# 8 bay TV Antenna question

#### gary350

##### Well-Known Member
WOW........... RG6 connectors are expensive, $1,$2, $5,$6, $7, each. That must be why I use to bought the reuseable connectors 10 years ago.$5 to \$12 postage.

Who has good prices and which ones do I buy?

#### Tesla23

##### Member
Another thing that occurred to me is that comparing the two feeding arrangements you show in your first post, one advantage of using open wire feeder to combine the arrays:

is that you know that you are combining the two arrays in phase. When you use separate baluns like this:

you run the risk that unless the baluns have some sort of polarity indicator (and the ones I have seen haven't), there is a possibility that the signals are out of phase. You should compare the received signal strengths with the connections to one of the baluns swapped and choose the arrangement that gives you the strongest signal. Do this for each pair of bays to ensure that each pair is working properly. When you then add the two pairs together, you need to again ensure that the signals are adding in phase. You can check this by reversing both baluns on one pair and see what happens to your combined signal strength.

Last edited:

#### gary350

##### Well-Known Member
Another thing that occurred to me is that comparing the two feeding arrangements you show in your first post, one advantage of using open wire feeder to combine the arrays:

View attachment 125013

is that you know that you are combining the two arrays in phase. When you use separate baluns like this:

View attachment 125014

you run the risk that unless the baluns have some sort of polarity indicator (and the ones I have seen haven't), there is a possibility that the signals are out of phase. You should compare the received signal strengths with the connections to one of the baluns swapped and choose the arrangement that gives you the strongest signal. Do this for each pair of bays to ensure that each pair is working properly. When you then add the two pairs together, you need to again ensure that the signals are adding in phase. You can check this by reversing both baluns on one pair and see what happens to your combined signal strength.
Yes you are right, I learned that 10 years ago. I learned to collect 1 balun to the left 4 bay side first then check field strength meter on TV. It turns out swaping wires on 1 balun gives a high & low signal, I did not expect that. With left side balun connected for high signal then connected right side balun and swapped wires on the right on it for best signal. With right side balun connected to T combiner high signal antenna produces the best signal on the field strength meter. I learned testing a weak signal is many times better than testing a strong signal because antenna improvements are often not noticed when already having a very strong signal. This antenna is 10 years old time to clean it up and test all the parts & make improvements. I like the antenna close to the ground so I can work on it but after getting it working better I want it higher up. I found this antenna in my amateur radio antenna book. There are no 2:1 baluns to be found and no plans to build one either.

#### gary350

##### Well-Known Member
Does it make any difference what type metal the reflector screen is made of ?

I am building a new antenna. I was going to use 1/4" hole hardware cloth because small holes work best for high frequency. Living in tornado alley 1/4" is probably too much wind resistance maybe 1/2" holes are better.

No steel tubing in this antenna I am using plastic. Any suggestions before I get started building ?

Book says, not to mount elements on the ends, only mount then at center. Flat sheet metal aluminum flapped in the wind mounted in center so I have always used steel screws on both ends of the elements. Copper wires has to be attached to aluminum elements. Aluminum screws will work if I can get some.

Book says, reflector screen should be .2 wave length from elements. This is a bit of a problem TV is a wide range of frequencies. Maybe I make notes for channel frequencies to decide what to do. I have tested, 5", 6", 10" that was 10 years ago.

Book says, reflector screen should be 1/4 wave length larger than elements to be effective. I learned from experiments larger screen works better, as the screen size changes signal strength chances from channel to channel. A screen size for 1 size fits all channels has not been easy.

Last edited:

#### gary350

##### Well-Known Member
TV Fool shows about 60 more channels than there was 10 years ago no exact count yet but about 90 channels. The channels I want to see are between 303 degrees compass magnetic to 330 degree. TV Fool shows station in rows by degrees from my house. Math shows antenna I already built is 540 Mhz

303 degree = channels 180 mhz and 550 mhz

313 degree = 195 545 MHz

318 degree = 545 575 640 Mhz

325 degree = 85 510 Mhz

329 degrees = 88 476 515 524 590 584 595 603 626 Mhz

330 degrees = 79 Mhz

Is it best to build antenna to target middle range of frequencies or target low end or high end ?

Last edited:

#### Tesla23

##### Member
You need a fairly broadband antenna to cover the UHF channels (510 - 640MHz), and you want to throw in a couple of VHF ones as well, down to 79MHz.

This means that you have to be careful in taking a narrow-band design, like that ARRL collinear, and using it for likely dimensions. Spacing to planar reflector - I'd suggest quarter wavelength or a bit less at the bottom of the band, but this is a starting value - it's a trade-off over the range. I could make similar suggestions on element lengths, spacings etc.., but in an array all these things interact and the only real way to know what is going on is to build and experiment or to simulate. If you just pick likely numbers for each parameter it will work in some fashion, but you would be lucky to get the best gain for your effort. I'd look for published designs, preferably with measured gains, or reliable simulations.

There seems to be a reasonable amount of activity out there with folk using antenna design tools to simulate and try to optimise this sort of design, over the frequency range you want to cover. e.g.
https://www.digitalhome.ca/forum/18...w-tie-tv-antenna-designs-ff4-m4-featured.html
User mclapp on that thread is doing NEC simulations of his designs and seems to have a reasonable number of happy devotees building his designs. The sort of data he produces: https://imageevent.com/holl_ands/multibay/8bayrefl/cm4228hdmod
is the sort of thing I'd expect from this sort of simulation, so maybe his results are trustworthy. No guarantees, but I suspect there are others out there.

As far as the reflector material is concerned, you need a good array of horizontal rods. At the top of your band the wavelength is about half a meter, so you don't need the rod spacing any less than lambda/10 = 50mm (2in). You do need continuous horizontal rods though (and not too skinny), whether these are held together by a few or many vertical ones is simply a construction preference.

The only data I have is from an old copy of Jasik (Antenna Engineering Handbook)

So these do provide useful broadband gain, but he does omit to tell us the spacing to the reflector.

#### gary350

##### Well-Known Member
When I bought my factory CM4228 it would not pick up low VHF channel 5 = 79 MHz or 85 MHz or 88 Mhz. After I added a reflector screen to the antenna horizontal rods signal was stronger but still no channel 5. After putting antenna up 10 ft higher to 30 ft channel 5 was watchable most of the time. I added a larger 48"x48" reflector screen signal was stronger and channel 5 was good. Then I sold the factory CM4228 and built my own signal was better. Solid sheet metal bow ties made a very big signal increase on the field strength meter. Smaller screen holes gave a stronger signal too. Adjusting the space between screen & bow ties I could tune in each channel by watching field strength meter get stronger. I was having trouble with low VHF so I adjusted screen bow ties spacing for the best signal and all the other channels were good too. That was 10 years ago. Antenna is getting rusty and signal seems to be getting worse. Antenna I have now works good, it receives all channels, it is old rusty antenna, i can probably build a better one.

Last edited:

#### gary350

##### Well-Known Member
Does anyone have a circuit drawing for a 150 ohm to 75 ohm balun ?

#### Nigel Goodwin

##### Super Moderator
Essentially what you're trying to do is combine two aerials into one - which gives you extra gain, by making the complete array more directional.

This is usually done with a 'phasing harness' - essentially three pieces of coax (of specific sizes) connected in a Y shape.

#### gary350

##### Well-Known Member
Essentially what you're trying to do is combine two aerials into one - which gives you extra gain, by making the complete array more directional.

This is usually done with a 'phasing harness' - essentially three pieces of coax (of specific sizes) connected in a Y shape.
I already have that. I am trying to reduce losses by eliminating 5 coax cable connects, T combiner, 2 baluns.

#### Nigel Goodwin

##### Super Moderator
I already have that. I am trying to reduce losses by eliminating 5 coax cable connects, T combiner, 2 baluns.
And is it accurately calculated and constructed? - you can't just stick bits of old coax together, it needs to be accurate, and presumably without spurious joins?.

They were available ready built from the aerial manufacturers, for combining either two or four aerials to provide greater gain.

#### gary350

##### Well-Known Member
And is it accurately calculated and constructed? - you can't just stick bits of old coax together, it needs to be accurate, and presumably without spurious joins?.

They were available ready built from the aerial manufacturers, for combining either two or four aerials to provide greater gain.
When I built my first antenna years ago I copied the CM4228 antenna, I assume factory antenna is correct it works? Somewhere I read coax from each 4 bay should be 1/4 wavelength each, 2 coax 10" long are 4" too short the 4 bays are 24" apart, I can not find that information now. I'm not exactly sure when calculating antenna MHz do I use total dipole length = 1/2 wave length or the length of 1 element = 1/4 wavelength? If I can make the 2 coax 1/2 wave length they will be long enough.

I was hoping to build the antenna that has the 4 bays connected with wires so I can eliminate 5 coax connectors & the T combiner but I don't have a 2:1 balun? Maybe I should build this with a 4"1 balun and test it. ?

ARRL book says, reflector screen should be 1/4 wavelength from the dipoles but CM4228 has them 5" apart while DB8 has them 4" apart. ? A few years ago I made screen adjustable for testing 10" spacing was a big gain increase on the field strength meter on certain channels and other channels were lost.

So far no problem receiving a good signal at Low VHF using the CM4228 design and DB8 design. I want to make a better 8 bay antenna what ever works best. It will be nice to get it right the first time and not have to keep changing it to make it work.

One thing I know about testing and making changes, NEVER change 2 things at the same time if it gets better or worse you don't know which change did it. Years ago when I was testing screen dipole spacing I wish I had made notes now I have to do that again and make notes.

Another things I don't know, does balun add length to a 1/4" wavelength coax cable?

Last edited:

#### Ylli

##### Active Member
Defining things in terms of wavelength only works if you are dealing with a single frequency or a very narrow range of frequencies. You are trying to over a wide range.

#### gary350

##### Well-Known Member
Defining things in terms of wavelength only works if you are dealing with a single frequency or a very narrow range of frequencies. You are trying to over a wide range.
Yes......if you compare 8 bay TV antennas CM4228 and DB8 to ARRL 8 bay antennas electronic rules do not apply. I hate to waste time building another copy of CM4228 and still have the same thing I already have. I wanted to build an antenna with only 1 balun but someone said, impedance matching works best like the antenna I already have. Only improvement I can actually make is build a new antenna with plastic frame not metal and use plastic screws not metal, maybe I can eliminate 2 coax connectors by soldering coax cable directly to the copper antenna wires. Putting antenna up another 10 ft higher might be better than trying to make improvements or build a better antenna.

#### gary350

##### Well-Known Member
You need a fairly broadband antenna to cover the UHF channels (510 - 640MHz), and you want to throw in a couple of VHF ones as well, down to 79MHz.

This means that you have to be careful in taking a narrow-band design, like that ARRL collinear, and using it for likely dimensions. Spacing to planar reflector - I'd suggest quarter wavelength or a bit less at the bottom of the band, but this is a starting value - it's a trade-off over the range. I could make similar suggestions on element lengths, spacings etc.., but in an array all these things interact and the only real way to know what is going on is to build and experiment or to simulate. If you just pick likely numbers for each parameter it will work in some fashion, but you would be lucky to get the best gain for your effort. I'd look for published designs, preferably with measured gains, or reliable simulations.

There seems to be a reasonable amount of activity out there with folk using antenna design tools to simulate and try to optimise this sort of design, over the frequency range you want to cover. e.g.
https://www.digitalhome.ca/forum/18...w-tie-tv-antenna-designs-ff4-m4-featured.html
User mclapp on that thread is doing NEC simulations of his designs and seems to have a reasonable number of happy devotees building his designs. The sort of data he produces: https://imageevent.com/holl_ands/multibay/8bayrefl/cm4228hdmod
is the sort of thing I'd expect from this sort of simulation, so maybe his results are trustworthy. No guarantees, but I suspect there are others out there.

As far as the reflector material is concerned, you need a good array of horizontal rods. At the top of your band the wavelength is about half a meter, so you don't need the rod spacing any less than lambda/10 = 50mm (2in). You do need continuous horizontal rods though (and not too skinny), whether these are held together by a few or many vertical ones is simply a construction preference.

The only data I have is from an old copy of Jasik (Antenna Engineering Handbook)

View attachment 125033
View attachment 125034

So these do provide useful broadband gain, but he does omit to tell us the spacing to the reflector.
I have read this several times and still trying to figure out what it is telling me?

A = 7 1/2" bow ties but what does 70" mean? Is that 70 inches or degrees? This has no screen.

B = dipole in front of a screen 22" x 19". The same 7 1/2" bow tie has a screen 19 x 22.

C = 2 bay of 8 stacked at 23". This mean there is 8 bow ties. What does stacked at 23" mean?

C has highest gain but falls low at 450 MHz but it is still better than B and A.

This 7.5" bow ties put the resonance frequency of the antenna at 720 MHz.

The highest freq. I am trying to receive is 640 MHz. Lowest UHF = 476 MHz. Mid Freq. = 558 MHz. If I make 10" bow ties = 540 MHz.

Graph shows 476 to 640 Mhz = 8.3 to 12 to 10.8 for 8 bow ties. My antenna is 8 bow ties. This looks good.

I am surprised my antenna picks up 79 85 88 MHz so well. I wonder if they could be a low harmonic frequency.

12.5" bow ties = 432 MHz.

Just thinking. ? ? ?

10" bow ties will be high gain at 540 MHz and a bit lower at 476 and 640 MHz.

Instead of 5" between each 4 bay maybe it should be 1/4 wavelength like ARRL books says = 10" spacing.

Last edited:

#### Tesla23

##### Member
You won't be able to optimise your antenna simply by assembling snippets of information. In an array like this the items interact, of particular significance when you want broadband operation. You have taken a commercial design (which may or may not have been reasonably optimised), and tweaked it by experiment to improve the performance in your situation. This is fine, but accept that many of the parameters of that design were probably the result of considerable experimentation, they may have started with some textbook formula, but then altered them to improve various performance parameters. They may have aimed at covering a different frequency range to what you need, and you are finding ways to boost the performance on particular channels at the expense of ones you don't want.

All I can give you is general advice to guide your experimentation. If you enjoy experimentation, by all means push on, but accept that with the tools you have you could easily end up 3dB (at least) down compared to a numerically optimised design. If you enjoy building but want it to work, then I'd suggest starting with a published, optimised design.

Looking at your pictures, and from your experiments with the combiners that suggest your impedances are way out, I think that you could improve your feeding structure.

Assuming the impedance at the feedpoint of the bowties is around 300 ohms, then in the diagram Zd = 300 ohms (Zd is the impedance seen looking upward into the top element, with the feedline disconnected.). What we want is that at Z1 we get the combination of both the top dipoles, and if they contribute equally we expect Z1 = 150 ohms. (Z1 is the impedance seen looking upwards into the two dipoles, with the lower feedline disconnected.) This is achieved if D1 is lambda/2 independent of the impedance of the transmission line you connect them with. I'll come back to that impedance later Just note that D1 is lambda/2 at 621MHz, so this part of the design is optimised at 621MHz.

So Z1 will be around 150 ohms, and what we want to do is to transform it to 600 ohms at Z2 so that when we parallel the bottom half of the array we get 300 ohms to feed to our balun. We do this by passing it through a lambda/4 section of 300 ohm transmission line (300 = sqrt(150 x 600)). Note that D2 is about lambda/4 at the frequencies you are interested in, but you need to get the impedance of the line to be about 300 ohms, and this is where your design seems to be way off.

There are many calculators out there to calculate the characteristic impedance of two-wire line, e.g.
https://hamwaves.com/zc.circular/en/index.html

and you find that for 300 ohm line where the center-center spacing of the conductors is 1.5in, you need a 1/4 in conductor diameter. Your wires look perhaps 10 awg, which at 1.5in gives over 400 ohms. This will transform your 150 ohms at Z1 to 400^2/150 = 1066 ohms, so the parallel combination of the upper and lower pair will give about 500 ohms at the junction.

So I'd suggest making the connections between the elements using 300 ohm line, which requires D/d = 6.15

Although the impedance of the upper transmission line (D1) seems less important (at the frequency where it is half a wavelength the impedance of this line doesn't matter at all), I suspect that making it a lower impedance will improve the broadband performance, I'd make it 300 ohms as well.

If I was designing this antenna using experimentation (cut and try) I'd be measuring the element impedances and mutuals so I could model the feed network - but you can't do that so all you can do is try to make sensible choices.

Also note that combining adjacent elements using a twisted feedline of length D1 works as the 180 degree phase reversal you get by passing along a half wavelength length of line is undone by the twist providing another phase reversal - hence the elements are added in phase. This is inherently a narrow-band process, fine for frequencies where D1 is around lambda/2. In your case you want to receive down to 80MHz where D1 is only about a 16th of a wavelength. This will really mess up your performance at these frequencies as twist in the feedline essentially adds the elements out of phase), probably costing you something like 10dB. You could probably improve performance at the lower end by altering the feed network, but only at the cost of complexity.

#### Attachments

• 130.1 KB Views: 1

#### gary350

##### Well-Known Member
I built a new 8 bay antenna today. I followed instructions in ARRL antenna book, I took my time and did a good job. ARRL page 17-21 fig 44 antenna should be 550 Mhz. About 90 minutes before dark I decided to hurry and get the wires on so I can test it and watch TV tonight, tomorrow I can do the wires like the book says. With antenna 7 ft high aimed at 329 degree and parallel to the ground field strength meter reads 97% on all channel expect channel 35 and 34. Channel 35 = 600 Mhz field strength meter reads 57% signal is good no flutter no fade out at all. 34 = 590 Mhz will not come in unless I rotate antenna 20 degrees left TVfool much be wrong channel 34 is not where it should be, signal is in the 60% range. Reflector screen is 10 inches from bow ties. Reflector screen is all I had no time to go buy a larger better screen until tomorrow. Signal is better than the old antenna. If I put correct wires on the antenna & larger screen tomorrow how can it get better than 97%. I still need to rescan to pick up channel 6 and figure out where channel 34 really is. Wood frame is pressure treated. All electrical wires & bow ties are on PVC or 1/2" thick plastic. Balun is 300 to 75

Last edited:

#### gary350

##### Well-Known Member
Question. If left side and right side are both 300 ohm with .186 diameter wire spaced 1.5" apart, when connected in parallel it becomes 150 ohms. Does wire spacing need to change for 150 ohms?