Elecraft › [K3]

[K3] Diversity Reception and Antenna Directivity

Classic

List

Threaded

20 messages Options

David Gilbert

[K3] Diversity Reception and Antenna Directivity

Several weeks ago, I believe that Wayne posted a message asking what
kind of different uses people were coming up with for their K3. One
thing I've been playing with lately is feeding the signals from two
horizontally polarized antennas at different heights on my tower into
the Main and Sub receivers of my K3 in diversity mode. Since the
relative phase between the two signals is preserved in the translation
to audio, I can feed the audio from the two receivers into the A and B
channels of my computer sound card and compare the relative phase using
a dual-trace sound card oscilloscope program like Zelscope. By knowing
the vertical distance between the two antennas I'm hoping to be able to
calculate the arrival angle of the signal in real time. I say "hoping
to" because so far I don't have a distant stable, unmodulated carrier to
work with ... the best DX carriers have come from 40m BC stations but
the modulation screws up the triggering. Once I get the methodology
worked out a bit better I'll ask someone in Europe to throw a carrier on
frequency for me.

Playing around with this stuff got me thinking, though. What if I fed
the output from two VERTICAL antennas into the K3 receivers in diversity
mode, fed the audio output of both receivers into the A and B channels
of the computer sound card, and used an application that introduced an
adjustable delay in one audio channel before summing the two channels
and doing the D/A translation back to monaural audio? Wouldn't that
have the exact same effect as being able to adjust the phase of the
incoming RF, and therefore the directivity of the 2 element vertical
array? I'm pretty sure that today's computers could certainly handle
the computation. There wouldn't be any constraints on the amount of
delay so the array should be continuously steerable through an entire
360 degrees, and since the delay would be imposed digitally there
wouldn't be any frequency dependency. Ideally the two feedlines would
be of equal construction and equal length, but even if they weren't it
would be fairly easy to characterize their relative phase delay as a
function of frequency. I think mutual coupling even become a non-issue
if the verticals are non-resonant. Non-resonant antennas might be the
way to go anyway since such unlimited control over phase means that
spacing between them would be less of an issue, and therefore the same
pair of verticals could be used on more than one band as long as the
spacing was wide enough.

Why wouldn't this work? The PCM data format is pretty straightforward
and I can't believe that the application would be that complicated to
write. I must be missing something but nothing jumps out at me. If it
worked, it could even be a feature in a next generation K3 (maybe even
the current one) .... all it would take is some means to adjust the
delay since everything else (two phase locked receivers, DSP processing
for both RF and audio) is already there.

In the case of the K3, all of this would only apply to reception, of
course, although it almost seems like a transceiver could be engineered
that used the desired delay determined from the receiver to set a
corresponding delay for two identical tone-modulated transmitter chains
driven from the same oscillator. I suspect a pair of phased
transmitters would have pretty limited appeal, though ... certainly
they'd be an expensive way to get just two or three db steerable gain.

Fun stuff to think about, in any case.

73,
Dave AB7E

______________________________________________________________
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

Jack Smith-6

Re: [K3] Diversity Reception and Antenna Directivity

Spectrum Laboratory by DL4YHF http://www.qsl.net/dl4yhf/spectra1.html
may be able to do this already.

If the K3's 15 KHz IF output were available from the main and sub
receivers, Spectrum Laboratory could use those to generate the automatic
phase display using the direction finding feature. Don't know how it
would work with demodulated audio and 0 Hz IF.

It may also be possible to use the DSP black box functionality in
Spectrum Laboratory to introduce variable delay into the audio chain.

Jack K8ZOA

On 9/16/2010 4:37 AM, David Gilbert wrote:

> Several weeks ago, I believe that Wayne posted a message asking what
> kind of different uses people were coming up with for their K3. One
> thing I've been playing with lately is feeding the signals from two
> horizontally polarized antennas at different heights on my tower into
> the Main and Sub receivers of my K3 in diversity mode. Since the
> relative phase between the two signals is preserved in the translation
> to audio, I can feed the audio from the two receivers into the A and B
> channels of my computer sound card and compare the relative phase using
> a dual-trace sound card oscilloscope program like Zelscope. By knowing
> the vertical distance between the two antennas I'm hoping to be able to
> calculate the arrival angle of the signal in real time. I say "hoping
> to" because so far I don't have a distant stable, unmodulated carrier to
> work with ... the best DX carriers have come from 40m BC stations but
> the modulation screws up the triggering. Once I get the methodology
> worked out a bit better I'll ask someone in Europe to throw a carrier on
> frequency for me.
>
> Playing around with this stuff got me thinking, though. What if I fed
> the output from two VERTICAL antennas into the K3 receivers in diversity
> mode, fed the audio output of both receivers into the A and B channels
> of the computer sound card, and used an application that introduced an
> adjustable delay in one audio channel before summing the two channels
> and doing the D/A translation back to monaural audio? Wouldn't that
> have the exact same effect as being able to adjust the phase of the
> incoming RF, and therefore the directivity of the 2 element vertical
> array? I'm pretty sure that today's computers could certainly handle
> the computation. There wouldn't be any constraints on the amount of
> delay so the array should be continuously steerable through an entire
> 360 degrees, and since the delay would be imposed digitally there
> wouldn't be any frequency dependency. Ideally the two feedlines would
> be of equal construction and equal length, but even if they weren't it
> would be fairly easy to characterize their relative phase delay as a
> function of frequency. I think mutual coupling even become a non-issue
> if the verticals are non-resonant. Non-resonant antennas might be the
> way to go anyway since such unlimited control over phase means that
> spacing between them would be less of an issue, and therefore the same
> pair of verticals could be used on more than one band as long as the
> spacing was wide enough.
>
> Why wouldn't this work? The PCM data format is pretty straightforward
> and I can't believe that the application would be that complicated to
> write. I must be missing something but nothing jumps out at me. If it
> worked, it could even be a feature in a next generation K3 (maybe even
> the current one) .... all it would take is some means to adjust the
> delay since everything else (two phase locked receivers, DSP processing
> for both RF and audio) is already there.
>
> In the case of the K3, all of this would only apply to reception, of
> course, although it almost seems like a transceiver could be engineered
> that used the desired delay determined from the receiver to set a
> corresponding delay for two identical tone-modulated transmitter chains
> driven from the same oscillator. I suspect a pair of phased
> transmitters would have pretty limited appeal, though ... certainly
> they'd be an expensive way to get just two or three db steerable gain.
>
> Fun stuff to think about, in any case.
>
> 73,
> Dave AB7E
>
>
> ______________________________________________________________
> 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
>

______________________________________________________________
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

Alan Bloom

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by David Gilbert

> What if I fed
> the output from two VERTICAL antennas into the K3 receivers in diversity
> mode, fed the audio output of both receivers into the A and B channels
> of the computer sound card, and used an application that introduced an
> adjustable delay in one audio channel before summing the two channels
> and doing the D/A translation back to monaural audio? Wouldn't that
> have the exact same effect as being able to adjust the phase of the
> incoming RF, and therefore the directivity of the 2 element vertical
> array?

Yes, I think that would work fine. There are two issues that I can
think of:

While the main and sub receivers are phase-coherent, the actual phase
difference between them is random. I believe that if you change
frequency it is not guaranteed that the phase difference will be the
same. So every time you change frequency you may have to re-adjust the
phase delay to get the antenna to "point" in the right direction.

The other issue is that the antennas must be no more than 1/2 wavelength
apart to get a clean, single-lobe response in the directivity pattern.

Al N1AL

On Thu, 2010-09-16 at 01:37 -0700, David Gilbert wrote:

Bill W4ZV

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by David Gilbert

David Gilbert wrote

Playing around with this stuff got me thinking, though. What if I fed
the output from two VERTICAL antennas into the K3 receivers in diversity
mode, fed the audio output of both receivers into the A and B channels
of the computer sound card, and used an application that introduced an
adjustable delay in one audio channel before summing the two channels
and doing the D/A translation back to monaural audio? Wouldn't that
have the exact same effect as being able to adjust the phase of the
incoming RF, and therefore the directivity of the 2 element vertical
array?

Good idea, already implemented by MFJ, DXE and others:

http://www.mfjenterprises.com/Product.php?productid=MFJ-1025
http://www.dxengineering.com/Parts.asp?ID=1227&PLID=215&SecID=114&DeptID=12&PartNo=DXE-NCC-1

This reminds me of one SDR guru who proposed using 4 SDR transceivers to emulate a 4-square. A multi-K$ solution to a $350 problem (i.e. one 4SQ controller box)...not to mention the need for 4 separate feedlines, 4 amplifiers, etc.

73, Bill

P.S. I believe one problem with what you proposed above is that phase would need to be readjusted every time you touched the VFO...i.e. phase is locked (i.e. meaning the phase differential is fixed) but the actual phase offset between the two RXs changes each time the synthesizer changes. Lyle can correct me if I'm wrong.

Jim Brown-10

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Alan Bloom

There is another VERY important thing that is being missed by this
discussion. A VERY large component of fading is due to multipath -- that is,
the arrival of more two or more wavefronts that travel different paths, and
thus have slightly different travel times. This produces a phase shift which
varies with frequency, position, and the path. When the two wavefronts are
precisely in phase and equal in amplitude, they add by 6dB; when they are
precisely 180 degrees out of phase and equal in amplitude, they cancel. At
any other value of phase difference, there is partial addition or partial
cancellation. And, of course, the more nearly equal the two wavefronts, the
deeper the cancellation.

This is WELL recognized as multipath on VHF and UHF, but few hams realize
that the same thing is happening on the lower bands. Think about it -- the
phase shift is a direct function of frequency, so the very fast "picket
fencing" at VHF/UHF of a mobile station or the flutter of a signal reflected
by a moving aircraft becomes MUCH slower fading on the lower bands. A fading
period on the order of a minute or two is not uncommon on 160M.

When there is multipath on HF and MF circuits, the paths (most?) often differ
by their vertical arrival angle. This suggests that one might expect a small
time offset between horizontal antennas at different heights based on their
vertical patterns and the vertical arrival angle of multiple wavefronts
favored by the directivity of one antenna or the other.

73, Jim Brown K9YC

______________________________________________________________
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 Gilbert

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Alan Bloom

Thanks for the reply, Alan.

That's interesting. I didn't realize that the two receivers in the K3
had a random phase difference between them even when locked. I noticed
differences in phase delay when I changed frequencies during my tests
with the two horizontal antennas on the tower, but I mostly attributed
that to different arrival angles for the different stations being
monitored, and different phase delay of the feed lines (expressed in
wavelengths) at the new frequencies.

However, that random phase difference between the two receivers could be
adjusted out by momentarily feeding the same antenna into both receivers
at each new frequency. That routing could be accomplished within the K3
if there was a quick and easy way to control whether or not the sub
receiver was switched to the AUX RF antenna when diversity mode is
active. If that routing option took the form of a command, the whole
process could be done in software (either external or internal to the
K3) each time the frequency was changed ... albeit of course with some
settling delay while the phase difference was determined and adjusted out.

73,
Dave AB7E

On 9/16/2010 8:42 AM, Alan Bloom wrote:
> Yes, I think that would work fine. There are two issues that I can
> think of: While the main and sub receivers are phase-coherent, the
> actual phase difference between them is random. I believe that if you
> change frequency it is not guaranteed that the phase difference will
> be the same. So every time you change frequency you may have to
> re-adjust the phase delay to get the antenna to "point" in the right
> direction.

______________________________________________________________
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

Alan Bloom-2

Re: [K3] Diversity Reception and Antenna Directivity

Yes, that should work, just switch the sub receiver antenna over to the
main receiver to calibrate the phase. I think you'd get the best
accuracy by looking for a null and then subtracting 180 degrees.

By the way, I think rather than an adjustable delay between the two
channels you need an adjustable phase. (If the phases of the two local
oscillators in the K3 differ by X degrees, than all audio frequencies
also differ by X degrees.) There are a number of ways to do that, but
probably the most straightforward is to re-convert each audio signal to
"RF" (a few kHz) using local oscillators of the same frequency but
different phases, and then convert back to baseband with a single LO.
That could be done either with hardware or in software.

Just a SMOP. (Small matter of programming :=)

Alan N1AL

On Thu, 2010-09-16 at 10:08 -0700, David Gilbert wrote:

> Thanks for the reply, Alan.
>
> That's interesting. I didn't realize that the two receivers in the K3
> had a random phase difference between them even when locked. I noticed
> differences in phase delay when I changed frequencies during my tests
> with the two horizontal antennas on the tower, but I mostly attributed
> that to different arrival angles for the different stations being
> monitored, and different phase delay of the feed lines (expressed in
> wavelengths) at the new frequencies.
>
> However, that random phase difference between the two receivers could be
> adjusted out by momentarily feeding the same antenna into both receivers
> at each new frequency. That routing could be accomplished within the K3
> if there was a quick and easy way to control whether or not the sub
> receiver was switched to the AUX RF antenna when diversity mode is
> active. If that routing option took the form of a command, the whole
> process could be done in software (either external or internal to the
> K3) each time the frequency was changed ... albeit of course with some
> settling delay while the phase difference was determined and adjusted out.
>
> 73,
> Dave AB7E
>
>
>
>
> On 9/16/2010 8:42 AM, Alan Bloom wrote:
> > Yes, I think that would work fine. There are two issues that I can
> > think of: While the main and sub receivers are phase-coherent, the
> > actual phase difference between them is random. I believe that if you
> > change frequency it is not guaranteed that the phase difference will
> > be the same. So every time you change frequency you may have to
> > re-adjust the phase delay to get the antenna to "point" in the right
> > direction.
>
> ______________________________________________________________
> 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

Kok Chen

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Jim Brown-10

On Sep 16, 2010, at 9/16 9:59 AM, Jim Brown wrote:

> A VERY large component of fading is due to multipath -- that is,
> the arrival of more two or more wavefronts that travel different
> paths, and
> thus have slightly different travel times.

Selective fading does not require multipath. The CCIR 520-2 profiles
for Raleigh fading are all single path models. Raleigh fading causes
selective fading.

http://www.itu.int/rec/R-REC-F.520-2-199203-W/en

Whenever you hear distorted AM signals on HF, chances are it is caused
by selective fading. You can see selective fading in a waterfall by
watching the fading occur as moving holes that sweep across the
spectrum (very visible when you tune in broadband signal such as a
Coast Guard weather FAX station on HF). You can also see it take away
individual tones in an Olivia signal in a waterfall. Selective fading
was also one of the primary impetus to switch from on-off keying to
FSK in the early days of RTTY.

The Watterson model for ionospheric propagation breaks up a path into
a complex signal with in-phase and quadrature components:

http://ieeexplore.ieee.org:80/xpl/freeabs_all.jsp?reload=true&arnumber=1090438

Each component passes through a independent scattering function that
have Gaussian random processes, both a Gaussian Doppler spreading term
and also a Gaussian amplitude term.

The modulus (i.e. "amplitude", or square root of power of the I and Q
components) of a bivariate Gaussian random process happens to have
Rayleigh distribution. See references here (the Rician distribution
is more general in that the mean of the components need not be zero):

http://en.wikipedia.org/wiki/Rayleigh_distribution
http://en.wikipedia.org/wiki/Rice_distribution

I.e., the amplitude of a signal whose I and Q components each have
independent Gaussian statistics, has a Rayleigh distribution.

As a consequence, a Rayleigh signal can cause selective fading without
the need for a second path. Since the Rayleigh probability density
function has finite probability of being infinitesimally small, the
fade has a chance of being very deep.

If you run a signal through an HF Channel Simulator (such as AE4JY's
PathSim or cocoaPath) set to rayleigh fading parameters, you will see
selective fading.

Watterson's paper also considers both the cases of multi- paths and
multi- magneto-ionic rays that are scattered by the ionosphere.

Multipath signals have a time delay, multiray signals do not have a
time delay between the rays.

It is a fascinating paper that hams interested in HF propagation
should read. Unfortunately, I have not found a free version on the
web that I can reference, even though the research was done using tax
payer's money at what is today NIST. But if you are an IEEE member,
you can download the paper for free.

73
Chen, W7AY

______________________________________________________________
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 Gilbert

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Bill W4ZV

Hi, Bill.

All true, but I thought that implementing it at audio might have some
advantages ... assuming at least that someone already had the sub receiver:

a. Cheaper. The DXE unit is $500 ... considerably more than that if
you buy the active antennas as well.
b. More versatile. Since the audio can be easily digitized using the
sound card, almost anything could be done with the result via software.
c. No need for separate receive and transmit antennas, or to maintain a
large spacing between receive and transmit antennas.

Alan has verified your point about relative phase of the two receivers
changing with frequency, so that indeed would have to be accounted for.
It's also true that something like a 4-Square would have a major
advantage by offering gain on transmit as well as pattern on receive.

In any case, I only suggested all of this for possible exploration. I
don't have anything at stake here if it turns out to be a dumb idea.

73,
Dave AB7E

On 9/16/2010 9:49 AM, Bill W4ZV wrote:

>
> David Gilbert wrote:
>>
>> Playing around with this stuff got me thinking, though. What if I fed
>> the output from two VERTICAL antennas into the K3 receivers in diversity
>> mode, fed the audio output of both receivers into the A and B channels
>> of the computer sound card, and used an application that introduced an
>> adjustable delay in one audio channel before summing the two channels
>> and doing the D/A translation back to monaural audio? Wouldn't that
>> have the exact same effect as being able to adjust the phase of the
>> incoming RF, and therefore the directivity of the 2 element vertical
>> array?
>>
> Good idea, already implemented by MFJ, DXE and others:
>
> http://www.mfjenterprises.com/Product.php?productid=MFJ-1025
> http://www.dxengineering.com/Parts.asp?ID=1227&PLID=215&SecID=114&DeptID=12&PartNo=DXE-NCC-1
>
> This reminds me of one SDR guru who proposed using 4 SDR transceivers to
> emulate a 4-square. A multi-K$ solution to a $350 problem (i.e. one 4SQ
> controller box)...not to mention the need for 4 separate feedlines, 4
> amplifiers, etc.
>
> 73, Bill
>
> P.S. I believe one problem with what you proposed above is that phase would
> need to be readjusted every time you touched the VFO...i.e. phase is locked
> (i.e. meaning the phase differential is fixed) but the actual phase offset
> between the two RXs changes each time the synthesizer changes. Lyle can
> correct me if I'm wrong.

Jack Smith-6

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Kok Chen

If you plot the fade statistics for HF signals, however, you find many
times that the result looks much more like Rician fading than Rayleigh.

Jack K8ZOA

On 9/16/2010 2:11 PM, Kok Chen wrote:

> On Sep 16, 2010, at 9/16 9:59 AM, Jim Brown wrote:
>
>> A VERY large component of fading is due to multipath -- that is,
>> the arrival of more two or more wavefronts that travel different
>> paths, and
>> thus have slightly different travel times.
> Selective fading does not require multipath. The CCIR 520-2 profiles
> for Raleigh fading are all single path models. Raleigh fading causes
> selective fading.
>
> http://www.itu.int/rec/R-REC-F.520-2-199203-W/en
>
> Whenever you hear distorted AM signals on HF, chances are it is caused
> by selective fading. You can see selective fading in a waterfall by
> watching the fading occur as moving holes that sweep across the
> spectrum (very visible when you tune in broadband signal such as a
> Coast Guard weather FAX station on HF). You can also see it take away
> individual tones in an Olivia signal in a waterfall. Selective fading
> was also one of the primary impetus to switch from on-off keying to
> FSK in the early days of RTTY.
>
> The Watterson model for ionospheric propagation breaks up a path into
> a complex signal with in-phase and quadrature components:
>
> http://ieeexplore.ieee.org:80/xpl/freeabs_all.jsp?reload=true&arnumber=1090438
>
> Each component passes through a independent scattering function that
> have Gaussian random processes, both a Gaussian Doppler spreading term
> and also a Gaussian amplitude term.
>
> The modulus (i.e. "amplitude", or square root of power of the I and Q
> components) of a bivariate Gaussian random process happens to have
> Rayleigh distribution. See references here (the Rician distribution
> is more general in that the mean of the components need not be zero):
>
> http://en.wikipedia.org/wiki/Rayleigh_distribution
> http://en.wikipedia.org/wiki/Rice_distribution
>
> I.e., the amplitude of a signal whose I and Q components each have
> independent Gaussian statistics, has a Rayleigh distribution.
>
> As a consequence, a Rayleigh signal can cause selective fading without
> the need for a second path. Since the Rayleigh probability density
> function has finite probability of being infinitesimally small, the
> fade has a chance of being very deep.
>
> If you run a signal through an HF Channel Simulator (such as AE4JY's
> PathSim or cocoaPath) set to rayleigh fading parameters, you will see
> selective fading.
>
> Watterson's paper also considers both the cases of multi- paths and
> multi- magneto-ionic rays that are scattered by the ionosphere.
>
> Multipath signals have a time delay, multiray signals do not have a
> time delay between the rays.
>
> It is a fascinating paper that hams interested in HF propagation
> should read. Unfortunately, I have not found a free version on the
> web that I can reference, even though the research was done using tax
> payer's money at what is today NIST. But if you are an IEEE member,
> you can download the paper for free.
>
> 73
> Chen, W7AY
>
>
>
>
>
> ______________________________________________________________
> 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 Gilbert

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Alan Bloom-2

Well, I was thinking mostly in terms of CW so I think delay and phase in
the context of a single frequency are essentially the same. I did give
a bit of thought to what it would take to make it work for SSB, though,
and while I am the last person on earth you'd ever want to consult on
math, it seemed to me that maybe adding the shift in the frequency
domain would work. I.e., perform an FFT, add the shift, and then
convert back to the time domain. Does that have the correct effect?
I'm just kind of guessing here.

LMOP (larger matter of programming ... but less hardware)

Thanks for humoring me on the discussion!

73,
Dave AB7E

On 9/16/2010 10:47 AM, Alan Bloom wrote:

> Yes, that should work, just switch the sub receiver antenna over to the
> main receiver to calibrate the phase. I think you'd get the best
> accuracy by looking for a null and then subtracting 180 degrees.
>
> By the way, I think rather than an adjustable delay between the two
> channels you need an adjustable phase. (If the phases of the two local
> oscillators in the K3 differ by X degrees, than all audio frequencies
> also differ by X degrees.) There are a number of ways to do that, but
> probably the most straightforward is to re-convert each audio signal to
> "RF" (a few kHz) using local oscillators of the same frequency but
> different phases, and then convert back to baseband with a single LO.
> That could be done either with hardware or in software.
>
> Just a SMOP. (Small matter of programming :=)
>
> Alan N1AL
>
>
> On Thu, 2010-09-16 at 10:08 -0700, David Gilbert wrote:
>> Thanks for the reply, Alan.
>>
>> That's interesting. I didn't realize that the two receivers in the K3
>> had a random phase difference between them even when locked. I noticed
>> differences in phase delay when I changed frequencies during my tests
>> with the two horizontal antennas on the tower, but I mostly attributed
>> that to different arrival angles for the different stations being
>> monitored, and different phase delay of the feed lines (expressed in
>> wavelengths) at the new frequencies.
>>
>> However, that random phase difference between the two receivers could be
>> adjusted out by momentarily feeding the same antenna into both receivers
>> at each new frequency. That routing could be accomplished within the K3
>> if there was a quick and easy way to control whether or not the sub
>> receiver was switched to the AUX RF antenna when diversity mode is
>> active. If that routing option took the form of a command, the whole
>> process could be done in software (either external or internal to the
>> K3) each time the frequency was changed ... albeit of course with some
>> settling delay while the phase difference was determined and adjusted out.
>>
>> 73,
>> Dave AB7E
>>
>>
>>
>>
>> On 9/16/2010 8:42 AM, Alan Bloom wrote:
>>> Yes, I think that would work fine. There are two issues that I can
>>> think of: While the main and sub receivers are phase-coherent, the
>>> actual phase difference between them is random. I believe that if you
>>> change frequency it is not guaranteed that the phase difference will
>>> be the same. So every time you change frequency you may have to
>>> re-adjust the phase delay to get the antenna to "point" in the right
>>> direction.
>>

KK7P

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by David Gilbert

> Alan has verified your point about relative phase of the two receivers
> changing with frequency, so that indeed would have to be accounted for.

A possible mod would be to feed the main synth to both receivers. This
would then only be good for diversity operation, but there would be no
random phase change when tuning.

The DSPs do not have this relative phase change issue. The dual A/D is
driven by one clock on the main DSP board. The ADC output is fed to the
Aux DSP, but not clocked by the Aux DSP. Thus, the sampling offset and
any delays are fixed and constant. DSP filter delays are also constant
as long as you do not use the IIR filters for 100 Hz and 50 Hz
bandwidth. I am not certain if the actions of AutoNotch and/or Noise
Reduction would introduce any relative phase changes -- assuming
activation in both receivers, of course!

73,

Lyle KK7P
______________________________________________________________
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 Gilbert

Re: [K3] Diversity Reception and Antenna Directivity

Hi, Lyle.

Thanks for the comment. If nothing else, I'm learning from this discussion.

Actually, until Alan pointed out that they were different, I thought
that the two receivers WERE driven by the same main synth in diversity
mode. If it were possible to have that option when in diversity mode,
that would seem to greatly simplify what I was suggesting.

73,
Dave AB7E

On 9/16/2010 11:35 AM, Lyle Johnson wrote:

>> Alan has verified your point about relative phase of the two receivers
>> changing with frequency, so that indeed would have to be accounted for.
> A possible mod would be to feed the main synth to both receivers. This
> would then only be good for diversity operation, but there would be no
> random phase change when tuning.
>
> The DSPs do not have this relative phase change issue. The dual A/D is
> driven by one clock on the main DSP board. The ADC output is fed to the
> Aux DSP, but not clocked by the Aux DSP. Thus, the sampling offset and
> any delays are fixed and constant. DSP filter delays are also constant
> as long as you do not use the IIR filters for 100 Hz and 50 Hz
> bandwidth. I am not certain if the actions of AutoNotch and/or Noise
> Reduction would introduce any relative phase changes -- assuming
> activation in both receivers, of course!
>
> 73,
>
> Lyle KK7P

Alan Bloom

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by David Gilbert

On Thu, 2010-09-16 at 11:29 -0700, David Gilbert wrote:
>
> Well, I was thinking mostly in terms of CW so I think delay and phase in
> the context of a single frequency are essentially the same. I did give
> a bit of thought to what it would take to make it work for SSB, though,
> and while I am the last person on earth you'd ever want to consult on
> math, it seemed to me that maybe adding the shift in the frequency
> domain would work. I.e., perform an FFT, add the shift, and then
> convert back to the time domain. Does that have the correct effect?
> I'm just kind of guessing here.

Yes, that would be another way to do it. Determine the phase of each
complex FFT frequency point (arctangent of imaginary/real), add the
proper phase, and then do an inverse FFT to get back to the time domain.
To get a continuous signal you have to choose the proper windowing
function and the amount of overlap of the sample sets in the time
domain.

Yet another way to do it is to design an all-pass network with a nearly
constant phase shift over the frequency band of interest. That could be
done with either an FIR or IIR digital filter, but the coefficients
would have to be re-calculated for each desired phase shift.

I'm thinking it would be easier to do the method where you heterodyne to
"RF" using LOs of different phase and then downconvert back to baseband.
That can be done all in software.

Alan

> LMOP (larger matter of programming ... but less hardware)
>
> Thanks for humoring me on the discussion!
>
> 73,
> Dave AB7E
>
>
>
> On 9/16/2010 10:47 AM, Alan Bloom wrote:
> > Yes, that should work, just switch the sub receiver antenna over to the
> > main receiver to calibrate the phase. I think you'd get the best
> > accuracy by looking for a null and then subtracting 180 degrees.
> >
> > By the way, I think rather than an adjustable delay between the two
> > channels you need an adjustable phase. (If the phases of the two local
> > oscillators in the K3 differ by X degrees, than all audio frequencies
> > also differ by X degrees.) There are a number of ways to do that, but
> > probably the most straightforward is to re-convert each audio signal to
> > "RF" (a few kHz) using local oscillators of the same frequency but
> > different phases, and then convert back to baseband with a single LO.
> > That could be done either with hardware or in software.
> >
> > Just a SMOP. (Small matter of programming :=)
> >
> > Alan N1AL
> >
> >
> > On Thu, 2010-09-16 at 10:08 -0700, David Gilbert wrote:
> >> Thanks for the reply, Alan.
> >>
> >> That's interesting. I didn't realize that the two receivers in the K3
> >> had a random phase difference between them even when locked. I noticed
> >> differences in phase delay when I changed frequencies during my tests
> >> with the two horizontal antennas on the tower, but I mostly attributed
> >> that to different arrival angles for the different stations being
> >> monitored, and different phase delay of the feed lines (expressed in
> >> wavelengths) at the new frequencies.
> >>
> >> However, that random phase difference between the two receivers could be
> >> adjusted out by momentarily feeding the same antenna into both receivers
> >> at each new frequency. That routing could be accomplished within the K3
> >> if there was a quick and easy way to control whether or not the sub
> >> receiver was switched to the AUX RF antenna when diversity mode is
> >> active. If that routing option took the form of a command, the whole
> >> process could be done in software (either external or internal to the
> >> K3) each time the frequency was changed ... albeit of course with some
> >> settling delay while the phase difference was determined and adjusted out.
> >>
> >> 73,
> >> Dave AB7E
> >>
> >>
> >>
> >>
> >> On 9/16/2010 8:42 AM, Alan Bloom wrote:
> >>> Yes, I think that would work fine. There are two issues that I can
> >>> think of: While the main and sub receivers are phase-coherent, the
> >>> actual phase difference between them is random. I believe that if you
> >>> change frequency it is not guaranteed that the phase difference will
> >>> be the same. So every time you change frequency you may have to
> >>> re-adjust the phase delay to get the antenna to "point" in the right
> >>> direction.
> >>
> ______________________________________________________________
> 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

Kok Chen

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Kok Chen

I just rummaged around in Google and found a different paper that
Watterson wrote related to HF propagation and digital modes (almost 10
years after his HF channel model paper):

http://www.its.bldrdoc.gov/pub/ntia-rpt/79-29/79-29.pdf

His conclusion definitely advocates the use of diversity reception.

Discussion of HF Channel model starts on page 35 (yes, gigantic
manuscript), discussion on using separate receiving antennas on page
46, use of matched filters on page 51, diversity reception on page 72.

73
Chen, W7AY

______________________________________________________________
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

Alan Bloom

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Alan Bloom

Hi Dave,

For some reason I got several copies of your message. I'll go ahead and
post my reply to the reflector since it may be of wider interest.

Yes, another way to get a constant phase shift at all frequencies is to
convert to an I/Q (in-phase/quadrature) signal. One way to do that is
to take the audio signal and run it through an all-pass network with a
constant 90-degree phase shift at all frequencies. That is then the "Q"
signal and the original non-shifted signal is the "I". You can then
obtain any desired phase shift by adding the two signals together with
the proper weighting factor for each. For example, if you weight I and
Q by the same factor you get 45 degrees.

But that still requires a wide-band 90-degree phase shift network. It
can be done with a FIR or IIR digital filter. I think there are some
free filter design programs on the web that can design an all-pass
network. It may need to be a big filter (i.e. lots of coefficients) to
get good amplitude and phase accuracy over a 10:1 frequency range (300
Hz to 3 kHz).

The other way to get the I/Q signals is with an I/Q modulator.
Basically you run the baseband audio signal into two mixers whose local
oscillators are 90 degrees out of phase. Now you have two RF signals 90
degrees out of phase. If you convert them back to baseband with the
same oscillator you get two audio signals 90 degrees out of phase and
can combine them as before. You then combine that signal with the
original to simulate rotation of the directional antenna.

I still think the easiest method is what I first suggested. Like the
I/Q modulator except that, instead of a 90-degree phase shift, you
adjust the relative phase of the two oscillators. When you convert back
to baseband the two signals are already at the correct phase, ready to
be combined to simulate the directional antenna.

Alan

______________________________________________________________
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 Gilbert

Re: [K3] Diversity Reception and Antenna Directivity

That sounds very good, Alan. As I say, I'm not very knowledgeable on
signal processing so all of this is very interesting to me. I guess
whether any of it ultimately turns out to be useful to anyone remains to
be seen.

Regarding the multiple posts ... my ISP has been acting very flaky this
afternoon and it kept telling me that the message upload had failed.
Not sure what's going on there, but I'm sorry for the extra clutter.

Thanks again es 73,
Dave AB7E

On 9/16/2010 1:55 PM, Alan Bloom wrote:

> Hi Dave,
>
> For some reason I got several copies of your message. I'll go ahead and
> post my reply to the reflector since it may be of wider interest.
>
> Yes, another way to get a constant phase shift at all frequencies is to
> convert to an I/Q (in-phase/quadrature) signal. One way to do that is
> to take the audio signal and run it through an all-pass network with a
> constant 90-degree phase shift at all frequencies. That is then the "Q"
> signal and the original non-shifted signal is the "I". You can then
> obtain any desired phase shift by adding the two signals together with
> the proper weighting factor for each. For example, if you weight I and
> Q by the same factor you get 45 degrees.
>
> But that still requires a wide-band 90-degree phase shift network. It
> can be done with a FIR or IIR digital filter. I think there are some
> free filter design programs on the web that can design an all-pass
> network. It may need to be a big filter (i.e. lots of coefficients) to
> get good amplitude and phase accuracy over a 10:1 frequency range (300
> Hz to 3 kHz).
>
> The other way to get the I/Q signals is with an I/Q modulator.
> Basically you run the baseband audio signal into two mixers whose local
> oscillators are 90 degrees out of phase. Now you have two RF signals 90
> degrees out of phase. If you convert them back to baseband with the
> same oscillator you get two audio signals 90 degrees out of phase and
> can combine them as before. You then combine that signal with the
> original to simulate rotation of the directional antenna.
>
> I still think the easiest method is what I first suggested. Like the
> I/Q modulator except that, instead of a 90-degree phase shift, you
> adjust the relative phase of the two oscillators. When you convert back
> to baseband the two signals are already at the correct phase, ready to
> be combined to simulate the directional antenna.
>
> Alan
>
>
>

Kok Chen

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Alan Bloom

On Sep 16, 2010, at 1:55 PM, Alan Bloom wrote:

> I think there are some free filter design programs on the web that can design an all-pass network. It may need to be a big filter (i.e. lots of coefficients) to get good amplitude and phase accuracy over a 10:1 frequency range (300 Hz to 3 kHz).

All-pass 90 degree networks (Hilbert Transforms) are really not that tough to implement anymore with the speed of today's computers.

Just a couple of weeks ago, I did an implementation of cocoaPath's Hilbert transformer using Grand Central Dispatch to see how many cores I can use up concurrently. I ended up with a Hilbert transformer for 3 kHz passband (100 Hz to 3.1 kHz) with 16000 samples/second that ran at 300x real time on an 8 core Intel processor. Source code is free if anyone is interested.

Unlike the good old analog days (you can find an analog design in Paul W1HFA's "Art of Electronics" book) you can get much more accurate quadratures using digital Hilbert transforms. The one is cocoaPath is good to 2.5 milli-degrees (yes, 2.5 thousandths of a degree deviation) over a span of 3 kHz. You can see the plots of the phase accuracy in Figures 3-2 and 3-3 here

http://homepage.mac.com/chen/w7ay/cocoaPath/Contents/technical.html

They are not a very long FIRs either -- two 511 tap filters, one to create the in-phase signal and one to create the quadrature signal.

If the passband that you need is much narrower than the sampling rate, it is probably cheaper to just remodulate using sine/cosine "local oscillators" followed by low pass decimation filters (like what the SoftRock analog hardware does, but more accurately in the digital world :-).

73
Chen, W7AY

______________________________________________________________
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

Jim Brown-10

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by David Gilbert

On Thu, 16 Sep 2010 12:06:23 -0700, David Gilbert wrote:

>I thought
>that the two receivers WERE driven by the same main synth in diversity
>mode. If it were possible to have that option when in diversity mode,
>that would seem to greatly simplify what I was suggesting.

YES!

Jim K9YC

______________________________________________________________
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 Woolley (E.L)

Re: [K3] Diversity Reception and Antenna Directivity

In reply to this post by Alan Bloom

Of course the K3 has already done most of this, and then thrown Q away
at the last minute. The upconvert is, of course, done by the transmitter.

Alan Bloom wrote:

> The other way to get the I/Q signals is with an I/Q modulator.
> Basically you run the baseband audio signal into two mixers whose local
> oscillators are 90 degrees out of phase. Now you have two RF signals 90
> degrees out of phase. If you convert them back to baseband with the
> same oscillator you get two audio signals 90 degrees out of phase and
> can combine them as before. You then combine that signal with the
> original to simulate rotation of the directional antenna.
>

--
David Woolley
"we do not overly restrict the subject matter on the list, and we
encourage postings on a wide range of amateur radio related topics"
List Guidelines <http://www.elecraft.com/elecraft_list_guidelines.htm>
______________________________________________________________
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