K3 SSB Power Control and ALC Metering

[wayne burdick]

There's been some recent discussion of how to properly set up the K3 in
SSB modes. I meant to respond sooner but was out of town for several
days.

MIC GAIN SETTING:  Adjust MIC gain for an ALC meter peak at around 5 to
7 bars during normal speech (rev. D owner's manual, page 28). This
applies even to TX TEST mode, meaning you can set it off-air. This
indication does not mean that you're "hitting the ALC really hard."
When you get to around 5 bars, you'll be "tickling" the DSP's ALC. More
on this below.

ALC METERING:  The K3's ALC meter is a bar graph, not an analog meter,
and we felt that 5 bars would provide the right granularity at the
target ALC level. The 5th bar of the ALC scale is heavier than the
others, serving as a reminder of the this level.

ALC DESIGN GOAL:  The K3 was designed to minimize transmit splatter and
other effects that plague some rigs. To achieve this, we apply all ALC
*before* the crystal filter, and minimize the application of ALC in
general. The levels have been carefully calibrated to that 5-to-7-bar
level mentioned above. The result is that the crystal and DSP filters
remove the types of artifacts that in many other rigs end up as
transmitted wideband noise or clicks. I believe this is why we continue
to hear excellent reports from K3 users regarding their on-the-air SSB
signals. (We've also heard, from some experts in the subject, that the
K3's compression is among the most effective they've ever measured.)

PER-BAND POWER CONTROL: With some vocalizations, speech energy may
build up within a narrow crystal filter to produce a slight peaking
effect. Because of this, the user must set the K3's power output level
such that it peaks at or below the safe level for any external
amplifier under all speech conditions. To facilitate this, you can use
per-band power control (refer to the PWR SET configuration menu entry).
We have many customers using this method. It compensates for per-band
gain variation in the external amp or the K3 itself, again reducing the
need for ALC. For example, all crystal filters have a small amount of
ripple (typ. +/- 0.5 to 1.0 dB) that can result in a slight difference
in average power output between LSB and USB with some voice
characteristics. But since you typically use only one of the two
sidebands on a given ham band, per-band power control can conveniently
compensate for small variations.

TX GAIN COMPENSATION FOR VOICE MODES (TXG VCE): If you find that your
speech peaks are routinely above the desired level (i.e., higher than
CW power output using TUNE), you can adjust the TXG VCE menu parameter
downward in 0.5 dB steps. Conversely, if your normal speech
characteristics result in lower power output compared to CW TUNE, you
can increase TXG VCE above the 0.0 dB level.

I'll be happy to answer any further questions. Our goal is to improve
the manual to make the above as clear as possible.

73,
Wayne
N6KR

* * *

OK?

W




---

http://www.elecraft.com

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[wayne burdick]

Hi Paul,

>> Been wondering just how the TX Gain calibration works. To quote from
>> the
>> manual - "The gain constant is updated whenever the TUNE function is
>> activated on a given band at one of three specific power levels:
>> 5.0W, 50W, and 1.00 milliwatt."
>
> To Stewart's point, does the gain of the K3 recalibrate every time
> TUNE is depressed at one of the three stated power levels... ?

Yes. It doesn't matter whether you select one of these power levels
manually using the PWR knob, or using the TUN PWR menu entry. It will
always recalibrate the gain constant during TUNE if you happen to be at
one of these power levels.

However, it will only do this if the SWR is 2.0:1 or less. This
provides plenty of safety margin.

> Finally, what prevents the re-loading of new gain data if as Stewart
> suggest, he sets the K3 power to exactly 5 watts or 50-watts and
> depresses TUNE for external ATU tuning purposes?  Should he instead
> stay away from those two values and use perhaps 49 or 51 watts?

No need to do this. The fact that it recalibrates every time you do
TUNE at one of these levels is not a problem. It just updates the
EEPROM constant if it has changed.

The TX gain constant is only there to compensate for small per-band
gain variation within the K3's transmitter. If you've done the power
calibration, then you can set PWR and be fairly close to the target
level on each band. Thus the MCU won't have to adjust power very much
in real time. In CW mode, this means we can apply our keying waveform
open-loop, preserving the target raised-cosine shape at all power
levels. In voice modes it helps achieve our goal of applying a minimum
of ALC.

73,
Wayne
N6KR


---

http://www.elecraft.com

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[Jim Brown-10]

Wayne,

I agree that the K3 can easily be made to transmit excellent audio,
and I've been getting glowing reports on mine. I have a question.
Exactly where in the signal chain is TXEQ with respect to Line In,
Mic In, Comp/Lim, and  the sensing for VOX?  The absolute best place
for TXEQ is after the point where the Mic In and Line In are summed
to feed VOX sensing, with Comp/Lim following that. Is that the
existing flow?  Are there different settings for Line In and Mic In?
I would expect (and hope) that TXEQ is disabled for digital modes.

73,

Jim K9YC


_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[KK7P]

MIC|LINE|DVR are summed, The result is sent to the VOX detector as well
as to the Tx Audio filter, which includes the TxEQ.

If the VOX detector were driven after the filtering, there would be
delay.  This way, the filter delay acts to improve VOX attack time.

Tx EQ is a global control at this time.  It does not currently change
with mode or source.  It is up to the operator of the radio to set it
correctly for the mode and source of the moment.  This is one of the
many items "on the list" for future enhancement.

73,

Lyle KK7P

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[Jim Brown-10]

On Tue, 23 Sep 2008 14:32:52 -0700, Lyle Johnson wrote:

>This way, the filter delay acts to improve VOX attack time.

Yes, but it also prevents the TXEQ filter from removing LF trash
(breathe pops, mic handling noise) that false triggers the VOX. And
I'm having that problem. For the same reason, we want the comp/lim
to be after the TXEQ.

Have you considered (and are you using) "feed forward" sensing for
comp/lim? This is widely used in pro audio. A few ms can do a lot to
catch fast peaks, and it's analogous to what you're doing with the
delay associated with EQ.

Jim



_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[KK7P]

Compression is after the filtering.  It is feed-forward.

VOX is before the filtering.

Tx EQ is inseparable from the rest of the Tx audio filtering/shaping.

73,

Lyle KK7P

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[Joe Subich, W4TV-3]

Wayne,

> The TX gain constant is only there to compensate for small per-band
> gain variation within the K3's transmitter. If you've done the power
> calibration, then you can set PWR and be fairly close to the target
> level on each band. Thus the MCU won't have to adjust power very much
> in real time.

Does the K3 include a "wattmeter slope" adjustment?  If so, I can't
find it.  

When the wattmeter is calibrated on 20 meters, I find a .75 to 1 dB
slope between 1.8 and 50 MHz when compared with both my LP-100 and
Bird wattmeters (power is about .25 dB high on 160 and a bit more
than .5 dB low on 50 MHz).  The slope appears to be consistent at
both 5 and 50 watts.

If a slope adjustment is not available, it might be worthwhile to
add the option to calibrate the wattmeter at two (160/10 or 160/6)
or three points (160/20/6) for frequency compensation.

73,

   ... Joe, W4TV
 

> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of wayne burdick
> Sent: Tuesday, September 23, 2008 4:42 PM
> To: Elecraft Reflector
> Subject: [Elecraft] Re: K3 - 'Unwanted' TX Gain calibrations?
>
>
> Hi Paul,
>
> >> Been wondering just how the TX Gain calibration works. To
> quote from
> >> the
> >> manual - "The gain constant is updated whenever the TUNE
> function is
> >> activated on a given band at one of three specific power levels:
> >> 5.0W, 50W, and 1.00 milliwatt."
> >
> > To Stewart's point, does the gain of the K3 recalibrate every time
> > TUNE is depressed at one of the three stated power levels... ?
>
> Yes. It doesn't matter whether you select one of these power levels
> manually using the PWR knob, or using the TUN PWR menu entry. It will
> always recalibrate the gain constant during TUNE if you
> happen to be at
> one of these power levels.
>
> However, it will only do this if the SWR is 2.0:1 or less. This
> provides plenty of safety margin.
>
> > Finally, what prevents the re-loading of new gain data if as Stewart
> > suggest, he sets the K3 power to exactly 5 watts or 50-watts and
> > depresses TUNE for external ATU tuning purposes?  Should he instead
> > stay away from those two values and use perhaps 49 or 51 watts?
>
> No need to do this. The fact that it recalibrates every time you do
> TUNE at one of these levels is not a problem. It just updates the
> EEPROM constant if it has changed.
>
> The TX gain constant is only there to compensate for small per-band
> gain variation within the K3's transmitter. If you've done the power
> calibration, then you can set PWR and be fairly close to the target
> level on each band. Thus the MCU won't have to adjust power very much
> in real time. In CW mode, this means we can apply our keying waveform
> open-loop, preserving the target raised-cosine shape at all power
> levels. In voice modes it helps achieve our goal of applying
> a minimum
> of ALC.
>
> 73,
> Wayne
> N6KR
>

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[P.B. Christensen]

Wayne,

Ahhh...I think I see now.  In effect, it's really an ALC loop gain
calibration, with ALC gain optimized for each band -- correct?  This sounds
very similar to what Ten Tec is using with their Omni VII, only their
calibration is conducted once at the factory to achieve optimum band-by-band
ALC gain.  Their rig also nicely preserves the raised-cosine CW waveform at
all power levels.

For some reason, I had this thought in my head that forward gain between say
the driver stage and PA was being changed in the firmware, irrespective of
the ALC.  Perhaps it was the "TX Gain" nomenclature...

Also, after the TX Gain calibration, I am seeing about 75-80W (with PWR set
to 100W) on 10M and 6M into my Bird TermaLine load and measured with LP-100
and Alpha 9510 WM.  No big deal but I just wanted to know if this is a
normal unit-by-unit variation?  This may, or may not be correlated with
Joe's comment concerning WM slope.

Paul, W9AC

----- Original Message -----
From: "wayne burdick" <[hidden email]>
To: "Elecraft Reflector" <[hidden email]>
Sent: Tuesday, September 23, 2008 4:41 PM
Subject: [Elecraft] Re: K3 - 'Unwanted' TX Gain calibrations?

> Hi Paul,
>
>>> Been wondering just how the TX Gain calibration works. To quote from the
>>> manual - "The gain constant is updated whenever the TUNE function is
>>> activated on a given band at one of three specific power levels: 5.0W,
>>> 50W, and 1.00 milliwatt."
>>
>> To Stewart's point, does the gain of the K3 recalibrate every time TUNE
>> is depressed at one of the three stated power levels... ?
>
> Yes. It doesn't matter whether you select one of these power levels
> manually using the PWR knob, or using the TUN PWR menu entry. It will
> always recalibrate the gain constant during TUNE if you happen to be at
> one of these power levels.
>
> However, it will only do this if the SWR is 2.0:1 or less. This provides
> plenty of safety margin.
>
>> Finally, what prevents the re-loading of new gain data if as Stewart
>> suggest, he sets the K3 power to exactly 5 watts or 50-watts and
>> depresses TUNE for external ATU tuning purposes?  Should he instead stay
>> away from those two values and use perhaps 49 or 51 watts?
>
> No need to do this. The fact that it recalibrates every time you do TUNE
> at one of these levels is not a problem. It just updates the EEPROM
> constant if it has changed.
>
> The TX gain constant is only there to compensate for small per-band gain
> variation within the K3's transmitter. If you've done the power
> calibration, then you can set PWR and be fairly close to the target level
> on each band. Thus the MCU won't have to adjust power very much in real
> time. In CW mode, this means we can apply our keying waveform open-loop,
> preserving the target raised-cosine shape at all power levels. In voice
> modes it helps achieve our goal of applying a minimum of ALC.
>
> 73,
> Wayne
> N6KR
>
>
> ---
>
> http://www.elecraft.com
>
> _______________________________________________
> Elecraft mailing list
> Post to: [hidden email]
> You must be a subscriber to post to the list.
> Subscriber Info (Addr. Change, sub, unsub etc.):
> http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/subscribers.htm
> Elecraft web page: http://www.elecraft.com
>

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[AD6XY]

Thanks for the detailed information. However, this is not my problem.

The most severe problem I have is:

a) the output power spikes and overshoots in CW or SSB or FM as you adjust the power control, so it is not advisable to use a sensitive, i.e solid state power amplifier with the K3. Especially not one needing less than 50W of drive. It needs a fast responding meter to see it.

A work around is to set the power without the amplifier in circuit, speak or key to let the spike happen, then switch on the PA and not adjust power or mode afterwards.

Does this not happen to anyone else?

A lesser problem is:
 
b) that the power output is unstable once set in SSB mode, showing variations of around 50-100% in the peak output depending on the power output setting. This is when set up as described. It is only really visible with a peak reading power meter.

[dj7mgq]

Hi,

> Does this not happen to anyone else?

Re: a) -> I've never noticed this. But it might be occurring.

Re: b) -> I've seen my K3 do this. If I set the power in CW or Tune  
mode, and don't touch the power control, usually the power level will  
stay constant in SSB. With my K3 the problem appears to be worse at  
lower power levels (e.g. 30W) than at higher ones (e.g. 90W). In newer  
versions of the firmware there has been, imho, distinct improvement  
but not a total solution. I haven't had the time to update to very  
newest firmware yet and test it.

Elecraft has in the past said that they are working on this, so I do  
hope they will solve this in the firmware soon, assuming that it's not  
something else in the hardware. Up to now, it has been more of a quirk  
than a real problem for me.

vy 73 de toby
--
DD5FZ ( ex 4n6fz, dj7mgq, dg5mgq, dd5fz)
K2 #885
K2/100 #3248
K3/100 #67



_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[DaveL G3TJP]

Thanks for that very useful guide, Wayne.  Much appreciated.

I searched to see if others had been written but failed to find any more, which is a pity because such gems seem worthy of  their own little archive "place".

73  DaveL  G3TJP
_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com