Single vs Stacked Dipole for Milair Reception?

Status
Not open for further replies.

MDScanFan

Member
Joined
Jan 23, 2010
Messages
359
Location
USA
I want to build a broadband dipole for 225-400 MHz reception using scrap material from a past project. The design is complete. Right now I am debating a single dipole vs a stacked configuration. With the right spacing the stacked version will give me around 3 dB more horizon gain than the single dipole. The mini circuits combiner I would use has about 0.3 dB insertion loss.

The trade off with the stacked configuration is a null in the pattern above the horizon. The angle varies with frequency between 30 (high freq) and 45 ( low freq) deg off the horizon. If I am thinking about this correctly, then that is a worthwhile trade for more horizon gain because any aircraft at those angles would be relatively short range and reception should not be a problem even in the null.

Any recommendations?
 

prcguy

Member
Joined
Jun 30, 2006
Messages
15,233
Location
So Cal - Richardson, TX - Tewksbury, MA
I would go for the most gain at the horizon and any aircraft at high angles would be fairly close with an unobstructed view. The problem is 225-400MHz is a huge spectrum and stacking antennas for 3dB gain will only happen across a much narrower range, then it causes upward and downward pattern shift above and below the target frequency where the spacing is correct.

This is why you don't see many stacked antennas for that band and when you do they are expensive. I have a 225-400 4-bay stacked array inside a fiberglass radome, so I can't see what they did. The patent for my antenna mentioned ferrite as part of the solution for pattern manipulation but I don't know how they used it.

I want to build a broadband dipole for 225-400 MHz reception using scrap material from a past project. The design is complete. Right now I am debating a single dipole vs a stacked configuration. With the right spacing the stacked version will give me around 3 dB more horizon gain than the single dipole. The mini circuits combiner I would use has about 0.3 dB insertion loss.

The trade off with the stacked configuration is a null in the pattern above the horizon. The angle varies with frequency between 30 (high freq) and 45 ( low freq) deg off the horizon. If I am thinking about this correctly, then that is a worthwhile trade for more horizon gain because any aircraft at those angles would be relatively short range and reception should not be a problem even in the null.

Any recommendations?
 

Ubbe

Member
Joined
Sep 8, 2006
Messages
8,946
Location
Stockholm, Sweden
I think you might be misreading the mini-circuits specification. It is 0,3dB loss above the theoretical 3dB loss that all combiners/splitters has.
So the 3dB gain by stacking will instead give you a 0,3dB loss compared to a single antenna. You will have to use stacking harness using coax cables to get almost no loss. You have look at the total loss figure.

Combiner-loss.jpg


/Ubbe
 

prcguy

Member
Joined
Jun 30, 2006
Messages
15,233
Location
So Cal - Richardson, TX - Tewksbury, MA
The MiniCircuits power divider will give the exact same results as using a coaxial cable stacking harness, they both have about 3dB loss as a device but will ignore that when combining two signals in phase. What is actually happening, and you wont find this documented very easily is....

Combing two antennas actually adds 6dB gain, not 3dB. Yes, you read this right and probably never heard about it, but in the process you always loose 3dB by dividing the signal and feeding both antennas to accomplish this, so the end result of combining two antennas with the proper spacing and phasing will result in a roughly 3dB gain after the roughly 3dB of divider losses.

The power divider for combining two antennas can be made from two critical lengths of 75 ohm coax with both coax's to a common point on the input side, or a Wilkinson Power Divider, or an RF transformer with single primary and two secondaries. These all have a similar 3dB loss as a power divider.

I think you might be misreading the mini-circuits specification. It is 0,3dB loss above the theoretical 3dB loss that all combiners/splitters has.
So the 3dB gain by stacking will instead give you a 0,3dB loss compared to a single antenna. You will have to use stacking harness using coax cables to get almost no loss. You have look at the total loss figure.

Combiner-loss.jpg


/Ubbe
 
Last edited:

Ubbe

Member
Joined
Sep 8, 2006
Messages
8,946
Location
Stockholm, Sweden
If you look at that combiner you'll see it has cooling fins. The 3dB disappears inside that combiner and if you transmit half the power will have to be dissapated as heat. No such cooling fins on stacking harness as the loss is much less. The stacking harness have to be done to a specific frequency and will not be very broadbanded. That combiner takes a whole frequency band, to the expense of 50% loss.

Compare to a basstation site that use hybrid combiners that will have more than 50% loss and a more narrow banded cavity combiner with much less loss but needs to be retuned if the transmit frequency are changed.

/Ubbe
 

prcguy

Member
Joined
Jun 30, 2006
Messages
15,233
Location
So Cal - Richardson, TX - Tewksbury, MA
A stacking harness and 3dB hybrid as used in a combiner and a power divider all have 3dB loss, there is only a fraction of a dB difference between them. You should know this. A hybrid combiner does need tuning but its a different component for a different purpose but in the end it splits power with 3dB loss at each port just like a cable stacking harness. The difference with a 3dB hybrid is there is a load involved so that helps provide a constant 50 ohm match in the system and if you use it as an antenna combiner and one antenna gets disconnected, the reflected power dumps into the load. A Wilkinson divider or stacking harness cannot do that so this is why 3dB hybrids are used for transmitter combining. And when tuning them you go for minimal power dumped into the load, maximizing the power to the antenna.

I've been using hybrid combiners for probably 40yrs and I just installed a two transmitter combiner for two repeaters in my garage a few weeks ago and had to go through the alignment on that one recently, so I'm very familiar with them.

The bottom line of this specific conversation is combining two antennas with everything else equal including the same power to both antennas as there was with one is you get 6dB gain by stacking two antennas and 3dB loss in the power dividing to the two antennas leaving a net gain of 3dB for combining two antennas. They type of power divider used, coax stacking harness, Wilkinson or transformer type is irrelevant and they all will have about 3dB loss.

Whenever you take a signal or anything (water, beans, sheep) and divide it in two, you have two of that thing but with only half as much as you started with in each individual thing. You can't divide something in two and have more that 50% of what you started with in each part.

If you look at that combiner you'll see it has cooling fins. The 3dB disappears inside that combiner and if you transmit half the power will have to be dissapated as heat. No such cooling fins on stacking harness as the loss is much less. The stacking harness have to be done to a specific frequency and will not be very broadbanded. That combiner takes a whole frequency band, to the expense of 50% loss.

Compare to a basstation site that use hybrid combiners that will have more than 50% loss and a more narrow banded cavity combiner with much less loss but needs to be retuned if the transmit frequency are changed.

/Ubbe
 
Last edited:

Ubbe

Member
Joined
Sep 8, 2006
Messages
8,946
Location
Stockholm, Sweden
Your reasoning are correct.

If you look at a stacking harness you transform one antennas 50 ohm impedance to 100 ohm by using a 1/4 wave 75 ohm coax to get a 100 ohm impedance and can connect to the other antennas 100 ohm coax to form a 50 ohm connection. Both coaxes are then connected to the 50 ohm one that goes to a receiver.

At the connection point one antenna sees the other antennas 100 ohm coax and the down lead coax of 50 ohm and will divide the signal between them, all coaxes see the same voltage at the common connection point. The lower impedance will get the majority of the signal as it is a bigger load, and the higher ohm antenna coax gets a smaller part. So the signal are not split even between them and are not feeding half the signal to both. I guess it 66% of the energy that goes into the 50 ohm coax and 33% into the 100 ohm coax.

The combiner have a 30dB isolation between the antenna ports and no signal can go from one antenna and out to the other antenna, so it could mean that all signal goes right thru the combiner without any loss. But the electric circuit that makes it a combiner have that 3dB loss, 50%, at all times.

Remember that when we receive from two antennas we combine the signals from from two sources into one and double the signal. If we generate one signal and split up into two antennas we split the beans with half of them going to each antenna.

I have personally never seen receiving antennas being connected together using combiners, it's always stacking harness being used and probably for a reason.

/Ubbe
 

MDScanFan

Member
Joined
Jan 23, 2010
Messages
359
Location
USA
I appreciate the replies but that went on a tangent from the original question. I am familiar with power combiner and how it is used for this type of application. What I am asking is in regard to the geometry of uhf airband reception. I know getting more gain on the horizon is a good thing. I am wondering if anyone has experience with the 30-45 deg pattern nulls causing any problems with overall airband reception. There is no free lunch, so with increased omni gain comes the trade of additional nulls in the pattern somewhere.
 

prcguy

Member
Joined
Jun 30, 2006
Messages
15,233
Location
So Cal - Richardson, TX - Tewksbury, MA
If you consider an aircraft at 35,000ft straight above you in the worst antenna null, its only 6.6 miles away with no obstructions. You should be able to pick that up with a paper clip on the end of some coax. With an aircraft at 35,000ft and a 45deg angle upward, its still not that far away but probably in the clear with no obstructions and a paper clip will probably work.

I have never had a problem with my big commercial air antennas and aircraft up at high angles, they usually pin the S meter. The problems come up when the aircraft is down at the horizon and you are looking through hundreds of trees and buildings. That's where you need the gain.

I appreciate the replies but that went on a tangent from the original question. I am familiar with power combiner and how it is used for this type of application. What I am asking is in regard to the geometry of uhf airband reception. I know getting more gain on the horizon is a good thing. I am wondering if anyone has experience with the 30-45 deg pattern nulls causing any problems with overall airband reception. There is no free lunch, so with increased omni gain comes the trade of additional nulls in the pattern somewhere.
 

MDScanFan

Member
Joined
Jan 23, 2010
Messages
359
Location
USA
Thanks, this makes sense. I was running through the numbers after my last post and the geometry does look fine. Assuming a flat earth I calculated the slant range distance between my location and an aircraft at an altitude of 30 kft. The lowest null of my stacked pattern is at 30 deg. To get within that null the aircraft would only be about 9 miles away. At that distance I should hear the signal even if it is in the null.

In light of this I will likely target a three stacked configuration instead. The hardware I have on hand would work for this size. That would place the null at a shallower angle, maybe 20 deg or so, but should be fine too. Based on the math it looks like only the horizon + about 5 degrees or so is the critical region.

If you consider an aircraft at 35,000ft straight above you in the worst antenna null, its only 6.6 miles away with no obstructions. You should be able to pick that up with a paper clip on the end of some coax. With an aircraft at 35,000ft and a 45deg angle upward, its still not that far away but probably in the clear with no obstructions and a paper clip will probably work.

I have never had a problem with my big commercial air antennas and aircraft up at high angles, they usually pin the S meter. The problems come up when the aircraft is down at the horizon and you are looking through hundreds of trees and buildings. That's where you need the gain.
 

prcguy

Member
Joined
Jun 30, 2006
Messages
15,233
Location
So Cal - Richardson, TX - Tewksbury, MA
Height can be more important than gain and you will be experimenting with an unproven design with unknown gain and pattern . A simpler antenna up higher might provide more predictable results.

As an extreme example I once did a distance test to a large aircraft heading to Hawaii where I was on an 800ft hill with a 5w handheld in the VHF air band connected to a 1/4 wave whip on my vehicle roof. The aircraft took off from LAX and punched his GPS as he passed over me and headed west. We communicated fine for quite some time as the pilot called out the distances and it was getting weak but perfectly readable at 338 nautical miles or about 389 statute miles when we finally gave up. I believe we could have hit 400 statue miles but we were getting bored.

I just remember another test I did with the same pilot but heading east from LAX and I was comparing a Discone to a huge TACO brand VHF air band Yagi. I think it had 10 or 11 elements. Anyway on many flights we communicated fine to about 200 miles then it dropped like a rock on both antennas at the same time. I didn't really notice much if any difference between the huge Yagi and the Discone because we never got far enough apart to get very noisy on the Discone. The radio I was using didn't have an S meter but I suspect the signal strength was much better on the Yagi, but terrain ultimately killed the signal on both antennas at the exact same distance.

Thanks, this makes sense. I was running through the numbers after my last post and the geometry does look fine. Assuming a flat earth I calculated the slant range distance between my location and an aircraft at an altitude of 30 kft. The lowest null of my stacked pattern is at 30 deg. To get within that null the aircraft would only be about 9 miles away. At that distance I should hear the signal even if it is in the null.

In light of this I will likely target a three stacked configuration instead. The hardware I have on hand would work for this size. That would place the null at a shallower angle, maybe 20 deg or so, but should be fine too. Based on the math it looks like only the horizon + about 5 degrees or so is the critical region.
 

MDScanFan

Member
Joined
Jan 23, 2010
Messages
359
Location
USA
I appreciate the anecdotes. And you bring up a very good point regarding height. I will ponder my plans a bit more before committing. I will report back on what I end up doing.
 

Ubbe

Member
Joined
Sep 8, 2006
Messages
8,946
Location
Stockholm, Sweden
I will report back on what I end up doing.
Please do. We are all interested in how it will work. If you get that combiner solution, and you probably need to as no stacking harness will have that huge frequency range needed, then before finalising the installation and when testing at a more easy to reach position, then compare signal levels (from a fixed transmitter site) stacked with a combiner and to a single dipole without any combiner loss.

One question, how did you manage to get the dipole that wide in frequency range? Is it a 3 inch thick element? It's a 30% bandwidth at a 310Mhz center frequency when a normal sized dipole have 10%.

/Ubbe
 

MDScanFan

Member
Joined
Jan 23, 2010
Messages
359
Location
USA
I am just going thick to get the bandwidth needed plus some marking at the band edges to account for tolerances. Right now it is a manageable ~3.5” thick.

I agree with the suggestion on the initial testing. What I will likely do is fabricate it as a three stacked configuration and initially just connect one of the elements for some reception testing. Then add the rest for comparison.

Please do. We are all interested in how it will work. If you get that combiner solution, and you probably need to as no stacking harness will have that huge frequency range needed, then before finalising the installation and when testing at a more easy to reach position, then compare signal levels (from a fixed transmitter site) stacked with a combiner and to a single dipole without any combiner loss.

One question, how did you manage to get the dipole that wide in frequency range? Is it a 3 inch thick element? It's a 30% bandwidth at a 310Mhz center frequency when a normal sized dipole have 10%.

/Ubbe
 

prcguy

Member
Joined
Jun 30, 2006
Messages
15,233
Location
So Cal - Richardson, TX - Tewksbury, MA
Once you have one element playing across the frequency range you want and are ready to test the others I would suggest testing the vertical radiation pattern. I have all the stuff to do this but you can get an idea of the pattern at home if you can come up with a signal generator and a directional antenna within the VHF air band.

I would place your test antenna as high in the air as you can and on its side on a rotatable plate like for a Lazy Susan. If you could get it 10 or 15ft in the air on a fiberglass ladder that's probably ok. Then set up the directional antenna a few feet off the ground, 30 or 40ft away and horizontal pointing up at your antenna to be tested. This reduces some ground bounce that can disturb the testing. The Lazy Susan thing should be tilted to match the angle back to your low directional antenna. Connect the signal generator to the low horizontal antenna and a receiver with an S meter to your antenna to be tested.

Fire everything up on a frequency in the middle of the VHF air band but in between channels and set the signal generator level so you have a usable signal with the test antenna broad side to your low directional antenna. Then rotate the the antenna being tested from broadside to + and - 90 degrees to see where the peaks and nulls are. If it peaks broadside then thats great. If it peaks higher as you go away from broadside, that's not good. You can also test a single element then add the others to see if the gain goes up and what happens to the pattern.

I have a nice 100MHz to 1GHz Log Periodic for this and some other things that make the job easier like a receiver that reads out in dBm and a spectrum analyzer that will do the same. With the spectrum analyzer it can be set up to draw a trace of the exact pattern if you can rotate the antenna at a very constant rate and sync that with the analyzer trace speed. Its fun to set up tests like this and see the results.



I am just going thick to get the bandwidth needed plus some marking at the band edges to account for tolerances. Right now it is a manageable ~3.5” thick.

I agree with the suggestion on the initial testing. What I will likely do is fabricate it as a three stacked configuration and initially just connect one of the elements for some reception testing. Then add the rest for comparison.
 

Ubbe

Member
Joined
Sep 8, 2006
Messages
8,946
Location
Stockholm, Sweden
I wonder if it would be possible to use a VNA to measure an antennas frequency response? If we use a LPDA antenna, that sends the same signal level over it's frequency range, or even a discone might work if we use a close distance of 30ft, and a VNA in transmission mode might send out a signal at -10dBm or -20dBm and we connect that to a coax that goes to that antenna and we connect the receive port of the VNA to the antenna to be measured, would that work?

A calculator shows that with two antennas with 0dB gain and 30ft of air between them will attenuate the signal by 50dB. It will still be well above the noise level and the VNA software should be able to plot a frequency response. Any transmitter signals within range would make a spike in the plot but that could easily be omitted as you concentrate on the levels between the spikes. Then we should get a real live and true result of the antennas performance. If we happen to have to identical antennas they could be used at both ends and the resulting plot would have to be dynamicly reduced to half. The strongest signal in the plot could be -60dBm and a signal that measures at a -80dBm level would have to be corrected to -70dBm as both antennas have attentuated that signal 10dB that will show up as a 20dB attenuation in the plot.

A standard preamplifier have it's compression point at an output of +20dBm so we could amplify that weak VNA signal at a -10dBm or -20dBm level to improve the signal/noise relationship. If a simple and inexpensive test setup can be made it would help immensely when doing antenna installations and modifications.

/Ubbe
 

prcguy

Member
Joined
Jun 30, 2006
Messages
15,233
Location
So Cal - Richardson, TX - Tewksbury, MA
Yes, this is what is done on most antenna ranges. Using a VNA and sometimes with a wide band amplifier, you calibrate the system (range) with an antenna of known gain, then substitute your antenna under test. This will give you a good idea of the gain over frequency and you must do some other tests like raise and low the antenna under test to see if you get the exact same results, then its pretty good data. If the gain numbers change over height then ground bounce is skewing the measurement and you must fix that before continuing.

I've even done this in my garage comparing a bucket of rubber duck type antennas. I pick one as my reference, doesn't really matter which one, then cal the system. Then I substitute other antennas and I will have the gain or loss compared to my reference antenna over frequency. Most scalar and vector network analyzer have a very specific modulation and the VNA receiver is looking for that specific modulation and should reject other carriers within the frequency range being tested. At least to some extent. Instead of a preamplifier on the antenna being tested its better to use a power amplifier on the VNA generator to bring the levels up for better SNR. You just have to be careful not to interfere with anyone's receiver.

My favorite VNA for this is the Agilent Field Fox as it also has a separate antenna analyzer function (S11) so I can get good VSWR data along with antenna range data.

I wonder if it would be possible to use a VNA to measure an antennas frequency response? If we use a LPDA antenna, that sends the same signal level over it's frequency range, or even a discone might work if we use a close distance of 30ft, and a VNA in transmission mode might send out a signal at -10dBm or -20dBm and we connect that to a coax that goes to that antenna and we connect the receive port of the VNA to the antenna to be measured, would that work?

A calculator shows that with two antennas with 0dB gain and 30ft of air between them will attenuate the signal by 50dB. It will still be well above the noise level and the VNA software should be able to plot a frequency response. Any transmitter signals within range would make a spike in the plot but that could easily be omitted as you concentrate on the levels between the spikes. Then we should get a real live and true result of the antennas performance. If we happen to have to identical antennas they could be used at both ends and the resulting plot would have to be dynamicly reduced to half. The strongest signal in the plot could be -60dBm and a signal that measures at a -80dBm level would have to be corrected to -70dBm as both antennas have attentuated that signal 10dB that will show up as a 20dB attenuation in the plot.

A standard preamplifier have it's compression point at an output of +20dBm so we could amplify that weak VNA signal at a -10dBm or -20dBm level to improve the signal/noise relationship. If a simple and inexpensive test setup can be made it would help immensely when doing antenna installations and modifications.

/Ubbe
 

Ubbe

Member
Joined
Sep 8, 2006
Messages
8,946
Location
Stockholm, Sweden
I doubt that a $50 VNA have any kind of interference protection, by modulation type or other ways, like a $1000 SiteMaster have.
You dont HAVE to measure SWR at the same time as plotting signal levels, you can do that at a seperate session.
The signal output are a sweeping signal so it should at worst be detected as a short blipp by a normal squelch circuit. If you manage to get a +10dBm signal into an antenna then that is 10mW input power that hardly have any range to reach a commercial receiver. But it is advisable to not continue tests at any prolonged periods of time.

/Ubbe
 
Status
Not open for further replies.
Top