New K3 Firmware: Pileup-inspired AGC changes

[wayne burdick]

There's been a lot of discussion recently about what happens to
multiple closely spaced signals with AGC turned on. We have a new
field-test firmware revision available that offers a significant
improvement in such situations by reducing the IMD caused by
conventional AGC. If you'd like to try it, please send an email to
[hidden email].

Details:

Some operators described a "merging" of multiple, close-spaced signals
when listening to pileups, making individual signals hard to copy.
Assuming that a narrow crystal filter is in use (close to the DSP
bandwidth), this is not "desense". In most cases, we've discovered, the
cause is IMD (intermodulation distortion) related to the AGC algorithm.

To simulate what happens under pileup conditions, we combined four
crystal oscillators with gaps between them of 5, 6, and 7 Hz (a total
spread of 18 Hz for the four signals). We then injected the signals
into the K3 at an equal level of about S5 and used spectral analysis to
examine what happens to the passband. With AGC-F turned on (AGC fast,
set to the K3's factory defaults), IMD products came up some 20-25 dB
as compared to AGC OFF. The situation could be far worse in an actual
on-air situation with more signals, noise, and key clicks. (AGC-S makes
only a small improvement.) We ran the same test on several other
radios, and verified the AGC-induced IMD characteristics are similar.

IMD is generated because AGC can cause mixing between the tones and
their harmonics. The faster the decay rate, in general, the more
products will be generated. With just one or two signals, most products
will be outside the DSP's narrow IF or AF passband. But in a pileup,
where signals are greater in number and can be nearly on top of each
other, there are many more opportunities for signals to combine. This
results in IMD products spaced at 1 to 20 Hz, depending on the time
constants, location of the signals, etc.

Turning AGC off is one strategy that operators have traditionally used
to combat the effects of AGC modulation. It then becomes necessary to
"ride" the gain controls. Neither Lyle nor I have that much patience :)
  So we took a different approach: keep the AGC on, but reduce its
modulation effects to negligible levels by adding a very long "hold"
time.

In the field-test firmware, two changes have been made. First, the AGC
HLD menu setting, which used to apply only to voice modes, now applies
to CW as well. (It still only affects AGC-S, the slow AGC setting.)
Second, we increased the range of AGC HLD to facilitate
experimentation. What we found is that an AGC hold time of about 1
second, combined with a slower AGC-S decay time, reduces in-band IMD
with closely spaced signals by 20-25 dB, making it very similar to
turning AGC off. The *attack* time is unaffected, so there is no
"thumping" as you might have with audio-derived AGC. Any new signal
peak above the one that starts the "hold" timer will restart the timer.

This "hold" AGC appears to be useful in many listening situations. It
can make voice signals sound cleaner, as well; after all, a voice
signal can have many components, and AGC-induced modulation can add IMD
products. It also helps with noisy band conditions.

The down side to AGC "hold" is that large signals can charge it up to a
high level, reducing gain for the duration of the hold time. But in
many cases this is far preferable to the AGC IMD effects, especially if
you're listening to many signals at similar amplitudes -- as in a
pileup.

73,
Wayne, N6KR
Lyle, KK7P


---

http://www.elecraft.com

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[wayne burdick]

We now have plenty of volunteers to try this experimental AGC firmware.
Thanks!

73,
Wayne
N6KR

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[wayne burdick]

To those of you who are testing this: Remember that AGC HLD only
affects slow AGC (AGC-S). So, once you have a long AGC HLD time set up
(0.3 to 1 sec, let's say), you can go back and forth between fast AGC
(with no hold), and slow AGC (with a long hold) just by tapping the AGC
switch.

Long-hold/slow AGC isn't going to be the best choice every situation.
Some experimentation will be required. Please let me know what settings
you're using, and what you've listened to (pileups, general QRM, QRN,
voice signals, etc.).

tnx
Wayne
N6KR


---

http://www.elecraft.com

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[.hank.]

Gee I wish this had been available last week - we just got back from a week
in Nicaragua.

We used a K3 in the ARRL DX  CW test for about 4700 QSOs.  Well over half of
them were under medium to heavy pileup conditions with AGC-F.  I found at
times turning on the attenuator helped .... probably not as much as AGC off
and 'ride' the RF gain control, but I don't have a whole lot of patience
with that either.
73    Hank    K8DD   (YN2DD)



On Wed, Feb 25, 2009 at 6:45 PM, wayne burdick <[hidden email]> wrote:

> There's been a lot of discussion recently about what happens to
> multiple closely spaced signals with AGC turned on. We have a new
> field-test firmware revision available that offers a significant
> improvement in such situations by reducing the IMD caused by
> conventional AGC. If you'd like to try it, please send an email to
> [hidden email].
>
> Details:
>
> Some operators described a "merging" of multiple, close-spaced signals
> when listening to pileups, making individual signals hard to copy.
> Assuming that a narrow crystal filter is in use (close to the DSP
> bandwidth), this is not "desense". In most cases, we've discovered, the
> cause is IMD (intermodulation distortion) related to the AGC algorithm.
>
> To simulate what happens under pileup conditions, we combined four
> crystal oscillators with gaps between them of 5, 6, and 7 Hz (a total
> spread of 18 Hz for the four signals). We then injected the signals
> into the K3 at an equal level of about S5 and used spectral analysis to
> examine what happens to the passband. With AGC-F turned on (AGC fast,
> set to the K3's factory defaults), IMD products came up some 20-25 dB
> as compared to AGC OFF. The situation could be far worse in an actual
> on-air situation with more signals, noise, and key clicks. (AGC-S makes
> only a small improvement.) We ran the same test on several other
> radios, and verified the AGC-induced IMD characteristics are similar.
>
> IMD is generated because AGC can cause mixing between the tones and
> their harmonics. The faster the decay rate, in general, the more
> products will be generated. With just one or two signals, most products
> will be outside the DSP's narrow IF or AF passband. But in a pileup,
> where signals are greater in number and can be nearly on top of each
> other, there are many more opportunities for signals to combine. This
> results in IMD products spaced at 1 to 20 Hz, depending on the time
> constants, location of the signals, etc.
>
> Turning AGC off is one strategy that operators have traditionally used
> to combat the effects of AGC modulation. It then becomes necessary to
> "ride" the gain controls. Neither Lyle nor I have that much patience :)
>  So we took a different approach: keep the AGC on, but reduce its
> modulation effects to negligible levels by adding a very long "hold"
> time.
>
> In the field-test firmware, two changes have been made. First, the AGC
> HLD menu setting, which used to apply only to voice modes, now applies
> to CW as well. (It still only affects AGC-S, the slow AGC setting.)
> Second, we increased the range of AGC HLD to facilitate
> experimentation. What we found is that an AGC hold time of about 1
> second, combined with a slower AGC-S decay time, reduces in-band IMD
> with closely spaced signals by 20-25 dB, making it very similar to
> turning AGC off. The *attack* time is unaffected, so there is no
> "thumping" as you might have with audio-derived AGC. Any new signal
> peak above the one that starts the "hold" timer will restart the timer.
>
> This "hold" AGC appears to be useful in many listening situations. It
> can make voice signals sound cleaner, as well; after all, a voice
> signal can have many components, and AGC-induced modulation can add IMD
> products. It also helps with noisy band conditions.
>
> The down side to AGC "hold" is that large signals can charge it up to a
> high level, reducing gain for the duration of the hold time. But in
> many cases this is far preferable to the AGC IMD effects, especially if
> you're listening to many signals at similar amplitudes -- as in a
> pileup.
>
> 73,
> Wayne, N6KR
> Lyle, KK7P
>
>
> ---
>
> http://www.elecraft.com
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
> <*> To visit your group on the web, go to:
>    http://groups.yahoo.com/group/Elecraft_K3/
>
> <*> Your email settings:
>    Individual Email | Traditional
>
> <*> To change settings online go to:
>    http://groups.yahoo.com/group/Elecraft_K3/join
>    (Yahoo! ID required)
>
> <*> To change settings via email:
>    mailto:[hidden email]
>    mailto:[hidden email]
>
> <*> To unsubscribe from this group, send an email to:
>    [hidden email]
>
> <*> Your use of Yahoo! Groups is subject to:
>    http://docs.yahoo.com/info/terms/
>
>

--
"I knew that God put me on this earth to be on the radio."    Ed Bradley

"Me too!"    K8DD
-
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html