[Elecraft] Fwd: Re: K3-P3 SSB spectrum

[Don Wilhelm]

Joe,

I disagree with you, but it may be only in semantics - there *must* be a 
control loop to obtain proper power control.  How that is implemented is 
different in the K2, K3, and KX3 as compared with other transceivers 
which only provide a knob to control the drive level and then depend on 
"ALC" which in their definition is detection of overload to reduce the 
drive level.

The K2, K3, KX3 does use a *closed* control loop for power control. The 
actual power output is compared with the power requested by the Power 
Knob, and the drive is adjusted by the MCU to make the two match.  The 
loop damping is set to minimize overshoot while also minimizing the 
oscillation of the control loop.  Ideally, that power control loop will 
be 'critically damped' which will result in no overshoot and no loop 
oscillation.  The K3 achieves that critically damped condition in all 
tests that I have done.

Yes, this enters into control loop theory, and the way the K2, K3, KX3 
control power is unlike that seen in other amateur transceivers.  Other 
transceivers depend on detecting an overload condition (ALC) and as a 
result reduce the drive level.  The Elecraft transceivers do it differently.
The K2, K3, KX3 start at a lower power level (after a band change or 
power level change) and ramp up to the requested level until they are equal.
The "downside" is that it takes some RF (a dit time or two, or a few SSB 
sylables) for the K3 to come up to the requested power.  The 'upside' is 
that it is done with no overshoot and does not depend on detection that 
some stage is being overdriven.
The actual output power is detected by the wattmeter which is right at 
the output.  That information is fed back to the MCU where it is 
compared with the requested power and the MCU alters the drive as 
required to make the two match.

That is why some operators have trouble with not providing adequate 
audio drive for digital modes.  If the audio level is too low, the drive 
will be ramped up to (hopefully) provide the requested power output, and 
that can result in "power hunting".  The "conventional" wisdom given on 
the internet for soundcard digital modes which is to run the power 
control of the transceiver at max and use the audio drive to control the 
power output does not work with Elecraft transceivers.  One must first 
set the audio level correctly (and then leave it alone) and adjust the 
power level with the power request knob.  Leave that "conventional" 
wisdom to other transceivers (where it does work well), but do not use 
it with your Elecraft gear.

73,
Don W3FPR

On 6/8/2015 7:44 PM, Joe Subich, W4TV wrote:
>
>> The internal power control loop then measures the power level of the
>> RF being made and adjusts it to meet the requested level.
>
> There is *no "power control loop"* in the K3 - that is the difference
> between the K3/K3S and traditional (YaeComWoood) transceivers. When
> you (or the factory) performs the TX Gain calibration, the rig learns
> how much (digital) gain is required in the DSP stage to drive the LPA
> (5W setting) and the KPA3 (50W setting) to reach the desired power
> level.
>
> Since the transmit chain of the K3/K3S is linear (unlike YaeComWood),
> when you request 100W instead of 50W the controller knows to increase
> the voltage output of the DSP by a factor of 4 (3dB) while if you
> request 20W the output of the DSP will be reduced by 4 dB (20/50 W =
> -3.98 dB).
>
> Because there is no "power control loop" the K3/K3S does not suffer
> from overshoot, ALC induced distortions, IF clipping, PA over-drive,
> etc. that plague traditional solid state transceivers.
>
> 73,
>
>    ... Joe, W4TV