[Elecraft] attenuator in amp was Re: K3 and Expert 1K-FA

[Tom W8JI]

<<<Since the input Z of most solid-state amps is 
well-controlled at 50-ohms,
that the input Z should not appreciably change as output Z 
changes, then
can't we simply use a feed-forward ALC system instead of 
feedback ALC?   A
fast RF voltage peak detector can be bridged across the 
amp's input Z.  When
the RF voltage exceeds ~ 70.7Vpk, correction bias is applied 
to the
solid-state PA. Simultaneously, a 3dB pad could then be 
inserted at the amp
input until average input power drops below 50W.   >>>

We can protect the amp if we can change bias or switch a pad 
a lot faster than the rise time of the envelope, but that 
does not help the bandwidth. It could actually make the 
bandwidth problem worse, and is unlikely to make it better.

The ideal solution would be to have a system that remembers 
the ALC required, and starts the transmitter with just a bit 
less than the expected voltage and makes slow or small 
adjustments or corrections from there.

Better yet, why are most radios made the way they are? Put 
the SSB ALC entirely in the audio system. For example we 
could split-frequency process the audio with far more 
effective results and less intermodulation or harmonic 
distortion  than any type of RF clipping or compression. The 
VOMAX did that, although not nearly as well as we could do 
with a DSP system, in the 70's or 80's. We should be doing 
all the gain control and processing at audio.

Trying to fix transmitter issues at the amplifier by 
wrapping the amp back into the transmitter controls just 
doesn't seem logical. RF clipping or limiting, unless it is 
gain reduction with a PIN diode attenuator, doesn't make 
much sense to me.

73 Tom