Compatable SSB

[W1EUJ]

How different does compatable SSB (SSB with some carrier) sound from true AM?

[The Slab Bacon]

Compatable AM is a carrier and one sideband. It sounds like crap, but can be detected with a standard diode detector and takes up a little less bandwidth than standard DSB AM.

                                                               the Slab Bacon

[W1EUJ]

Is the poor audio quality inherent in the scheme, or is it because of the way its done (say, adjusting/pulling the balance off in the balanced modulator)?

[KA1ZGC]

Listen to WBCQ on 9330 and judge for yourself. That's single-sideband reduced carrier. Not sure about 5105, but that might be as well. 7415 is full AM, and I beleive 17495 is also.

So-called "compatible SSB" is just that: single-sideband, reduced-carrier. Plate detectors don't always handle it well. It'll sound like a constant selective-fade because it's particularly sidebandy.

--Thom
Keep Away One Zorched Ground Conductor

[k4kyv]

The distortion results from inherent properties of the modulation process.

Actually, "compatible" SSB is not the same  thing as SSB with carrier.  It was tried on the AM broadcast band sometime in the 60's with mixed results.  It carried the full bandwidth of a regular AM signal, but everything was pushed to one side of the carrier, with an extra set of sidebands on the same side as the fundamental.  As I recall, 100% modulation with a sine wave reduced the amplitude of the carrier, produced a 1st order sideband of greater amplitude than the carrier, and a 2nd order sideband of the same amplitude as the carrier.

Ideally, the net result was a distortion-free tone when demodulated by a diode detector.  What you actually had was the 1st order sideband functioning as the carrier, while the carrier and 2nd order sideband served as sidebands, so that it looked like a conventional AM signal offset by the frequency of the tone. 

With complex modulation such as music and voice, the carrier and sidebands were all muddled up, but supposedly distortion-free on an envelope detector.  The signal was a complex mixture of phase and amplitude modulation.  Its purpose was to alleviate adjacent channel interference.  IIRC, WBZ and KDKA tried using it for a while to clear up mutual interference they were having with their skywave coverage.  I seem to recall that it was eventually abandoned because it didn't hold up well under selective fading skywave conditions.

Proceedings of the IRE Oct. 1961

[WA1GFZ]

1 BA AM

[Bacon, WA3WDR]

With 100% modulated SSB+C, there will be envelope distortion that will produce significant audio distortion when a linear detector is used. A square-law detector at the receiver will give linear throughput and clean audio with 100% modulated SSB+C, as long as there is not phase distortion in the signal path.

You know how an AM envelope looks when a good transmitter is modulated with a sine wave audio tone.  100% modulated SSB with carrier looks like a 'two-tone' test pattern for a linear amplifier - the positive peaks are seriously rounded down, and at 100% modulation the negative peaks are chisel-sharp.

[k4kyv]

With 100% modulated SSB+C, there will be envelope distortion that will produce significant audio distortion when a linear detector is used. A square-law detector at the receiver will give linear throughput and clean audio with 100% modulated SSB+C, as long as there is not phase distortion in the signal path.

I have often wondered how well an old buzzard 1931 vintage Philco would demodulate SSB+C.  Those receivers used a triode detector biased to the square-law mode.  That's why they have that peculiar distortion on a normal AM signal.  The early 30's BC receivers took a trade-off; they tolerated the inherent distortion of the square-law detector for the increased sensitivity.

[Bacon, WA3WDR]

This has been bothering me for some time.  I know the square law removes the distortion from 100% modulated SSB+C, but it seemed to me that there would be a DC shift with modulation.  So I put a spreadsheet together, two sine waves separated by a difference frequency, and then added together and squared.  I set it up to vary the level of one of the sine waves, and I tried a few levels against the fixed 'carrier', to simulate SSB+C with a varying amplitude sine-wave modulating tone.

Using a square-law detector on SSB+C, the demodulated audio is a sine wave at all levels, but at 100% modulation there is a 2:1 upward DC shift.  (AGC needs to be derived from a separate linear detector, or it will thrash around with modulation.)  The demodulated audio level is directly proportional to the SSB level, but the carrier shift makes the modulation percentage of lower level audio higher than it would otherwise be, because the carrier is lower at lower audio levels.  After squaring, with the SSB equal to the carrier (both unity level), the carrier appeared to be 6dB stronger than unmodulated carrier level, and average modulation was 100%.  Reducing the SSB 6 dB only cut the modulation to about 80%.  Reducing the SSB another 6 dB (to -12 dB) cut the modulation to about 53%.  Reducing the SSB another 6 dB (to -18 dB) cut the modulation to about 25%.  If there is AGC and it follows the carrier, it will compress the audio by about 6 dB.

Without AGC, the carrier shift will cause some degree of thumping on sudden transients.  With AGC, there would still be thumping, and the audio would be compressed, sort of soggy 6 dB peak-limiting.  But on steady-state modulating signals such as sine waves, there would be no distortion.

With decent transmit audio processing, I find SSB+C to be listenable on a typical receiver (using a semi-linear detector), but it's definitely not the best AM.  With a square law detector it would sound better, but there would be some thumping from that DC shift.  Probably most things would sound ok, but it just wouldn't be as good as real AM.  And when SSB+C falls into a selective fade, the distortion is decidedly SSBish, as opposed to the DSBish distortion of real AM.  This difference is secondary - but to me, it makes selective fades even more unpleasant.

[Rob K2CU]

I have followed this thread with interest. The inherent second harmonic distortion from the popular square law detector led me to experiment with the precision AM detector and incorporation of same into my R4-C. There was something alluring in the idea of SSB+ C. Mostly it was the fact that instead of 375W plus 93.5 W in each of two sidebands for DSB-AM, you could do 375 W of carrier and 375 W in one sideband for SSB + C. And the nearly regular intrusion of SSB signals on one side of an AM QSO I am listening to, has found me playing with tuning to one side and effectively making the signal SSB + C. Sure, if you tune too far, the carrier slide down the slope on one side and you get into trouble. CHU at 7335 KHZ is USB +C. If you tune it in, it sounds quite natural. It does require a bit more tuning accuracy than DSB-AM.

I decide to run some simple tests this morning to compare the harmonic distortion of several of the receivers here. I used Spectrum Lab with my PC to examine the detected audio output for harmonics. The DSB-AM signal was generated by an HP-8640 set to 90% AM. I didn't want it's potential problems at 100% to enter the picture. IF there is a clammering for it, I rerun the tests.  I set the RF level to 50 uV (S9) at 3876 KHz. The receivers were tuned to center the carrier in the passband. I repeated the test using just the carrier of the 8640 plus a crystal oscillator at 3875 KHz, both at 50 uV. The receiver was tuned such that both signals were well within the passband, which was the AM passband of the receiver being tested.

The receivers tested were Collins 51J-4, TR-7, FT-897, and the R4-C with the precision detector. The others were with their respective stock AM detectors.

The measurements of second harmonic distortion (2 KHz) compared to the desired 1 KHz signal for the two types of signals
are as follows:

Receiver      Two tone        DSB-AM
51J-4           -21 dB            -17 dB
TR-7            -23 dB            -19 dB
FT-897         -22 dB            -25 dB
R4-C            -25 dB            -56 dB

ON air testing was done in late morning when there were strong QSO's on 3875 KHz and 3885 KHz. Full bodied signal with great vocals from  'HLR and 'INR  were coming into Hebron, CT at 20 over S9 or more. With each receiver other than the R4-C, tuning to either side so as to suppress one sideband resulted in a clearer sounding signal, especially with regard to the rich bass components in their speech. They was richer bass. OF course, each receiver has a different speaker and amplifier as well as slightly different AM mode passbands. But I listened for quite a while. The only drawback was the need to slightly retune if a station was off frequency a little in the round table.

Do I prefer it to DSB-AM? The jury is still out. I think some on air testing is needed. I know that if I were to join in a QSO with the TR-7, I would probably set the carrier to  -!KHz from the QSO frequency so as to ease the listening by others. If anyone else is interested in exploring this, let me know.

[W1EUJ]

The source of my original question came about when I was asked by another friend to fix up the Tempo One (a.k.a FT-200) and write an article about it.

The FT-200 uses a 7360 modulator tube for SSB. For AM and CW, it uses a 3M pot on one of the deflector plates to produce an imbalance and allow carrier through the balanced modulator circuit. In addition, a small 5pF cap is switched into the carrier oscillator to shift the carrier by 200 Hz, into the 2.7 KHz crystal filter passband. The FT-200, like many radios, shares part of the IF section for transmit and recieve - it is the same crystal filter used in reception.

I was thinking, for improving AM reception, of installing a switched 6 KHz filter, selected in AM mode. Thinking it through, this would also have the effect of passing the whole AM-modulated signal produced in the imbalanced balanced modulator, through the now-wider filter, then off to the driver circuits.

I have seen a number of mods for older SSB rigs, adding diode ring or Gilbert Cell mixers. Is a particular reason for adding a seperate/new modulator?

David Goncalves
W1EUJ

[WU2D]

Actually Don runs double sideband EXALTED carrier.

Mike WU2D