Elecraft › [K3]

Re: K3 Rx IMD

Classic

List

Threaded

11 messages Options

Geoffrey Mackenzie-Kennedy

Re: K3 Rx IMD

Hi Ramiro,

Unfortunately I do not have a K3 to test because my receivers are
homebrewed.

I still believe that the IMD is generated by something outside of the
radios, one reason being that your results from the K3 and the FT-897 are
not very different. The differences between the AGC-S, AGC-F and AGC-OFF
results could be a symptom of some other problem.

73,
Geoff
GM4ESD

Ramiro Aceves wrote on Tuesday, November 23, 2010, at 9:41 PM

> If you have the time, can you make a 10Hz spaced test AGC-F, AGC-S and
> AGC-off? I will appreciate that. Of course if you hace free time for that.

______________________________________________________________
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

Ramiro Aceves

Re: K3 Rx IMD

Geoffrey Mackenzie-Kennedy escribió:
> Hi Ramiro,
>
> Unfortunately I do not have a K3 to test because my receivers are
> homebrewed.

That is a very interesting thing. I have also homebrewed receivers. They
are far from perfect but I love them, hi.

>
> I still believe that the IMD is generated by something outside of the
> radios, one reason being that your results from the K3 and the FT-897
> are not very different. The differences between the AGC-S, AGC-F and
> AGC-OFF results could be a symptom of some other problem.

That makes sense. Once generated they would be increased by the K3 "AGC
pump" How can I test that?. Can the toroid inside the antenna T tunner
be the culprit? Any ideas to test will be welcome.
Anyway, Kok tests so that AGC-F makes IMD.

73, Ramiro.

>
> 73,
> Geoff
> GM4ESD
>
>
> Ramiro Aceves wrote on Tuesday, November 23, 2010, at 9:41 PM
>
>> If you have the time, can you make a 10Hz spaced test AGC-F, AGC-S and
>> AGC-off? I will appreciate that. Of course if you hace free time for
>> that.
>
>
>
>
>

______________________________________________________________
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

Geoffrey Mackenzie-Kennedy

Re: K3 Rx IMD

Sorry for being slow to reply.

Oh yes, a coil wound on a core made from "Ferrite" or "Powdered Iron" will
be a source of IMD products, the culprit being the core. The level of the
IMD products generated by a core depends on a number of factors, which we
could discuss "Off List" if you wish.

While I understand Lyle's point concerning the AGC, I cannot shake my
suspicion that something in or near your antenna is behaving as a mixer, and
is the source of IMD products. The problem with adding the DDS' signal by
coupling to the antenna is that you have no control over "unwanted" mixers
in the vicinity. To disprove (or confirm) my suspicion, my suggestion is
that you build a hybrid combiner, and use it to inject the BC and DDS
signals into the K3.

73,

Geoff
GM4ESD

On Wednesday, November 24, 2010, at 7:59 AM, Ramiro Aceves wrote?

> Can the toroid inside the antenna T tunner
> be the culprit? Any ideas to test will be welcome.

______________________________________________________________
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

David Cutter

Re: K3 Rx IMD

Please stay on the list, guys, this is interesting stuff that I'm sure many
would like to hear more of.

David
G3UNA

> Sorry for being slow to reply.
>
> Oh yes, a coil wound on a core made from "Ferrite" or "Powdered Iron" will
> be a source of IMD products, the culprit being the core. The level of the
> IMD products generated by a core depends on a number of factors, which we
> could discuss "Off List" if you wish.
>
> While I understand Lyle's point concerning the AGC, I cannot shake my
> suspicion that something in or near your antenna is behaving as a mixer,
> and
> is the source of IMD products. The problem with adding the DDS' signal by
> coupling to the antenna is that you have no control over "unwanted" mixers
> in the vicinity. To disprove (or confirm) my suspicion, my suggestion is
> that you build a hybrid combiner, and use it to inject the BC and DDS
> signals into the K3.
>
> 73,
>
> Geoff
> GM4ESD
>
>
> On Wednesday, November 24, 2010, at 7:59 AM, Ramiro Aceves wrote?
>
>
>> Can the toroid inside the antenna T tunner
>> be the culprit? Any ideas to test will be welcome.
>
>
>

Ramiro Aceves

Re: K3 Rx IMD

Hello all

As David suggested we continue the discussion on the list. I have done
some new poor man tests, using a local AM carrier on Medium Wave at 917
KHz. For generating the other tone, I am using the same home made DDS
generator as before. At this lower frequencies the signal generated by
the DDS is much more cleaner. The antenna is a 6.5m long dipole on the
balcony and an open wire feeder as in the earlier tests, but using only
one half of the dipole by just directly connecting one open wire end to
the K3 antenna connector. The reasons of doing that is that my T match
tunner does not work below 1.8 MHz. This eliminates the posibility of
IMD products produced on the tuner toroids. It does not remove, just in
case, other no-linearities found everywhere as suggested by (sorry for
my poor english, I hope you understand me).

10 Hz spacing, ACG-F:

http://ea1abz.ure.es/acg-F.png

Now with AGC-S, notice the strong reduction of IMD:

http://ea1abz.ure.es/agc-S.png

Now see the AGS-S to AGC-S-soft transition on the middle of the capture:

http://ea1abz.ure.es/agc-S_vs_agc-soft.png

Now see the IMD at 100Hz spacing, AGC-F to AGC-S transition at the
middle of the screen:

http://ea1abz.ure.es/100Hz.png

Sorry for the poor noise floor, is the Madrid band noise received by the
antenna. Hi.

I see that the IMD problem is worse with small carrier distances. At 100
Hz spacing they are much more weaker and not problematic.

73, Ramiro.

EA4NZ
______________________________________________________________
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

Guy, K2AV

Re: K3 Rx IMD

You do know that you cannot do valid IMD sweep measurements with AGC on?

First thing I have to do for any measurements is turn off the AGC and set
fixed levels. Particularly very fast config of fast AGC won't work. It
won't be valid for absolute measurements, but the bandwidth needs to be at
least four times the spacing of the two to prevent AGC severely warping the
content. And then you will need to provide your single tone sweep
selectivity EXTERNAL to the K3. Your fast AGC is pulling up the IMD tones
by changing the RX gain during the sweep.

73, Guy.

On Wed, Nov 24, 2010 at 3:07 PM, Ramiro Aceves <[hidden email]> wrote:

> Hello all
>
> As David suggested we continue the discussion on the list. I have done
> some new poor man tests, using a local AM carrier on Medium Wave at 917
> KHz. For generating the other tone, I am using the same home made DDS
> generator as before. At this lower frequencies the signal generated by
> the DDS is much more cleaner. The antenna is a 6.5m long dipole on the
> balcony and an open wire feeder as in the earlier tests, but using only
> one half of the dipole by just directly connecting one open wire end to
> the K3 antenna connector. The reasons of doing that is that my T match
> tunner does not work below 1.8 MHz. This eliminates the posibility of
> IMD products produced on the tuner toroids. It does not remove, just in
> case, other no-linearities found everywhere as suggested by (sorry for
> my poor english, I hope you understand me).
>
> 10 Hz spacing, ACG-F:
>
> http://ea1abz.ure.es/acg-F.png
>
>
> Now with AGC-S, notice the strong reduction of IMD:
>
> http://ea1abz.ure.es/agc-S.png
>
>
>
> Now see the AGS-S to AGC-S-soft transition on the middle of the capture:
>
> http://ea1abz.ure.es/agc-S_vs_agc-soft.png
>
>
>
> Now see the IMD at 100Hz spacing, AGC-F to AGC-S transition at the
> middle of the screen:
>
> http://ea1abz.ure.es/100Hz.png
>
>
> Sorry for the poor noise floor, is the Madrid band noise received by the
> antenna. Hi.
>
> I see that the IMD problem is worse with small carrier distances. At 100
> Hz spacing they are much more weaker and not problematic.
>
>
> 73, Ramiro.
>
> EA4NZ
> ______________________________________________________________
> 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
>

Ramiro Aceves

Re: K3 Rx IMD

Thanks Guy for the explanations about IMD measurements. I need a signal
generator to do better tests and get rid of the need of receiving a
broadcast carrier. I was thinking in building a simple Xtal oscillator
and some attenuators.

73, Ramiro. EA4NZ

Guy Olinger K2AV escribió:

> You do know that you cannot do valid IMD sweep measurements with AGC on?
>
> First thing I have to do for any measurements is turn off the AGC and
> set fixed levels. Particularly very fast config of fast AGC won't
> work. It won't be valid for absolute measurements, but the bandwidth
> needs to be at least four times the spacing of the two to prevent AGC
> severely warping the content. And then you will need to provide your
> single tone sweep selectivity EXTERNAL to the K3. Your fast AGC is
> pulling up the IMD tones by changing the RX gain during the sweep.
>
> 73, Guy.
>
> On Wed, Nov 24, 2010 at 3:07 PM, Ramiro Aceves <[hidden email]
> <mailto:[hidden email]>> wrote:
>
> Hello all
>
> As David suggested we continue the discussion on the list. I have done
> some new poor man tests, using a local AM carrier on Medium Wave at 917
> KHz. For generating the other tone, I am using the same home made DDS
> generator as before. At this lower frequencies the signal generated by
> the DDS is much more cleaner. The antenna is a 6.5m long dipole on the
> balcony and an open wire feeder as in the earlier tests, but using only
> one half of the dipole by just directly connecting one open wire end to
> the K3 antenna connector. The reasons of doing that is that my T match
> tunner does not work below 1.8 MHz. This eliminates the posibility of
> IMD products produced on the tuner toroids. It does not remove, just in
> case, other no-linearities found everywhere as suggested by (sorry for
> my poor english, I hope you understand me).
>
> 10 Hz spacing, ACG-F:
>
> http://ea1abz.ure.es/acg-F.png
>
>
> Now with AGC-S, notice the strong reduction of IMD:
>
> http://ea1abz.ure.es/agc-S.png
>
>
>
> Now see the AGS-S to AGC-S-soft transition on the middle of the capture:
>
> http://ea1abz.ure.es/agc-S_vs_agc-soft.png
>
>
>
> Now see the IMD at 100Hz spacing, AGC-F to AGC-S transition at the
> middle of the screen:
>
> http://ea1abz.ure.es/100Hz.png
>
>
> Sorry for the poor noise floor, is the Madrid band noise received by the
> antenna. Hi.
>
> I see that the IMD problem is worse with small carrier distances. At 100
> Hz spacing they are much more weaker and not problematic.
>
>
> 73, Ramiro.
>
> EA4NZ
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email] <mailto:[hidden email]>
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>
>

Ramiro Aceves

Re: K3 Rx IMD

In reply to this post by Guy, K2AV

Hi

I have tried to improve my IMD test setup. It is far from perfect, but I
have got rid of the broadcast station by bulding "quick and dirty " Xtal
4 MHz oscillator. It is built using a BC547 NPN bipolar transistor.

http://ea1abz.ure.es/XTAL.jpg

This is the beaufitul 4 MHz signal as received in the K3:

http://ea1abz.ure.es/xtal.png

As I do not have attenuatores nor combiners at the moment, the ugly
setup is:

http://ea1abz.ure.es/DDSandXTAL.jpg

50 cm wire conected at K3 antenna input.
4 MHz Xtal oscillador placed near the K3 wire antenna to produce a 9+20
signal.
DDS generator tuned 10 Hz over the Xtal frequency, same level as Xtal
signal.

Here is the graph AGC-OFF, AGC-S, AGC-F comparison. I reduced RF gain in
AGC-OFF mode to keep two tone audio signals level at the same value.
Aproximate measured 3rd order products, relative to one signal level is:

AGC-F: -17 dB
AGC-S: -23 dB
AGC-OFF: -45 dB.

http://ea1abz.ure.es/AGC-off_AGC-S_AGC-F.png

Let me know what do you think. This it what I can do with this poor
setup. In the future, I plan to build attenuators and hybrid combiner to
make a better test.

73, Ramiro. EA4NZ

______________________________________________________________
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

Geoffrey Mackenzie-Kennedy

Re: K3 Rx IMD

In reply to this post by Geoffrey Mackenzie-Kennedy

Ramiro,

Some very good information about Combiners is published on Jack Smith's
website (Clifton Laboratories), also how to perform "IMD" tests.

The reason for using a lowpass filter between each generator and the
combiner is to prevent the second harmonic of generator "A", for example,
getting into generator "B", mixing with the fundamental of "B" to produce a
third order product 2A - B. Similarly to keep the second harmonic of "B" out
of "A". Any mixing of other harmonics can also result in "alien" third order
products.

If you build a "pretty" version, I would suggest that each generator be
built in its own metal box, with all DC supply leads filtered. What you do
not want is the generators talking to each other :-)

Good luck with your project.

73,
Geoff
GM4ESD

Ramiro Aceves wrote on Thursday, November 25, 2010, at 7:52 AM:

> No problem. I cant resist what you propose. I am going to build Xtal
> oscillator. I have a 4 MHz cristal here. The the combiner.
> Keep tuned!
>
> Ramiro

______________________________________________________________
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

Ramiro Aceves

Re: K3 Rx IMD

Thanks Geoffrey for the interesting information about IMD measuring. I
see that it is a difficult task, it implies using carefully made
equipment and care. Everything is out of my possibilities at the moment.
I have made these easy experiments just trying to have an explanation of
what occured in my K3 when tried to see aircraft reflections and got
duplicated reflections. While they are not perfect, these tests have
teached me implications of AGC-F/S and have duplicated a IMD problem
under more "controlled" circumstances.

Trying to obtain a decent test gear is a difficult task for me, and
perhaps it makes no sense to invest time and money on it. I am not a
receiver tester, just a curious person who likes to know why things
happen.

Anyway, since I changed AGC-F to AGC-S or even better AGC-off, I have no
longer received duplicated traces on the screen. If I change to AGC-F I
get consistent duplicated traces. I do not certainly know if the spurs
are generated out or inside the K3, I tend to think that the latter. But
who knows, perhaps the K3 receiver is not as perfect as many of us think.

Thank you, will keep in touch.

Ramiro. EA4NZ.

Geoffrey Mackenzie-Kennedy escribió:

> Ramiro,
>
> Some very good information about Combiners is published on Jack Smith's
> website (Clifton Laboratories), also how to perform "IMD" tests.
>
> The reason for using a lowpass filter between each generator and the
> combiner is to prevent the second harmonic of generator "A", for
> example, getting into generator "B", mixing with the fundamental of "B"
> to produce a third order product 2A - B. Similarly to keep the second
> harmonic of "B" out of "A". Any mixing of other harmonics can also
> result in "alien" third order products.
>
> If you build a "pretty" version, I would suggest that each generator be
> built in its own metal box, with all DC supply leads filtered. What you
> do not want is the generators talking to each other :-)
>
> Good luck with your project.
>
> 73,
> Geoff
> GM4ESD
>
>
> Ramiro Aceves wrote on Thursday, November 25, 2010, at 7:52 AM:
>
>> No problem. I cant resist what you propose. I am going to build Xtal
>> oscillator. I have a 4 MHz cristal here. The the combiner.
>> Keep tuned!
>>
>> Ramiro
>
>
>
>

Joe Subich, W4TV-4

Re: K3 Rx IMD

Ramiro,

> Anyway, since I changed AGC-F to AGC-S or even better AGC-off, I have
> no longer received duplicated traces on the screen. If I change to
> AGC-F I get consistent duplicated traces. I do not certainly know if
> the spurs are generated out or inside the K3, I tend to think that
> the latter.

Try setting AGC DCY = SoFt as recommended by Lyle, KK7P. You are
seeing the effect of the BEAT between the two signals modulating
the DSP AGC. The SOFT AGC Decay was designed specifically for the
situation when the two signals were within the AGC loop filter!

I believe you noted that the effect was almost non-existent when
the two signals were 100 Hz apart vs. 10 Hz. This is the nature
of nearly every receiver with DSP based AGC. The DSP is so much
FASTER than conventional analog AGC that these effects (AGC induced
IMD) begin to appear.

Fortunately Lyle and Wayne recognized this issue some time ago and
have provided the tools to deal with it.

73,

... Joe, W4TV

On 11/25/2010 4:56 PM, Ramiro Aceves wrote:

> Thanks Geoffrey for the interesting information about IMD measuring. I
> see that it is a difficult task, it implies using carefully made
> equipment and care. Everything is out of my possibilities at the moment.
> I have made these easy experiments just trying to have an explanation of
> what occured in my K3 when tried to see aircraft reflections and got
> duplicated reflections. While they are not perfect, these tests have
> teached me implications of AGC-F/S and have duplicated a IMD problem
> under more "controlled" circumstances.
>
> Trying to obtain a decent test gear is a difficult task for me, and
> perhaps it makes no sense to invest time and money on it. I am not a
> receiver tester, just a curious person who likes to know why things
> happen.
>
> Anyway, since I changed AGC-F to AGC-S or even better AGC-off, I have no
> longer received duplicated traces on the screen. If I change to AGC-F I
> get consistent duplicated traces. I do not certainly know if the spurs
> are generated out or inside the K3, I tend to think that the latter. But
> who knows, perhaps the K3 receiver is not as perfect as many of us think.
>
> Thank you, will keep in touch.
>
> Ramiro. EA4NZ.
>
>
>
> Geoffrey Mackenzie-Kennedy escribió:
>> Ramiro,
>>
>> Some very good information about Combiners is published on Jack Smith's
>> website (Clifton Laboratories), also how to perform "IMD" tests.
>>
>> The reason for using a lowpass filter between each generator and the
>> combiner is to prevent the second harmonic of generator "A", for
>> example, getting into generator "B", mixing with the fundamental of "B"
>> to produce a third order product 2A - B. Similarly to keep the second
>> harmonic of "B" out of "A". Any mixing of other harmonics can also
>> result in "alien" third order products.
>>
>> If you build a "pretty" version, I would suggest that each generator be
>> built in its own metal box, with all DC supply leads filtered. What you
>> do not want is the generators talking to each other :-)
>>
>> Good luck with your project.
>>
>> 73,
>> Geoff
>> GM4ESD
>>
>>
>> Ramiro Aceves wrote on Thursday, November 25, 2010, at 7:52 AM:
>>
>>> No problem. I cant resist what you propose. I am going to build Xtal
>>> oscillator. I have a 4 MHz cristal here. The the combiner.
>>> Keep tuned!
>>>
>>> Ramiro
>>
>>
>>
>>
>
> ______________________________________________________________
> 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
>