Comparison of conventional AM to DSB + inserted carrier (Flex Radio)

[AB2EZ]

In the past few weeks I have listened to and/or participated in a number of discussions of the differences between conventional AM (plate modulation, screen modulation, ...) and the type of AM that is produced by a Flex Radio (double sideband + inserted carrier). Naturally, DSB + inserted carrier has been around since the first SSB transmitters appeared.

To allow for more than 3 attached .jpg slides, I will break this posting into two postings.

The purpose of these postings is to clarify the difference between conventional AM and DSB + inserted carrier in two contexts:

1. What does the modulated carrier look like (e.g. on a scope) if the modulation percentage exceeds 100% (i.e. not enough carrier in the case of DSB + injected carrier)

2. How does this affect the response of a traditional diode peak detector (or other type of envelope detector)

The attached .jpg slides show the modulation (chosen to be a triangle wave for ease of drawing) and the modulated carrier of a conventional AM signal, for 2 levels of modulation: 50% and 125% (positive).

Note a very important aspect of traditional AM, that is highlighted on the slides:

In traditional AM, represented mathematically... the amplitude of the modulated carrier is [1 + m(t)] only if [1 + m(t)] is greater than 0. If [1 + m(t)] is less than zero (corresponding mathematically to greater than a 100% negative peak) then the amplitude of the modulated carrier is zero. In these equations, m(t) is the waveform of the modulating audio signal.

[AB2EZ]

This post contains attachments that show what the modulated envelope of a DSB + inserted carrier signal looks like (e.g. if observed on an oscilloscope) for the cases of 50% modulation and 125% (positive peak) modulation.

Note that for the case of DSB + inserted carrier, the mathematical representation of the modulated carrier is [1 + m(t)] x the carrier, even if 1 + m(t) is less than 0

Also note, on the 2nd attachment, the green-highlighted envelope of the modulated carrier, for the case of 125% modulation. Note that a traditional diode peak detector (or any other kind of peak detector) would respond to the green waveform.

Typical power meters, and modulation monitors also respond to the green waveform.

For example, if you are using a Flex Radio, and you adjust the modulation level and the inserted carrier to produce 125% positive peaks, with a symmetrical modulation waveform... then the carrier level (as displayed by a modulation monitor) will appear to increase, when you apply the modulation.

If you adjust a traditional AM transmitter to produce 125% positive peaks, with a symmetrical modulation waveform... then the carrier level (as displayed by a modulation monitor)  will also appear to increase, when you apply the modulation... but not as much as it will in the case of DSB + inserted carrier.

[k4kyv]

The opposite is true for traditional AM. If you adjust a traditional AM transmitter to produce 125% positive peaks, with a symmetrical modulation waveform... then the carrier level (as displayed by a modulation monitor)  will appear to decrease, when you apply the modulation.

The carrier level meter in a traditional mod monitor is basically a rectifier type rf voltmeter that responds to the average amplitude of the signal, but not to the phase of the envelope.  So it would respond to the DSB signal exactly the same way as it would to a conventional AM signal with "ultramodulation" or negative peak limiting.  This would cause the carrier level to appear to increase with modulation, since the energy above the audio baseline (top edge of the unmodulated carrier envelope) is higher than that below.

OTOH, if the transmitter is modulated with an inherently asymmetrical waveform such as the human voice, phased so that the higher amplitude peaks go in the positive direction, and no rectifier circuitry such as the 3-diode negative peak limiter is in place, the peak amplitude may exceed that of "normal" modulation, but the carrier will not appear to increase.  The reason is that without some kind of rectifier circuitry, the average a.c. voltage of the modulating waveform is unchanged by the asymmetry.  The higher peaks are of proportionally shorter duration than the lower amplitude peaks on the other side of baseline, so that the area bounded between the waveform and baseline is the same on both sides.

[AB2EZ]

Don

Yes, I corrected my post... but I still think the effect (with symmetrical modulation) will be greater in the case of DSB + inserted carrier.

With traditional AM, the negative-going excursions of the envelope are cut off when the negative peak reaches 100%. This increases the average value of the envelope... as you pointed out.

In DSB + inserted carrier, the negative going excursions of the envelope are flipped over (adding to the average rather than subtracting from the average)... causing twice the effect v. just cutting them off.

Stu

[WD5JKO]

Stu,

   The two links below are from my Central electronics 20a at 130% sine wave modulation:

http://pages.prodigy.net/jcandela/Central%20Electronics%2020a%20QRO%20%232/130_precent.jpg
http://pages.prodigy.net/jcandela/Central%20Electronics%2020a%20QRO%20%232/130_percent_Trapezoid.jpg

   The balanced modulator in the 20a behaves just like the Flex does.

   I look forward to your analysis of what happens at the diode detector when we have DSB + carrier and the AM negative modulation swaps phase at the baseline. My experience with monitoring myself or others doing the same, is the diode detector has a very apparent "click-type" distortion that is very unpleasant to listen to. Then switch on the product detector, or better yet a SAM detector and the signal sounds fine.

Jim
WD5JKO

[N2DTS]

Did some experiments with dsb, and with dsb and very little carrier today with w1oxp, w1vtp and kb2rsk.
Only Garry was very strong into south jersey, so it was a good test.

Dsb was fine, but not as good as ssb, as there were phaze changes which sound like qsb on a signal, ssb sounded the same fidelity, less noise, no qsb effects, background static without the carrier to quiet things.
DSB with carrier and a sync detector was ok to about 200% modulation, above that, the inserted carrier was not enough to remove severe distortion. People did sound louder though.

The flex radios can do dsb, or you can insert any amount of carrier you want, tests show that a regular very good AM detector falls apart way before the sync one does at higher audio levels, but the sync detector would need to be changed to insert more carrier than normal.
What the flex likely does is insert a carrier at the same level as the average carrier level (to remove qsb effects) but was not designed to replace a carrier that is way down.

With modern radios (flex 5000's and 3000's) there was no problem being right on frequency, everyone sounded the same on all modes when switching between them, AM still sounds nicer because of the removal of background noise, and the ability to tell when someone unkeyed, fidelity seemed the same (freq response, distortion).
As the typical AM operators, it was a bit crazy when people just stopped talking, you cant tell without an 'over' that anything happened....over.

Brett

[k4kyv]

As the typical AM operators, it was a bit crazy when people just stopped talking, you cant tell without an 'over' that anything happened....over

I hate it when people just stop talking, unless it is a full-fledged fast break-in type of QSO.  This seems to be a recent phenomenon with AM, probably as a result of more SSB ops just discovering AM and running it on riceboxes.

The biggest problem I have with the sync detector is people in a roundtable not being on the same frequency.  It takes the detector a second or so to lock in on the new frequency, and often the first few words of a transmission are  lost.  This is particularly a problem when a new station breaks in and they have already given their callsign before the detector locks onto their signal.

[w1vtp]

I sort of agree with Brett - with a possible exception -- you can actually hear a sort of pseudo zero beat while the information of the two sidebands are canceling or adding the to AF power going to the speaker.  I looked at the phase display with Gary WA1OXT and I seem to always be 3 Hz different in either AM or DSB even though we both calibrate our timebases.  When I added in the usual 3 Hz, the phase rotation went away and that "QSB" sound went away also.  Another thing that showed up was that your really have to go in and do a custom timing on the AVC with the Flex or it is going to be really annoying.  I have a custom "hang" of 3 seconds set up for either HIFI SSB or in this DSB so the background isn't in your face all the time.

Don (KYV)  I agree with your assertion that there needs to be some sort of timebase sync when using regular AM detect or pilot carrier when using SAM detect for this to really work.

However, with HIFI SSB this phasing fading and adding is not a problem as there are no vector components from the two sidebands playing against with each other.  My conclusion -- if you are going to use DSB use a pilot carrier for SAM to lock onto or use common timebases for regular AM.  Otherwise, HIFI AM or HIFI SSB is the way to go.

Al

[w1vtp]

Brett -- how do you insert a carrier in DSB on a Flex? With a CE20a  You just have to offbalance the carrier

[N2DTS]

To adjust the carrier, you run AM mode, and in the TX menu, you select the amount of carrier you want.
You can turn it down to zero and have dsb, or put in anything from 0% to 100%.

Brett

[k4kyv]

Don (KYV)  I agree with your assertion that there needs to be some sort of timebase sync when using regular AM detect or pilot carrier when using SAM detect for this to really work.

However, with HIFI SSB this phasing fading and adding is not a problem as there are no vector components from the two sidebands playing against with each other.  My conclusion -- if you are going to use DSB use a pilot carrier for SAM to lock onto or use common timebases for regular AM.  Otherwise, HIFI AM or HIFI SSB is the way to go.

With HIFI SSB the purpose of the pilot carrier is to eliminate the frequency error.  The exact phase is not important. The listener may not know the nominal frequency of the suppressed carrier station even if both share a synched time base. Tuning strictly by ear can be difficult or easy depending on the nature of the signal and the aural acuity of the listener.

Last night some clown was playing what sounded like Nazi-era German drinking songs on slopbucket on 3875.  The audio was so hyelloweey that I had a hard time tuning it in well enough to figure out what it was.

[w1vtp]

To adjust the carrier, you run AM mode, and in the TX menu, you select the amount of carrier you want.
You can turn it down to zero and have dsb, or put in anything from 0% to 100%.

Brett

Ah!  I thought of that afterwards but thanks for confirming it.  I'd like to try that trick next time and see if SAM will lock up on the somewhat suppressed carrier.

[N2DTS]

It does, at least in the flex radios, but you cant go too far.
Other sync detectors might do better, or worse.

We did not really test the low limit, Eric did 1% carrier and that did not work any better than trying to copy dsb on the sync AM mode.

25% might work, I think that would be about 400% modulation.
100 is 100% modulation, 50% is 200% modulation, etc...

Brett

[N2DTS]

Don,
Heard you in the getto tonight, you were strapping and sounded very good...

Brett

[AB2EZ]

To illustrate what Don and Brett and Al are describing... regarding the demodulation of DSB vs SSB, if you don't have a small amount of transmitted carrier for your synchronous detector to lock on to (or if you don't have a synchronous detector):

Consider the special case, where the modulation is a pure sine wave (test tone) at 1 kHz.

Assume that you receive this by using a BFO (locally generated surrogate carrier) that is 0.5 Hz higher in frequency than a perfect surrogate carrier.

For LSB single sideband... the output of you receiver will be a tone at 1000.5 Hz. For USB single sideband, the output of your receiver will be a tone at 999.5 Hz. Either way, it will sound (to you, as the listener) like a tone at around 1000 Hz.

Now consider DSB. In that case, you will get the sum of two tones at the output of your receiver: 1000.5 Hz and 999.5 Hz. These will "beat" with each other (just as when you use a tuning fork to tune a musical instrument) . You will hear a tone that fades up and down at the beat frequency: 1000.5 Hz - 999.5 Hz = 1 Hz. On a scope, the output of the detector will look like a 1 kHz, 100% modulated sine wave... that is fading down to zero once per second.

When you are listening to speech, with DSB,  you would probably adjust your receiver to be several several Hz off of the ideal carrier frequency... to make this fading faster (and therefore less damaging to intelligibility*).... or you would adjust your i.f. filter to reduce the strength of one of the received sidebands.

*Sort of like two people singing ... but not exactly "on pitch"

Best regards
Stu

[w1vtp]

Don, Brett, Stu et al

It was a fun experiment nonetheless.  I like your explanation Stu - makes sense to me.  It was weird to hear it.  Guys, I'd like to try different settings of carrier insertion to see where things lock up in SAM.

Al

[AB2EZ]

The SB-10 is a beautiful piece of gear. I built one to go with my DX-100b when I was in high school. It is a great example of using the phasing method ...  a method that is now reborn in the Flex Radio equipment, and most of the software defined radios (except direct sampling at the incoming radio frequency).

My parents didn't have much money (living in rented apartments in the Bronx, no car, etc.) but, what they could afford (maybe more than they could afford), they gladly spent of things to feed my pre-engineering interests. I owe much to them for that.

Anyway, I have often thought of re-acquiring an SB-10; including a few unsuccessful bids on SB-10s listed on Ebay.

Just like comparing a 2N107 (which I have) to a 2Gigabyte USB drive (~2 transistors per bit x 8 bits per byte x 2 billion bytes ~ 32 billion transistors): inflation-adjusted price for the 2N107: $10.00; price for the USB drive: $10.00

...  It would be fun to show my students a vacuum tube-based phasing rig (inflation adjusted price ~ $750.00) and an SDR-1500. In particular, comparing the 90 degree phase shift network in the SB-10 (for producing the Hilbert transform of the audio signal) ... v. doing the Hilbert transform on any plain vanilla pc using digital signal processing and free software.

Stu

 

[WD5JKO]

Its still way more fun to run DSB with suppressed carrier using an old vacuum tube adaptor though.

Brian,

   I'm run a CE 20a and yes I have a blast with it with both AM and SSB. I'm curious though about your comment. It seems that the AM crowd would not like this since DSB SC is difficult to tune, and the SSB folks might say, hay OM, you have two sidebands. Add a little VFO drift and the SSB folks will throw you under the bus.

So who do you talk to when running DSB SC? Where and on what band do you run DSB SC?

Jim
WD5JKO

[N2DTS]

DSB-SC is not hard to tune at all, if you dont supress the carrier much.

Its just like AM with 150% modulation.

You can do that with various rigs, the modified FT102, the flex radios, maybe others....

Brett

[k4kyv]

To illustrate what Don and Brett and Al are describing... regarding the demodulation of DSB vs SSB, if you don't have a small amount of transmitted carrier for your synchronous detector to lock on to (or if you don't have a synchronous detector):

Consider the special case, where the modulation is a pure sine wave (test tone) at 1 kHz.

Assume that you receive this by using a BFO (locally generated surrogate carrier) that is 0.5 Hz higher in frequency than a perfect surrogate carrier.

For LSB single sideband... the output of you receiver will be a tone at 1000.5 Hz. For USB single sideband, the output of your receiver will be a tone at 999.5 Hz. Either way, it will sound (to you, as the listener) like a tone at around 1000 Hz.

Now consider DSB. In that case, you will get the sum of two tones at the output of your receiver: 1000.5 Hz and 999.5 Hz. These will "beat" with each other (just as when you use a tuning fork to tune a musical instrument) . You will hear a tone that fades up and down at the beat frequency: 1000.5 Hz - 999.5 Hz = 1 Hz. On a scope, the output of the detector will look like a 1 kHz, 100% modulated sine wave... that is fading down to zero once per second.

When you are listening to speech, with DSB,  you would probably adjust your receiver to be several several Hz off of the ideal carrier frequency... to make this fading faster (and therefore less damaging to intelligibility*).... or you would adjust your i.f. filter to reduce the strength of one of the received sidebands.

*Sort of like two people singing ... but not exactly "on pitch"

If the carrier is completely suppressed, it can still be reconstituted from the coherent nature of the two sidebands, using a circuit known as the Costas Loop. The greatest problem I see with DSBSC is that the local oscillator may  go out of lock during extended periods of "dead air", such as pauses in speech or silent intervals in a musical presentation, resulting in artefacts such as a clicking sound when audio resumes and the BFO re-locks. The pilot carrier method would seem easier and more practical, or better still (but more complicated) a doubly redundant system incorporating both methods.

The paper "Synchronous Communications," originally published in this journal in December 1956 by John P. Costas, has had a profound effect on modem digital communications. At the time of its publication, virtually all communications were analog in nature and, in fact, the stated goal of the paper was to demonstrate an amplitude-modulated (AM) system that could provide significantly improved performance for analog signalling and to dispel the notion that single sideband (SSB) had inherent performance advantages over properly constructed AM signaling systems. This paper is particularly significant because it was the first to demonstrate that carrier phase could be reliably recovered from the received signal using a structure that has come to be known as the Costas Loop.

It can be shown that present usage of amplitude modulation does not permit the inherent capabilities of the modulation process to be realized. In order to achieve the ultimate performance of which AM is capable synchronous or coherent detection techniques must be used at the receiver and carrier suppression must be employed at the transmitter. When a performance comparison is made between a synchronous AM system and a single-sideband system it is shown that many of the advantages normally attributed to single sideband no longer exist. SSB has no power advantage over the synchronous AM (DSB) system and SSB is shown to be more susceptible to jamming. The performance of the two systems with regard to multipath or selective fading conditions is also discussed. The DSB system shows a decided advantage over SSB with regard to system complexity, especially at the transmitter The bandwidth saving of SSB over DSB is considered and it is shown that factors other than signal bandwidth must be considered. The number of usable channels is not necessarily doubled by the use of SSB and in many practical situations no increase in the number of usable channels results from the use of SSB. The transmitting and receiving equipment which has been developed under Air Force sponsorship is discussed. The receiving system design involves a local oscillator phase-control system which derives carrier phase information from the sidebands alone and does not require the use of a pilot carrier or synchronizing tone. The avoidance of superheterodyne techniques in this receiver is explained and the versatility of such a receiving system with regard to the reception of many different types of signals is pointed out. System test results to date are presented and discussed.

http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?tp=&arnumber=1037570&isnumber=22249

What is synchronous detection and how does it work?  (a classic example of why the QST/QEX split was a terrible idea)

http://amfone.net/Amforum/index.php?action=printpage;topic=18379.0

[k4kyv]

That’s the reason for occasionally running it because the smart SSB operators will just zero beat your AM carrier and keep going. One of the reasons why speaking softly and turning up the wick doesn't always work. It just makes it easier for them to continue.

Now, what are they going to do when you don’t have a carrier for them to home-in on or zero beat

It's the dumb ones who do that.  The smart ones will QSY a few kc/s to a clear spot.

If you are running a carrier, you can always use the old "exit stage left" tactic.

[Pete, WA2CWA]

The SB-10 works really good, but you occasionally have to keep nulling out the carrier. As with any vacuum tube unit there is always a bit of drift involved. If you just let it go and don’t keep you eye on it the SB-10 will start to sound like an old Drake and you will hear a bit of carrier on it at the receiving end.

Its still way more fun to run DSB with suppressed carrier using an old vacuum tube adaptor though.

I've had a  SB-10 to mate with my Apache since 1960. By 1980 all the caps in the RF Phase Shift Network had to be replaced. Depending on the band, the phase shift was all over the place. Also, replaced the 12AT7's in the balanced modulator with 6201's. Produced better carrier null. Also replaced the standard carrier null pots with 10-turn units. Made it a lot easier to get a deep null.

[N2DTS]

Ten turn pots are the bomb!


Brett

[N2DTS]

Maybe build some book cases for W1UJR while you are at it...

That is the good thing about homebrew, you can build it just how you want it.

To much in one box can get hot though...

Brett

[Pete, WA2CWA]

Here's a great article from CQ Mag.
Upgrading your SB-10, May 1970, page 22 and a follow-up July 1970, page 10

And, if you're really ambitious:
Put your SB-10 on 160M
73 Mag, Feb. 1976, page 22 and a follow-up Feb. 1977, page 160