On Thu, 2006-12-21 at 14:18 -0800, Ron Castro wrote:
> First off, I would think that the TT final amp can take quite a bit more
> peak output and still stay within its linear range. Certainly an occasional
> transient peak exceeding 100 Watts by 1 dB wouldn't "break the bank" and
> cause splatter. That being said, the average power output under a specific
> peak output restriction could be increased by a couple of the methods you
> mentioned.
But can it do 1 dB more output without increasing intermod products by
many more than 3 or 5 dB. E.g. is that 1 dB extra range (25%) still as
linear as the power range up to the rated power? I suspect its not.
>
> Split spectrum processing is certainly one way to increase average power,
> and this has been done as part of the old Alpha Vomax scheme, also it is
> used in the TS-870 processing, as well as in commercial broadcasting, both
> AM and FM where high average power is very much desired.
Drat, its hard to invent a new wheel!
>
> The comparisons that have been made with the Orions are usually against rigs
> that are in common use that show ALC level on an analog or digital meter of
> some sort. These rigs achieve their higher average to peak ratio by
> fast-acting "syllabic rate" ALC time constants. I recall that Collins had a
> Service Bulletin for the KWM-380 that increased ALC time constants to
> increase "talk power".
Yes there was a popular mod for the Yaesu FT-726 of hanging 10 uf on the
ALC line. What it did was delay the attack time so the rig ran into
distortion before the ALC could cut back the power. It put a spike on
the leading edge of the SSB envelope. Not a good mod or a friendly mod
for satellite users. Just like the vintage Tentec rigs with audio
derived AGC that put a click to the speaker on the leading edge of
received signals due to filter and AGC detector delay.
> Clipping with appropriate post-clipping filtering is also a useful way to
> bring up the average. Setting the SP carefully on the O2 seems to bring up
> the average level before distortion becomes noticeable, but you can't get
> very 'deep' before it gets dirty-sounding. It's OK for DX and contests,
> especially when the bass is curtailed, but not for every-day rag-chew.
its been known a long time that ordinary audio clipping makes for a
spikey SSB RF envelope. That it hurts the talk power wishes. What it
does is creates odd harmonics of the voice fundamentals and so puts MORE
tones to have that peaky vector sum I described before. Langevin used to
make a distribute clipper that didn't introduce so much noticeable audio
distortion (and VOA may still be using them) but the original I saw in
1965 was bigger than the Orion. Clipping at the IF is more effective,
but running that clipped signal through another SSB filter to clean up
the splatter introduces more distortion.
>
> The current thought in broadcasting is to use all of the above methods.
> Usually a wide-band AGC, followed by 3 to 6 spectral bands of AGC, followed
> by 3 to 6 spectral bands of peak limiting, followed by one or two stages of
> either very fast wide band peak limiting or hard clipping or both. All of
> those processes are DSP functions. That pretty well describes the
> characteristics of the best Orban and Omnia brands of processors. Obviously
> overkill for ham radio!
And what works for AM (especially high level plate or switching
modulation) is not the same for SSB and linear amplification. Brute
force clipping of AM audio didn't cause spikes in the envelope, like it
does for SSB. On AM clipping just clips the modulation peaks. And the
clipping looks the same at the RF envelope as it did for the audio
envelope. For the Collins 821A-1 (250 KW) VOA specified 12 dB of
clipping. When VOA came to test they cranked up the audio and ran more
than 20 dB of clipping. So running a sinewave the RF output looked like
a CW transmitter with the bug stuck on dots. The Langevin nicely rounded
the envelope corners even with a 50 Hz sinewave. Rolling along at a
megawatt peak. Would have been good for working pile ups on 20 meters,
even on AM! But OH the power bill.
>
> I think the problem could be fixed by changing the time constant on the ALC
> when in the SSB mode. The big question that still looms is whether this is
> a purely analog hardware function, or if it is controlled in any way the
> firmware.
Having a fast ALC turns the radio into an RF envelope compressor which
distorts the audio by emphasizing the weak components while reducing the
transmitted strong components and it does it based on time, not
frequency. Might as well let the PA clip, you get nearly the same
distortion products.
I suspect the Orion has analog RF detection which controls a gain stage
in the DSP, so likely the analog RF detector output goes to an A/D input
because gain control at digital level is super easy to accomplish with
any desired time constant. What I don't find in the Orion is a way to
manually set the drive so that ALC isn't working hard on some bands and
not so hard on others so it puts a hard corner on the CW envelope. Doug
Smith has a web page on that subject and there's an article in the
January QST about it too.
>
> Ron N6IE
> www.N6IE.com
> (Formerly N6AHA)
>
--
73, Jerry, K0CQ,
All content copyright Dr. Gerald N. Johnson, electrical engineer
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|