[Elecraft] K3 - SPE Expert 1K-FA and ALC

[Tom W8JI]

*We really suggest to always leave the transceiver's output 
power manual*
*control to its maximum level because the EXPERT, thanks to 
its exclusive*
*design, will provide to control that level as long as the 
required optimal*
*value will be reached.*
*Elecraft has only to implement a very fast input able to 
deal with negative
*voltages as all the other manufacturers already do: SPE's 
amplifier will*
*adapt to that in order to work for the best possible 
linearity.*

The ALC loop response can never be faster than group delay 
through the transmitter chain. While I have never measured 
the Elecraft filter, some crystal filters have a few mS 
delay. The propagation through the rest of the circuitry 
adds to the delay, as does the inevitable delay time in the 
actual ALC system. All of those delays, from the ALC control 
point forward to the ALC detection, and back from that point 
to the AGC controlled stage, add together.

Elecraft did a VERY smart thing by having the ALC detection, 
or at least part of the ALC detection, ahead of the narrow 
filters and everything else. Running the radio full on and 
relying on the amplifier ALC would undo this, and may not 
bring the results assumed.

Not only that, running other radios full on and depending 
100% on amplifier ALC is probably not a good idea either. 
The less "stuff" between the power sampling point and the 
gain controlled stage, the better off we are.

The whole reason some radios have overshoot is they run gain 
wide open and depend on a detector at the antenna port to 
reduce power, and the sum of delays from the gain controlled 
stage to the antenna, through the detector, through a 
comparator, and back to the gain controlled stage is what 
causes the problem in the first place. Moving the detection 
point out of the radio is never a benefit.

I can't think of a reason in the world to use the amplifier 
ALC for anything other than a safety limit, especially with 
a rig like the K3 that does gain control and ALC ahead of 
the filters and components that cause most of the delay in 
the ALC loop.

Can anyone else see why running the radio wide open and 
using 100% external detection is a good idea?

73 Tom