Thank you for this note Tom. Very informative.
I am working on an SDR/DSP "smart antenna" design now for deployment at
Joel's and I thank him for allowing me to do ALL of this development when I
had no real track record of my own in this band.
I believe we are at place now where we can build a smart array that is self
calibrating, using your original elements (top-hat short verticals) and
combine all 8 is my goal. It can be done reasonably inexpensively on 160m
but 80m is right on the cusp of having to go from inexpensive to much more
expensive.
For those who don't know, Dan McGuire AC6LA, is about to release an
seriously updated version of MultiNEC which he calls AUTOEZ and it works
with late versions of EZNEC 5. I need tools to do the arithmetic since my
intuition and experience is nothing akin to yours.
Bob
HY
On Thu, Feb 14, 2013 at 8:22 AM, Tom W8JI <w8ji@w8ji.com> wrote:
> Thanks for the insightful reply. Looks like there is no free lunch here- I
>> can't get better performance by simply using a different circle array
>> element. Bummer, I will have to stick with the verticals.
>>
>
> Not with a simple element.
>
>
> If you have a significant null performance shortfall in some direction,
> which could be caused by an array error or abnormal noise levels in a
> certain direction, directive cells for elements can help provided they
> force a null where you need the additional help. But this is a lot of work
> and hardware if you want to maintain all directions, and you would have to
> be careful to add nulls where you need them. As K9DX John said:
>
> There would be little advantage with the loops because once the side/rear
>>> response is 20 or so db down, the >>RDF is determined by the width of the
>>> forward lobe.
>>> John K9DX
>>>
>>
> Which I often say the same thing in a different way....
>
> Antennas obtain directive gain by creating nulls. They do this by pattern
> multiplication of the directive cells into an array. We can't add
> directivity by forcing a null unless it is in a direction of significant
> radiation. If there is nothing to remove from an area, attempting to remove
> something more won't help.
>
> On receiving we talk about directivity, on transmitting gain. They are
> different because directivity does not include efficiency. This is why two
> close-spaced broadside lossy antennas, like Beverages, can have 3 dB gain
> at almost any spacing without any significant change in pattern or
> directivity. The same applies to other arrays, so for receiving we should
> always look at directivity change and not gain change.
>
>
> If space is not an issue, will enlarging the circle diameter make any
>> difference? Note, however, that I am considering the 3-band commercial
>> solutions (160, 80, and 40m 8 circles).
>>
>>
> The diameter limitation is where the circumference causes end-fire spacing
> to be around 1/4 wavelength. There are some complicated tricks than can
> obtain another ~1 dB directivity where more elements are used inside that
> circle, but they make phase and level distribution much more complex and
> critical. Because of that, the small change rarely materializes in the real
> world and can easily go negative.
>
> When the circle gets larger than that diameter limit, then the obvious
> next step is more elements inside the circle. The problem is this drives
> complexity up.
>
> I experimented here with very large arrays years ago by watching signals
> and watching phase, and I found skywave 160 signals commonly have phase and
> level variations between arrays centered about 1-2 wavelengths apart. This
> is at the root of the very reason we use spatial diversity, and why spatial
> diversity works. It stands to reason if phase and level were stable,
> spatial diversity would not work.
>
> I could not combine very large arrays in fixed phase with the stable,
> predictable, results I hoped to have. I decided the most reliable system
> was to combine in stereo and use my brain to ignore the slow phase errors,
> although it may not be beyond DSP technology now.
>
> I used the Drake R4C, because it had access to external oscillator ports.
> This easily allowed phase locking of both channels. The K3 is the only
> receiver I'm aware of that comes close to doing this in an ideal fashion.
> With a locked oscillator the signals could be summed in audio by shifting
> phase to correct errors, and the results would be identical to shifting
> phase at radio frequencies. But this requires locked oscillators and
> identical channels, something most receiver designers miss or do not
> understand.
>
> The K3 is less than ideal, but much closer to ideal than anything else.
>
>
> 73 Tom
> _________________
> Topband Reflector
>
--
Bob McGwier
Owner and Technical Director, Allied Communication, LLC
Facebook: N4HYBob
ARS: N4HY
_________________
Topband Reflector
|