Building an 8 bay from two 4 bay’s
 

I have two winegard 4400 4 bay antennas. Currently I have one on a mast at my house. I have multipath problems that the directivity of an 8 bay would probably help. I want to bolt them together with angle iron pointing the same direction.

My plan is to use 450 ohm window line to create quarter wave impedance transformers, roughly taking the 300 ohm terminals on the antennas and giving me a 600 ohm point. Then I will parallel the two 600 ohm points back to 300 ohms and use a balun to send the signal to my preamp.

My first question is what frequency to cut the quarter segments for.

At my house, the Baltimore UHF stations come booming in, so efficiency on their channels is not critical. If I had to pick the DC stations that most mattered to me, it would be on real 34,35 and 36. At the same time, there are stations from the DC direction as low as real 27 and as high as real 50 I would like to get.

If I wanted to make an impedance transformer assembly focused on 34-36, would I cut it for 599 MHz?
Is wanting to make quarter wave segments that would work from real 27 to real 50 asking for too much? That would be centered at 620 MHz.

The supplied velocity factor for the 450 ohm line I have is 91%. (I do not have the equipment to measure the velocity factor myself). If I did my math correctly for a 620 MHz quarter wavelength section, I would need to use 4.3” (4 5/16”) lengths of 450 ohm line.

The antennas are too wide to use the 4.3” segments of line to join the two. I have some Radio Shack 300 ohm twin lead (the kind with the thick black outer covering and the foam core). Can I use equal lengths of that to connect the quarter wave 450 ohm line segments to the antennas? Would I want the lengths of the 300 ohm line to be some specific length, such as 1 wavelength at the frequency of my 450 ohm transformers?

I would end up with:
Antenna 1-300 ohm line-450 ohm line-balun-450 ohm line-300 ohm line-antenna 2
(The two 450 ohm line segments would be connected in parallel to the balanced side of the balun).

I could buy a DB8e or some other 8 bay antenna, but I have the 2 4400’s, the 450 ohm line, the 300 ohm line, and balun already. All I would need to buy is some angle iron to hold the two 4400’s next to each other and parallel.

Without test equipment, is there anything I can do to optimize lengths of lines, spacing’s, etc?

Am I making this too complicated? Would I get similar results just taking the two antennas, putting a balun on each of them, and connecting them to a splitter in reverse?

Thank you for your help.

(Location details are in http://forum.tvfool.com/showthread.php?t=13197 suggestions from that thread were noted, but there is a money problem preventing me from buying a DB8e.)

Start long, about 5.5" per 450Ω 1/4 wave section... experiment. Also start with each 300Ω link cut for about 1/2 wavelength in the middle of the band (around CH-33). Keep the twin lead stable... and both sides the same distance from metal. PVC pipe can be used to fabricate standoffs.

I have an HD8800 that's not in service at the moment. Give me a couple of days and I'll get some photos & dimensions of the combiner/phasing lines. They're just open parallel wire...

HD8800 phasing harness
 

Here is a drawing based on my HD8800...
http://forum.tvfool.com/attachment.p...1&d=1377822402

You may already have this, but here is the manufactures technical sheet: http://www.winegarddirect.com/pdf/spec_PR8800.pdf

Would a ¾ wavelength section of 450 ohm lead perform the same impedance transformation as a ¼ wavelength section? If I could just use 450 ohm lead, it would simplify the harness construction.

I am not looking forward to soldering 300 ohm lead to 450 ohm lead. I have all the hardware to put the two 4400’s together, and I have all the wire to do it. Right now, I am just waiting on time to get the installed 4400 down, and to bolt the two of them together.

No. I believe you would effectively be inserting three step up transformations in series.

3/4 Wave Length Transformation
 

Both seem correct.

That is, a 3/4 wavelength could transform similar to a 1/4 wavelength section. And yes, 3 transformations is one way of looking at it.

Roughly,

First 1/4 wave transforms up or down depending on whether the line impedance (Z) is above or below the input impedance. Then, the 2nd 1/4 wave transforms back to original input impedance.

Two, 1/4 wavelength lines in series equals one 1/2 wavelength line.

Recall, 1/2 wavelength provides no transformation regardless of Z of the transmission line.

The 3rd 1/4 wavelength starts over again.

There is a bit more:

A 1/4 wavelength line is only 1/4 wavelength at one frequency. The transformation becomes more complex when the frequency is above or below 1/4 wavelength frequency. Also, if the in/out impedance is not purely resistive, the transformation is not as straightforward.

The preceding constraints give rise to this:

Three 1/4 lines in series have a bandwidth that is less than a single 1/4 line.

Bandwidth continues to shrink if the line is lengthened further (5/4 wavelength etc.).

I read about 8 bay antennas and the flaws in the 8800 at hdtvprimer.com
Poor feed 8800 http://www.hdtvprimer.com/ANTENNAS/w8800.html
8-bay discussion http://www.hdtvprimer.com/ANTENNAS/TemporaryPage.html

I decided to try what HDTV Primer suggested and used two baluns, 2 equal lengths of coax, and a combiner to gang the 4 bays into an 8 bay rather than trying to make an impedance matching feed.

I used channel master baluns, 3’ lengths of coax, and a 1-2 splitter used in reverse as a 2-1 combiner. The result is higher signal quality, presumably from less multipath, and reliable reception of DC stations.

If I have some time in the future I may work on a feed system with twin lead, but for now what I have works quite well.