[Elecraft] Erratic transmit power -- ALC oscillating?

[Don Wilhelm]

  Bob,

The spikes you are seeing in the ALC circuit are normal, and are not 
associated with any problem you may be having.

Is the variation in power observed when you press the TUNE button?  If 
so, that is normal too.  The level is not controlled as tightly in TUNE 
as it is with normal keying as a help to those who would be using a 
manual tuner with the K2.

You apparently have an oscilloscope - have you observed the keying 
waveform by connecting the 'scope across a 50 ohm dummy load?  Does the 
waveform have a shape like that shown in the Keying Waveform Mod 
instructions?  If so, that is OK.

Does the power output oscillate with normal keying?  The LED bargraph or 
your 'scope across a dummy load should reveal the instantaneous RF 
level, and it should not vary.  If it does, there is an application note 
that will tell you how to determine the optimum value of RF Board R98 
for your particular K2.

Yes, you should have used the capacitor from the Keying Waveshape mod at 
CB C31.

Are you having problems with HiCur messages and low output power?  I am 
confused by your statements about the PA and LPF - there was no problem 
with that part explicitly stated - any problem there is not related to 
the rest of your post.

73,
Don W3FPR

On 2/10/2011 4:00 PM, n7zf wrote:
> This is a new build of a rev A K2 (installed rev B mods), rework eliminators,
> no options installed, several other mods installed including the CW waveform
> keying mod (VALC circuit).  Receiver is working well.
>
> If I set the output power to 5 watts,  the usually K2 reports the power as
> 4.2 watts or 6.3 watts -- sometimes 5.1 watts.  I am seeing some unexpected
> things in the ALC circuit on the control board and am wondering if the ALC
> is causing the transmit power to fluctuate.
>
> In both transmit (cw) and receive mode, on all bands, there is a 45 hertz
> 0.7 volt spike coming out of U10A (LMC660), causing Q8 to turn on for about
> 3.5 msecs every 22 msecs.  The B-E junction of Q8 is limiting the spike
> voltage.  See photo of signal on pin 1 of U10 -- same as base of Q8.  Using
> a 10x probe, so vertical scale is 200 mV per division, horizontal scale is 5
> msec per division.
>
> At the junction of R22 (82K) and R21 (270K) I am seeing a 1volt peak-to-peak
> signal being fed back by C46 (.01).  Is this expected?  I see this 'ripple'
> continue as a 150 mV peak at Test Point 2, and also on the output of the
> Xmit Mixer U10-pin4 during transmit.  I have removed the control board and
> verified all the correct components against the Rev. H schematic.  I have
> traced all the connections in that circuit.  I have rechecked against the
> keying mod instructions.  The CW key waveform modifications appear to be
> working correctly but is the ALC circuit supposed to be oscillating?
>
> The CW keying mod (rev A 2/3/2004) specifies C31 to be .01 uF.  The Rev
> A-to-B upgrade (rev E 4/29/2003) specifies C31 to be .047 uF.  I assumed
> that the CW keying mod was a more robust change and followed it (so C31 is
> .01 uF).
>
> Same problem running off a power supply or a battery.  I have installed new
> power amp finals and new power amp bias transistors.  I have carefully
> inspected T1-T4 in the final, including removing T3 and rewinding it.  I
> have tested the low pass filters with a network analyzer and found no
> problems.  I have measured and inspected the RF Output Detector circuit (and
> have the 1% resistor at R68 which measures to be 225.0 ohms).
>
>