next QST and help stop this myth from being perpetuated.
 
Steve Johnston, WD8DAS
 
sbjohnston@aol.com
 

[k4kyv]

A thread has started on QRZ.com, and a similar one is going on the AM Reflector, regarding another article in that same issue of QST.  This article appears on page 71 and is also part of the QST Workbench section, titled "Getting on the Air: Selecting Your First HF Transceiver".

The author basically directs people to buy new equipment.  He implies that it is beyond the capability of today's ham to learn how to tune up a tube radio, then he goes on to state how "rare" finals tubes are, followed by some discussion of the "dangerous" high voltages involved.

Well... read for yourself some of the responses on those forums.

[AF9J]

Yeah, I replied to that thread.  Basically, I stated that the claims made by the article's author, were inaccurate and overblown.

73,
Ellen - AF9J
Plugging away at work, checking circuit boards

[W3SLK]

QST blows!!!

[W9GT]

Editor, QST -
 
The May 2008, QST "Doctor is In" column opened with a question about so-called "30 kHz wide" AM signals. The answer provided in your column accepted these reports of wide AM signals at face value, but I suspect that those making such reports (Official Observers?) need to think again about how bandwidth is measured.
 
Occupied bandwidth is NOT measured by tuning a communications receiver back and forth across a signal to see "how far you can hear it". The differing selectivity characteristics of various receivers, their dynamic range capabilities, and the strength of the incoming signal will produce HUGE differences in the results, and the results will almost always be greater than the actual bandwidth of the signal. A spectrum analyzer, a special type of test instrument with a scanning receiver and standardized bandwidth and display characteristics, would be a much more appropriate device for bandwidth measurement.
 
But if you do accept the flawed premise that an amateur receiver would a good way to judge the bandwidth of a signal, you should perform the same experiment on SSB and CW signals to make an appropriate comparison. You'll find that the 2700 Hz-wide SSB signal described in the column will be audible for a much wider tuning range than just 2700 Hz - typically about 6 to 7 kHz depending on filtering in use in the receiver and signal strength. And a CW signal, often claimed to have almost no bandwidth, will be audible for several hundreds or thousands of Hertz depending on the filtering and signal strength. Try it yourself!
 
Does that mean that the CW signal is 2 kHz wide? Or that the SSB signal is 7 kHz wide? Of course not! But those are the results when you apply the same method as is being used to judge the supposedly "30 kHz wide" AM signals. Turn on a noise blanker or use a receiver that is getting overloaded and the results are even worse!
 
Please make this clear in the next QST and help stop this myth from being perpetuated.
 
Steve Johnston, WD8DAS
 
sbjohnston@aol.com
 

Very well stated and absolutely accurate!  Bravo Steve!

This myth has also unfortunately even been perpetuated by some AMers who do not seem to understand what happens in a receiver that is overloaded by strong signals.  You can also be 60 dB down  5 KHZ away from a 40 over S-9 signal and still be very strong (S-5) in an adjacent channel situation.  Even the spectrum analyzer must be used carefully to avoid misinterpreting the results due to overloading.

73,  Jack, W9GT

[wd8das]

Response from QST:

- - - - - - - - - -

Steve,

You are quite correct, I didn't address the issue of how to actually
measure the bandwidth, but rather discussed the needed bandwidth for
the different modes and some ways in which wider signals can be
generated. I viewed that as the more informative topic, since we are
all bound by "good amateur practice."

The question of spectrum measurement is another interesting one, and
worthy of a note by the "Doctor" as well, in my view.

Interestingly, the June and July issues will have a two part series,
first a tutorial and then hands-on article, on spectrum analysis by
noted RF engineer John Stanley, K4ERO. He makes exactly your point,
although I'm not sure if it was mentioned in part 1 or 2 at this point.

The June Doctor column has one foot out the door, so I'll raise and
discuss your point in the July column.

BTW, with a 3 kHz (rectangular) receiving bandwidth and a 6 kHz AM
signal, the "observed" bandwidth will be 12 kHz, so that must not be
the full story of the "30 kHz bandwidth" signal. OTOH, who knows what
that reporter was using.

Personally, I use a Johnson Viking II for my AM work, mostly occasional
check-ins with the Sunday afternoon AWA net on 75. Yes, I've increased
the size of my bypass and coupling capacitors to improve my LF response
(from my D-104, often pictured in QST <g>), but I doubt if my gravelly
voice or my "old iron" gets much above 3 kHz. I also know I can't even
quite hit 100% modulation, so it probably isn't me <g>!

Thanks again for taking the time to write and for the good "Doctor"
suggestion!

GL & 73,    The "Doctor"

- - - - - - - - - -

[Tom WA3KLR]

Wouldn't a 6 kHz wide signal tuned on a receiver with a 3 kHz. brickwall filter yield a 9 kHz. "observed" bandwidth?

[k4kyv]

Exactly.

Response from QST:

- - - - - - - - - -
...BTW, with a 3 kHz (rectangular) receiving bandwidth and a 6 kHz AM
signal, the "observed" bandwidth will be 12 kHz, so that must not be
the full story of the "30 kHz bandwidth" signal. OTOH, who knows what
that reporter was using.

Is he overstating our position, or is this a typo?  With an ideally rectangular receiving bandwidth and a 6 kHz-wide AM signal with an ideally rectangular shaped transmitted bandpass characteristic, wouldn't the observed bandwidth be 9 kHz?  In an actual practical case, he is right; it would be closer to 12 kHz because of the combined effects of the sloping bandpass characteristic of any practical receiving filter, and the more gradual cutoff at the extreme edges of transmitted sidebands.

The "observed" bandwidth of a  signal is the sum of transmitted bandwidth and receiving bandwidth.

[w3jn]

And herein lies the the ultimate illustration of why any regulation by bandwidth is doomed to failure.   Even several very tech-savvy guys have different perspectives on what it is and how to measure it.  Think of Wally, Rich'rt, 'n' Fahrball trying to cipher this one 

[wd8das]

There are just too many variables when using communications receivers to try to judge bandwidth in a way that can be compared with others' results.  Just signal strength variations are enough - a strong signal will seem wider as it will be heard farther out on the slope of the filters.

But I was very glad to see "The Doctor" acknowledge this and be willing to discuss it in his column.  Might be helpful in the "myth busting" that it needed.

Steve WD8DAS

[W3SLK]

Unfortunately, the damage has been done. Those out there that are anti-AMer's will continue to point back to this 'official statement' by the (be)League(d). Regardless of how much back tracking is offered by
QS(heissen)T.

[ab3al]

30 khz hell my flex radio is set for 40khz... want to touch my monkey kaumawn

[WD8BIL]

No tellin' where that munky has bin !!

[k4kyv]

30 khz hell my flex radio is set for 40khz... want to touch my monkey kaumawn

If your audio  chain is distortion-free, unless maybe you use 10's of dB's of high frequency boost, that shouldn't matter.  Although frequencies above 3000~ contribute substantially to signal intelligibility, the power density of the sideband products representing modulating frequencies above 4 or 5 kHz is so low as to contribute negligibly to the interference caused by your signal.  If the band is relatively quiet and uncongested, those upper frequencies will be heard and appreciated on a good receiver with a wide enough passband.  Otherwise they will simply be lost in the noise floor of the band.

Of course if you are coming in 60 dB over S9, sideband products that are 60 dB down will still be coming in S9.  But try to convince the plastic appliance crowd of that, even if they can comprehend what you are talking about.

[Jim, W5JO]

Of course if you are coming in 60 dB over S9, sideband products that are 60 dB down will still be coming in S9.  But try to convince the plastic appliance crowd of that, even if they can comprehend what you are talking about.

Ha! Fat chance of convincing them or the League.

[Steve - WB3HUZ]

Doesn't matter what your filter types or bandwidth may be, if the AGC is on when making the "measurement", it will be incorrect.