Tweet
Re: Waller flag antenna
In attesa di tue notizie , posto un po' di cose interessanti, uno scambio di idee fra Tom, w8ji e N4IS, JC.
Magari segui già e sei nella mail list di top band reflector, ma credo possa essere utile a qualcuno questo scambio di impressioni, oltretutto discutono proprio di un problema, che è alla base della Waller Flag, il preamplificatore, si parla anche del preamplificatore di kd9sv distribuito da dx enginnering.
Ho estrapolato alcuni messaggi, la discussione andrebbe letta tutta compreso gli interventi di K9YC e G8LUG. Ma veniva lunga...
Dentro un messaggio di JC ho trovato un link interessante sulle prove della Waller Flag, che ancora non avevo letto:
http://www.qsl.net/k4fk/presentation...nas-SFDXAs.pdf
Spero di essere stato utile.
Saluti
In attesa di tue notizie , posto un po' di cose interessanti, uno scambio di idee fra Tom, w8ji e N4IS, JC.
Magari segui già e sei nella mail list di top band reflector, ma credo possa essere utile a qualcuno questo scambio di impressioni, oltretutto discutono proprio di un problema, che è alla base della Waller Flag, il preamplificatore, si parla anche del preamplificatore di kd9sv distribuito da dx enginnering.
From: "Tom W8JI" <w8ji@w8ji.com>
To: <topband@contesting.com>
Subject: Re: Topband: Waller Flag Question
Thank you, JC:
> I don't know where this coming from ,but the gain you need for a VWF
> modest
> size is 20db for vertical polarization and for horizontal HWF you need
> 40db
> on 160m, on 80m divide this by 2, you need only 20 dB and on 40m 10 dB a
> NORTON preamp is enough. All situations you need a band pass filter.
>
20 dB is a realistic gain figure.
The very low sensitivity of horizontal polarization, because at low heights
in wavelength the earth "shorts the electric field", and because at low
heights the earth's reflection nulls the antenna peak response, causes great
difficulty.
If local site noise is high, and if care is taken in balance, the horizontal
system can be built but 40 dB gain is unlikely to be needed unless the
receiver is dead.
The reason is pretty simple. Most receivers are in the minus 130-140 dBm
noise floor range. If you added 40 dB gain to that, the noise figure of the
required front end would be an impossible negative noise figure in the -20
dB or more noise figure range. Of course anything less than 1 dB is very
difficult, and below 1/2 dB starts to be impossible. Even if you obtain that
noise figure, cable leakages and common mode would overwhelm the low antenna
level.
20 dB is about the limit for most receivers, although a dead receiver could
use 40. If the receiver is stone deaf, 40 dB would allow a workable noise
figure at the front end. :-)
This low sensitivity is why K6STI's antenna met with such limited reports of
success. If the site is very noisy with local distant noise, then the
antenna's noise floor is high enough to limit system noise floor. Otherwise,
the cables and input amplifier would set noise floor.
I have a similar thing here with a commercial loop antenna. Even though
vertically polarized, it is noise limited at my location by internal
amplifier noise. Now if I move it into a noisy location, it limits by
outside noise.
No matter what we try to do, we are not going to have a 0 dB noise figure.
When we start making the antenna sensitivity so low it requires gain with a
normal receiver so unrealistic that it limits on the front end noise, it is
useless. What good would seeing the S meter at S-2 or S-4 from amplifier
noise do? That is what the popular commercial loop I have does. In a quiet
location, it limits on its own internal amplifier noise. Six dB less gain
does not change S/N ratio one bit.
We should all question systems that need 40dB with normal receivers. 20 dB
is more rational.
73 Tom
To: <topband@contesting.com>
Subject: Re: Topband: Waller Flag Question
Thank you, JC:
> I don't know where this coming from ,but the gain you need for a VWF
> modest
> size is 20db for vertical polarization and for horizontal HWF you need
> 40db
> on 160m, on 80m divide this by 2, you need only 20 dB and on 40m 10 dB a
> NORTON preamp is enough. All situations you need a band pass filter.
>
20 dB is a realistic gain figure.
The very low sensitivity of horizontal polarization, because at low heights
in wavelength the earth "shorts the electric field", and because at low
heights the earth's reflection nulls the antenna peak response, causes great
difficulty.
If local site noise is high, and if care is taken in balance, the horizontal
system can be built but 40 dB gain is unlikely to be needed unless the
receiver is dead.
The reason is pretty simple. Most receivers are in the minus 130-140 dBm
noise floor range. If you added 40 dB gain to that, the noise figure of the
required front end would be an impossible negative noise figure in the -20
dB or more noise figure range. Of course anything less than 1 dB is very
difficult, and below 1/2 dB starts to be impossible. Even if you obtain that
noise figure, cable leakages and common mode would overwhelm the low antenna
level.
20 dB is about the limit for most receivers, although a dead receiver could
use 40. If the receiver is stone deaf, 40 dB would allow a workable noise
figure at the front end. :-)
This low sensitivity is why K6STI's antenna met with such limited reports of
success. If the site is very noisy with local distant noise, then the
antenna's noise floor is high enough to limit system noise floor. Otherwise,
the cables and input amplifier would set noise floor.
I have a similar thing here with a commercial loop antenna. Even though
vertically polarized, it is noise limited at my location by internal
amplifier noise. Now if I move it into a noisy location, it limits by
outside noise.
No matter what we try to do, we are not going to have a 0 dB noise figure.
When we start making the antenna sensitivity so low it requires gain with a
normal receiver so unrealistic that it limits on the front end noise, it is
useless. What good would seeing the S meter at S-2 or S-4 from amplifier
noise do? That is what the popular commercial loop I have does. In a quiet
location, it limits on its own internal amplifier noise. Six dB less gain
does not change S/N ratio one bit.
We should all question systems that need 40dB with normal receivers. 20 dB
is more rational.
73 Tom
N4IS JC Wrote:>>
At -140 dBm and 250 Hz noise bandwidth, the system would require a 1 dB
noise figure front end. That's about 35 deg K noise temperature.
>>>
Tom is as usual 100% right, the RX system gain should be near 1 dB, it means
the preamp at 1.8 MHz should be .5 dB NF the input filter and the feed line
< then .5 dB att. Together, not each.
To make things more complicated, when there is no atmospheric noise, like
we have in some winter days, the band noise can drop to 100K, and at that
point the gain of the WF and the NF of the system should be designed to no
more than 3 dB deterioration on signal to noise ratio, it means the 1 dB is
not enough, the solution for that is a bigger WF.
>>
Besides that, if the gain is so far negative the coaxial cable will easily
become more of an antenna than the thing we call an antenna.
<<
This is most common problem for all flag . EWE; WF and low gain antennas.
The only way to overcome this is using good quality twist par UNSHILDED
cable, or choke the cable as much possible. If you know what you are doing.
However detuning any structure or antenna at the same frequency is a must,
it can deteriorate the directivity of the RX antenna to make it useless. No
free beef here.
>>
40 dB gain in front of a receiver is pure fantasy, unless the receiver is
dead as a door nail.
<<
Tom, I'm afraid I disagree but agree with some of that, I am using a 43dB
gain preamp since 2010 with not a single failure yet, but I understand your
point. It is so delicate to implement that most of fellow that try it fail.
Even aluminum enclosure does not shield it enough, 40 dB gain is 10.000
voltage gain, it needs a dual shield with steel to cut magnetic field, the
feed lines must be decoupled over 80 dB, relays must be 100dB or more in
isolation, and much more details that I won't cover.
It is not a weekend project.
Regards
JC
N4IS
At -140 dBm and 250 Hz noise bandwidth, the system would require a 1 dB
noise figure front end. That's about 35 deg K noise temperature.
>>>
Tom is as usual 100% right, the RX system gain should be near 1 dB, it means
the preamp at 1.8 MHz should be .5 dB NF the input filter and the feed line
< then .5 dB att. Together, not each.
To make things more complicated, when there is no atmospheric noise, like
we have in some winter days, the band noise can drop to 100K, and at that
point the gain of the WF and the NF of the system should be designed to no
more than 3 dB deterioration on signal to noise ratio, it means the 1 dB is
not enough, the solution for that is a bigger WF.
>>
Besides that, if the gain is so far negative the coaxial cable will easily
become more of an antenna than the thing we call an antenna.
<<
This is most common problem for all flag . EWE; WF and low gain antennas.
The only way to overcome this is using good quality twist par UNSHILDED
cable, or choke the cable as much possible. If you know what you are doing.
However detuning any structure or antenna at the same frequency is a must,
it can deteriorate the directivity of the RX antenna to make it useless. No
free beef here.
>>
40 dB gain in front of a receiver is pure fantasy, unless the receiver is
dead as a door nail.
<<
Tom, I'm afraid I disagree but agree with some of that, I am using a 43dB
gain preamp since 2010 with not a single failure yet, but I understand your
point. It is so delicate to implement that most of fellow that try it fail.
Even aluminum enclosure does not shield it enough, 40 dB gain is 10.000
voltage gain, it needs a dual shield with steel to cut magnetic field, the
feed lines must be decoupled over 80 dB, relays must be 100dB or more in
isolation, and much more details that I won't cover.
It is not a weekend project.
Regards
JC
N4IS
Date: Tue, 8 Sep 2015 10:45:37 -0400
From: "Tom W8JI"
> Tom, I'm afraid I disagree but agree with some of that, I am using a
> 43dB
> gain preamp since 2010 with not a single failure yet, but I understand
> your
> point. It is so delicate to implement that most of fellow that try it
> fail.
> Even aluminum enclosure does not shield it enough, 40 dB gain is 10.000
> voltage gain, it needs a dual shield with steel to cut magnetic field, the
> feed lines must be decoupled over 80 dB, relays must be 100dB or more in
> isolation, and much more details that I won't cover.
>
> It is not a weekend project.
JC,
The problem is gain and noise figure, not shielding.
Let's assume a receiver with 250 Hz bandwidth has a MDS (3dB S+S/N) of -135
dBm. This is a 15 dB receiver noise figure.
A 0.5 dB noise figure front end amplifier with NO other losses would
produce -149.5 dB MDS. That is the absolute maximum MDS sensitivity
obtainable with 250 Hz BW and 0.5dB total input noise figure.
If we include the receiver's noise figure, 14.5 dBm gain would result is a
system composite noise figure of 3.44 dB. Increasing amplifier gain (with no
change in amplifier 0.5 dB noise figure) results in the following system
composite noise figures:
14.5dB = 3.44 dB
20 dB = 1.55 dB
25 dB = 0.86 dB
30 dB = 0.62 dB
35 dB = 0.54 dB
40 dB = 0.51 dB
At someplace around 20-25 dB, you get into system limits. The improvement
from 30 dB to 40 dB is only 0.11 dB. No one will notice that.
This of course varies with the receiver, but few receivers are worse than
this example.
Let's say we have an input stage NF of .5 dB with 15 dB gain. In order to
have a cascade NF of .7dB the second stage has to have about a 2 dB NF.
All of this is peanuts. A .6dB noise figure is a 3 dB MDS of -149.4
dBm, while a 2 dB NF is a MDS of -148 dBm with 250 Hz BW.
There is a point where inevitable system flaws make using an antenna with
such negative gain to require less than 1 dB NF impossible for "copy this
plan". This is why Beazley's out-of-phase small horizontal elements were
mostly met with didn't work. The problem with models is we can build perfect
systems that we cannot repeat in the real world.
Again my example of the small commercial loop I have. It limits by loop
internal noise by many dB at my location, and common mode on the cable is
very evident. If I moved the same antenna to a location with 20 dB more
external ambient noise floor, it would limit on external noise.
It seems unlikely most compact antennas are being used in locations so quiet
they need 30 dB gain, or .6 dB NF. If they had room, they would not have
as much ambient site noise.
I'm not being disagreeable, just describing the practical limits.
73 Tom
------------------------------
Message: 14
Date: Tue, 8 Sep 2015 11:37:22 -0400
From: "JC" <n4is@comcast.net>
To: "'Tom W8JI'" <w8ji@w8ji.com>, <topband@contesting.com>
Subject: Topband: [WARNING: ATTACHMENT(S) MAY CONTAIN MALWARE]RE:
Waller Flag Question
Message-ID: <000001d0ea4c$429046b0$c7b0d410$@comcast.net>
Content-Type: text/plain; charset="us-ascii"
>>
A 0.5 dB noise figure front end amplifier with NO other losses would produce
-149.5 dB MDS. That is the absolute maximum MDS sensitivity obtainable with
250 Hz BW and 0.5dB total input noise figure.
<<
Correct agree
>>
If we include the receiver's noise figure, 14.5 dBm gain would result is a
system composite noise figure of 3.44 dB.
At someplace around 20-25 dB, you get into system limits. The improvement
from 30 dB to 40 dB is only 0.11 dB. No one will notice that.
>>
Agree when the signal is above noise floor
The issue is most internal preamps are not designed for 160m, they need to
cover up to 50 MHz Even the ICOM Norton preamp does not have the muscle to
handle the signals level on 160m. If you turn on you internal preamp and use
a better preamp outside the radio with a good BPF or pre-selector you will
always have the flexibility to adjust the system NF reducing RF gain. But
there is no way to improve NF it is all about less degradation of the signal
to noise ratio.
The low NF is required only when the propagation noise is very low, the
concept of degradation is based on human skills and vary from operator to
operator, Most of us can copy CW signal 3 db above noise, 2 db above noise
is hard to copy, and 0 db SNR is very hard but some of us can compensate
that in their brain.
What I don't agree is that :
When the signal is at noise level. 3 db improvement on signal to noise ratio
can make a QSO possible, you can hear the DX early and for a long time
before it fade. 3 db degradation means 50% power noise and 50% signal power
noise. Near the MDS 3 db signal to noise is close to 3db NF improvement.
At this point the difference is copy or no copy. QSO or no QSO. Log the DX
or wait the next DX expedition.
<<
There is a point where inevitable system flaws make using an antenna with
such negative gain to require less than 1 dB NF impossible for "copy this
plan". This is why Beazley's out-of-phase small horizontal elements were
mostly met with didn't work. The problem with models is we can build perfect
systems that we cannot repeat in the real world.
>>
Agree , the practical noise figure for the system is around 2 dB NF, The way
to drop the noise floor few db more is reducing the BW to 100 Hz or 50 Hz.
An EME experienced CW operator knows that some signals you can copy using 50
Hz and no copy with 100 Hz. It is that simple in real weak signal DX.
The MDS can be low as -157dbm for 50 Hz BW, and this is almost -s9, if
using 6db for each S unit. There is a lot of dynamic range bellow s0.
3db makes a lot of difference, I fight for every .1 dBm I can get.
<<
Again my example of the small commercial loop I have.
>>
The HWF is not so small, for the WF300 (300 sq Ft.) I use 44 ft boom and
loops 24ft.x12ft.
However for a noise location there is one ideal size, the trick is to drop
local manmade noise bellow or near the MDS. During the day the HWF looks
like a dummy load noiseless antenna. My system's uses different gain for
each band and tuned for a radios with 20 NF (no internal preamp on).
For noise locations the WF200 with 24 ft boom is a good choice.
Like I said , it's not a weekend project but can be done and it performs
very well. Polarization filter is a powerful tool for urban stations that
cannot be disregarded.
There is no commercial loop compatible or comparable to the HWF antenna, it
is a different ball game.
The results worth the time and money to implement it.
It all a good debate and we agree in most of it. Or at least we agree we
disagree.
Regards
JC
N4IS
From: "Tom W8JI"
> Tom, I'm afraid I disagree but agree with some of that, I am using a
> 43dB
> gain preamp since 2010 with not a single failure yet, but I understand
> your
> point. It is so delicate to implement that most of fellow that try it
> fail.
> Even aluminum enclosure does not shield it enough, 40 dB gain is 10.000
> voltage gain, it needs a dual shield with steel to cut magnetic field, the
> feed lines must be decoupled over 80 dB, relays must be 100dB or more in
> isolation, and much more details that I won't cover.
>
> It is not a weekend project.
JC,
The problem is gain and noise figure, not shielding.
Let's assume a receiver with 250 Hz bandwidth has a MDS (3dB S+S/N) of -135
dBm. This is a 15 dB receiver noise figure.
A 0.5 dB noise figure front end amplifier with NO other losses would
produce -149.5 dB MDS. That is the absolute maximum MDS sensitivity
obtainable with 250 Hz BW and 0.5dB total input noise figure.
If we include the receiver's noise figure, 14.5 dBm gain would result is a
system composite noise figure of 3.44 dB. Increasing amplifier gain (with no
change in amplifier 0.5 dB noise figure) results in the following system
composite noise figures:
14.5dB = 3.44 dB
20 dB = 1.55 dB
25 dB = 0.86 dB
30 dB = 0.62 dB
35 dB = 0.54 dB
40 dB = 0.51 dB
At someplace around 20-25 dB, you get into system limits. The improvement
from 30 dB to 40 dB is only 0.11 dB. No one will notice that.
This of course varies with the receiver, but few receivers are worse than
this example.
Let's say we have an input stage NF of .5 dB with 15 dB gain. In order to
have a cascade NF of .7dB the second stage has to have about a 2 dB NF.
All of this is peanuts. A .6dB noise figure is a 3 dB MDS of -149.4
dBm, while a 2 dB NF is a MDS of -148 dBm with 250 Hz BW.
There is a point where inevitable system flaws make using an antenna with
such negative gain to require less than 1 dB NF impossible for "copy this
plan". This is why Beazley's out-of-phase small horizontal elements were
mostly met with didn't work. The problem with models is we can build perfect
systems that we cannot repeat in the real world.
Again my example of the small commercial loop I have. It limits by loop
internal noise by many dB at my location, and common mode on the cable is
very evident. If I moved the same antenna to a location with 20 dB more
external ambient noise floor, it would limit on external noise.
It seems unlikely most compact antennas are being used in locations so quiet
they need 30 dB gain, or .6 dB NF. If they had room, they would not have
as much ambient site noise.
I'm not being disagreeable, just describing the practical limits.
73 Tom
------------------------------
Message: 14
Date: Tue, 8 Sep 2015 11:37:22 -0400
From: "JC" <n4is@comcast.net>
To: "'Tom W8JI'" <w8ji@w8ji.com>, <topband@contesting.com>
Subject: Topband: [WARNING: ATTACHMENT(S) MAY CONTAIN MALWARE]RE:
Waller Flag Question
Message-ID: <000001d0ea4c$429046b0$c7b0d410$@comcast.net>
Content-Type: text/plain; charset="us-ascii"
>>
A 0.5 dB noise figure front end amplifier with NO other losses would produce
-149.5 dB MDS. That is the absolute maximum MDS sensitivity obtainable with
250 Hz BW and 0.5dB total input noise figure.
<<
Correct agree
>>
If we include the receiver's noise figure, 14.5 dBm gain would result is a
system composite noise figure of 3.44 dB.
At someplace around 20-25 dB, you get into system limits. The improvement
from 30 dB to 40 dB is only 0.11 dB. No one will notice that.
>>
Agree when the signal is above noise floor
The issue is most internal preamps are not designed for 160m, they need to
cover up to 50 MHz Even the ICOM Norton preamp does not have the muscle to
handle the signals level on 160m. If you turn on you internal preamp and use
a better preamp outside the radio with a good BPF or pre-selector you will
always have the flexibility to adjust the system NF reducing RF gain. But
there is no way to improve NF it is all about less degradation of the signal
to noise ratio.
The low NF is required only when the propagation noise is very low, the
concept of degradation is based on human skills and vary from operator to
operator, Most of us can copy CW signal 3 db above noise, 2 db above noise
is hard to copy, and 0 db SNR is very hard but some of us can compensate
that in their brain.
What I don't agree is that :
When the signal is at noise level. 3 db improvement on signal to noise ratio
can make a QSO possible, you can hear the DX early and for a long time
before it fade. 3 db degradation means 50% power noise and 50% signal power
noise. Near the MDS 3 db signal to noise is close to 3db NF improvement.
At this point the difference is copy or no copy. QSO or no QSO. Log the DX
or wait the next DX expedition.
<<
There is a point where inevitable system flaws make using an antenna with
such negative gain to require less than 1 dB NF impossible for "copy this
plan". This is why Beazley's out-of-phase small horizontal elements were
mostly met with didn't work. The problem with models is we can build perfect
systems that we cannot repeat in the real world.
>>
Agree , the practical noise figure for the system is around 2 dB NF, The way
to drop the noise floor few db more is reducing the BW to 100 Hz or 50 Hz.
An EME experienced CW operator knows that some signals you can copy using 50
Hz and no copy with 100 Hz. It is that simple in real weak signal DX.
The MDS can be low as -157dbm for 50 Hz BW, and this is almost -s9, if
using 6db for each S unit. There is a lot of dynamic range bellow s0.
3db makes a lot of difference, I fight for every .1 dBm I can get.
<<
Again my example of the small commercial loop I have.
>>
The HWF is not so small, for the WF300 (300 sq Ft.) I use 44 ft boom and
loops 24ft.x12ft.
However for a noise location there is one ideal size, the trick is to drop
local manmade noise bellow or near the MDS. During the day the HWF looks
like a dummy load noiseless antenna. My system's uses different gain for
each band and tuned for a radios with 20 NF (no internal preamp on).
For noise locations the WF200 with 24 ft boom is a good choice.
Like I said , it's not a weekend project but can be done and it performs
very well. Polarization filter is a powerful tool for urban stations that
cannot be disregarded.
There is no commercial loop compatible or comparable to the HWF antenna, it
is a different ball game.
The results worth the time and money to implement it.
It all a good debate and we agree in most of it. Or at least we agree we
disagree.
Regards
JC
N4IS
Dentro un messaggio di JC ho trovato un link interessante sulle prove della Waller Flag, che ancora non avevo letto:
http://www.qsl.net/k4fk/presentation...nas-SFDXAs.pdf
Spero di essere stato utile.
Saluti
Come avevo promesso posto le foto dell'antenna completamente installata oggi pomeriggio.
Commenta