Canadian TV, Computing and Home Theatre Forums banner

1 - 20 of 26 Posts

·
Registered
Joined
·
111 Posts
Discussion Starter · #1 ·
Well, so far it's not a success story. Far from it I'd say.

I need help. How to make a proper 1/4 wave transformer?

I have been reading all along the best way to combine antennas is simply not using a combiner /splitter. From what I understand, you're lucky to get anything out of combining the two antennas if you use a splitter to combine them. Well, to me that's working better than my attempts at 1/4 wave transformers.



So as far as I understand, the most common now it's things are in order to combine 2 antennas for gain, you must have several important features present.

#1. The antennas must be exactly the same, from the same manufacturer, part number. Essentially the same production. Mine are 2 91XG's but were purchased 3 years apart. They do have some differences. After several balun, cable and reflector swaps, I said screw it and ordered 2 new 91 XG's to be delivered today. Test them today also.

#2. The coax cables must be the exact same length, the same kind for example : RG6 dual shield Etc... and should be the same manufacturer and part number. Might as well take it off the same spool. Mine are the same to a 1/8" and the same cable.

#3. Both antennas must be mounted at the same height but can be stacked over each other with the proper distance or side by side, where the separation isn't a big deal. Mine are side by side Booms are separated 50"

#4. Both antennas need to be aimed at exactly the same target in order to make the gain. Mine are.

#5. There's no point combining antennas for gain if you are going to have a lot of loss in the system. Combiners/splitters eat signal. So, this is where the quarter wave 50 ohm "matching" (stub) section comes in handy. OK, this is my problem.

So here's my saga. Get a beer, maybe some popcorn relax and read on.

I want a virtually lossless array. So, I did my research, talked to a few people on some other forums and found out that either RG 8 or RG58 can be used to create these "matching transformer" sections. As long as it's 50 ohm. A friend of mine just happened to have a roll of foam filled RG8 and has donated several lengths of RG8 that I thought I would never need. Honestly, for my new arrays, one VHF pair and one UHF pair I figured that would be enough. However it didn't turn out that way.

I had first read that quarter wave matching section for UHF made from RG 8 should be 10 inches. Including connectors. Okay. No problem, I butchered a combiner, got the cover off tore all the guts out and ran my two equal antenna leads into the sides stripped the RG 8 and somehow got it to slip into a RG 11 coax connector. (I drilled it). I then used some solder to join the three together inside the combiner. It was my first time so it took me sometime, but it seemingly turned out fine.



So, I climb the 22 foot ladder. I install my new piece, I'm excited! I carefully measure the spacing and aim my 91XG's. There are no leaves on the trees yet. I check all my connections, climb down the ladder and run a scan. What? How can this be. I run the scan again, I climb back up the ladder I take everything back apart I retighten. Nope.

I find out later that I made my matching section too short. 10" is wrong for foam filled RG 8. That's for polyethylene filled RG 8. I didn't know there was two kinds. By doing the calculations, my trusted friend Rabbit figures out for me that it should be 12" because the velocity Factor is different. Foam filled is slightly "faster" then the other. No problem, if that's all I need to do I'll make another section. This time I decided not to play around with combiner boxes I figured out a way to make an RG8 cable work, look and install just like a piece of RG 11. I use a "T" connector.





I make my 12" RG 8 matching section. Back up the ladder, I need the exercise anyhow. I attach it run down in my basement and run my scans. Another disaster. I recheck all my connections, I'm back up the ladder several times trying to find out what the problem is. This just isn't working.

Okay, brain fart. Guess what I forgot? I forgot to tell the guys it was UHF that I would be combining when I got the measurements? I assumed the matching section would be the same.. Originally, the measurements I got were for a VHF matching section. However, combining my VHFs didn't go so smoothly either. Another story. Maybe that'll be another post for another day. Or I'll incorporated into this thread.

Since I didn't know the matching sections would be different between UHF and VHF, my mistake and because I had done the exact same thing with my VHF a few weeks earlier with no success and I had several different lengths of matching sections between 9 and 1/2 in + 13 and 1/2 in. None of it worked.



So, I find out that for UHF, its 3 and 1/2 inch RG 8 matching section if it's polyethylene insulated but it's 4.6" if it's foam filled. I'm thinking to myself, no wonder I had such poor results! My matching sections were three times longer than they needed to be!

I make a new matching section stub. Doesn't work, so I make another a little longer.



I climb the ladder, I'm pumped! I take the 12" matching section off of my Kitz KT200 amp. By the way, I don't know if you've ever taken a coax out of a Kitz Tech amp but it's not easy with your fingertips. Especially RG 11. It's an awesome designed Pre-amp but it takes some time to master reaching inside of it and starting the RG11 cable connectors especially when these are the Belden Aqua Tights. I've literally done this 50 times in the last 2 weeks with different testing I'm still not good at it. And you think I would have bent the 7/16" wrench by now.



Okay, where was I? Oh yeah, I put my freshly made 4.75" matching section in, checked all my connections twice, ran down in my basement and ran the scans again. Disaster strikes again.

After several private messages to different friends, it was suggested that maybe, just maybe I'm not getting the shielding into the connector correctly, yes I have been trimming shielding back because the connector goes deep under the skin of the RG 8 anyhow where I'm sure the connector makes contact with the shielding. No problem so here it is Memorial Day. Since it started out raining this morning I decided I would make one more attempt at making a 4.66" matching section.

I mentioned this before, but even before I made my first matching section I was told to include the length of the connectors. However it was never fully understood that I was adapting an RG 11 connector to the RG8. You know, inside the RG11 connectors there's someone in pain that forms the actual contact whatever you're screwing the cable onto. So I figured I would outsmart myself and make this one longer. Because really, isn't it the function of RG8 length that matters?

So this time, I decided to make the cable itself 4 and 1/2 inches long and then solder on the contact tips from the RG11 connectors. Now, the overall length is 6.5" .




However, this didn't work either. I'm at a loss.. Could it be the "T" connectors? I've used several new ones. This was one I'd made up for the VHF, it's 12":





It's not the cables and although antennas aren't really balanced, noted during previous testing a simple Holland GHS Pro splitter outperforms any type of matching wave section I have conjured up.





I'll retest today, I'll try and match the 2 new 91XGs to each other. If that's a go I will combine them.
 

·
Registered
Joined
·
1,279 Posts
The home made splitter that you made will not work! A T fitting will not work. Both will mess up the impedance of the coax. When all is in proper phase, identical antennas pointed in same direction, and combining coax identical length and same velocity factors, them a splitter/combiner will actually add the 2 antenna signals and create a gain rather that a loss!
 

·
Registered
Joined
·
111 Posts
Discussion Starter · #3 ·
The home made splitter that you made will not work! A T fitting will not work. Both will mess up the impedance of the coax. When all is in proper phase, identical antennas pointed in same direction, and combining coax identical length and same velocity factors, them a splitter/combiner will actually add the 2 antenna signals and create a gain rather that a loss!
Thanks for your reply.
So what is your recommendation for a no loss/ low loss combiner. ? A splitter will cause some loss.
That's why I'm experimenting with 1/4 wave matching.

In my testing, the single 91XG outperformed the Combined 91XGs :
 

·
Registered
Joined
·
5,062 Posts
Your fundamental problem is that a quarter wave anything (transformer, stub, whatever) is only a quarter wavelength at one frequency. Your antennas are broadband. With identical broadband antennas you need to combine them with combiners that are broadband also, using jumpers that are of identical electrical length. For the reasons Jorgek pointed out. I see you are stacking Horizontally.
If you have the vertical space (a tower??), I wonder how much better you'd make out stacking them vertically? That way you'd maintain the same Horizontal beamwidth as one antenna.

Stacking guideline for identical antennas:
Horizontal Stacking: Horiz. BW narrows by roughly 1/2, while vertical BW remains the same. Might be more difficult to aim azimuth wise, but is usually easier for most people to accomplish.
Vertical Stacking: Vertical BW narrows by roughly 1/2, while Horizontal BW remains the same. Makes it easier to aim azimuth wise, but wanna ensure your vertical beam is aimed at the Horizon (eg antennas are plumb, without up or down tilt). Most people probably don't have the vertical space necessary to safely accomplish (eg wind load, etc).

In both cases, you need to use electrically identical jumpers to connect to your combiner.

Edit: we need to be careful when reading about techniques commonly used in Amateur Radio, because what works for them is only possible due to the extremely small bandwidth frequency bands they are dealing with. I'll bet they wish they had 100 MHz to 200 MHz + of UHF spectrum to play around in...
 

·
Registered
Joined
·
1,279 Posts
I'm also a ham radio operator for over 40 years. Majortom hit the nail on the head.
 
  • Like
Reactions: Chiwaukee

·
Registered
Joined
·
111 Posts
Discussion Starter · #6 ·
I'm also a ham radio operator for over 40 years. Majortom hit the nail on the head.
I don't doubt your skills or your experience I hope I didn't come off the wrong way. Actually that's what I'm looking for, I was looking for suggestions. I'm told that combiners will have some loss.

It seems that way when I run a scan of the combined antennas versus a single antenna. I was using a Holland GHS Pro. So I probably have some sort of issue even though the jumpers are exactly the same length and made from the same spool of cable.




 

·
Registered
Joined
·
111 Posts
Discussion Starter · #7 ·
Your fundamental problem is that a quarter wave anything (transformer, stub, whatever) is only a quarter wavelength at one frequency. Your antennas are broadband. With identical broadband antennas you need to combine them with combiners that are broadband also, using jumpers that are of identical electrical length. For the reasons Jorgek pointed out. I see you are stacking Horizontally.
If you have the vertical space (a tower??), I wonder how much better you'd make out stacking them vertically? That way you'd maintain the same Horizontal beamwidth as one antenna.

Stacking guideline for identical antennas:
Horizontal Stacking: Horiz. BW narrows by roughly 1/2, while vertical BW remains the same. Might be more difficult to aim azimuth wise, but is usually easier for most people to accomplish.
Vertical Stacking: Vertical BW narrows by roughly 1/2, while Horizontal BW remains the same. Makes it easier to aim azimuth wise, but wanna ensure your vertical beam is aimed at the Horizon (eg antennas are plumb, without up or down tilt). Most people probably don't have the vertical space necessary to safely accomplish (eg wind load, etc).

In both cases, you need to use electrically identical jumpers to connect to your combiner.

Edit: we need to be careful when reading about techniques commonly used in Amateur Radio, because what works for them is only possible due to the extremely small bandwidth frequency bands they are dealing with. I'll bet they wish they had 100 MHz to 200 MHz + of UHF spectrum to play around in...
Major Tom thank you for taking the time to enlighten me. I may just try to combined them one over the other. Do you know what the spacing should be one above the other from boom to Boom?
 

·
Registered
Joined
·
5,062 Posts

·
Registered
Joined
·
5,062 Posts
BTW, This is what will happen when Horizontally stacked, Aimed in the same direction, but the two antennas are fed to the combiner 180 degrees out of phase (eg if one of the two baluns has their leads flipped). A guy could use that to his advantage if he had two distant markets to aim for in roughly those directions and strong local stations right in the main boresight of the antenna that you wouldn't mind being in the null.
11289
 

·
Registered
Joined
·
111 Posts
Discussion Starter · #10 ·
Try using this stacking distance calculator. All you need to know is the Horizontal and Vertical beamwidth of the identical antennas, and I would use at the lowest frequency of operation, 470 MHz. This will spit out both a suggested Vertical and Horizontal stacking distance.

Very nice, I'll search around for the beam width of my 91XG's and my 30-2476s.
Or do you know where to find it?

I've seen that "two antenna trick" before on the HD primer page. Very interesting also. Thank you so much for taking the interest in my project.

In your mind what is the best way to combine two identical TV antennas ?
 

·
Registered
Joined
·
5,062 Posts
Download the specs for your 91xg antennas straight from antennas direct.
They are pretty comprehensive.
The Stellar Labs antenna specs are also right there at Newark where you presumably bought them. Just download it.
Simply read the specs, plug in a few figures. and have at it... give it a whirl in the field, then see how you make out.
 

·
Registered
Joined
·
111 Posts
Discussion Starter · #12 · (Edited)
Well here it is a full 10 months later and I can say I've had successes. Actually the success came much earlier but I haven't posted it. It was a busy summer....

11921


As it turns out the main problem was antenna alignment. Being way up on the ladder and looking down wasn't the best way to assure alignment. However, spotting it from Below was key. So I made an alignment tool. Once the front of the booms were spaced evenly and the two booms were parallel, everything worked out perfect. I would say at the front of the antenna the alignment was a couple inches off.
11922


11923


11924


11925


I then played with different lengths on the 50 ohm Transformer sections and the UHF and VHF. The UHF sections are much shorter:
11926


11927



11928



There was an improvement over just using a good quality combiner / splitter:

11932


Success! I want to give some credit to Calaveras over on AVS and to Rabbit here on Canadian home digital and AVS.
 

·
Registered
Joined
·
406 Posts
I personally consider all this splitter stuff from shops as bullshit (from microwave point of view. It work perfect for what it is inteded - mechanically splitting ultra strong signals after LNA, in cable networks). They have no predictable S-parameters.

Also I buy, disassemble, make 3D drawings (simulate in HFSS), and measure with vector impedance dozens of antennas, including UHF TV.

Many many industrial TV antennas (including 91XG) use fake balun. It is typically lambda/2 loop for 800 MHz
First photo is 91XG
At frequencies other-than-800 MHz it neither produce 75 Ohm output, neither block common mode currents (coax outer shield work as antenna, distorting radiation pattern and introducing more mismatch)
20200512_184426_copy_768x1024.jpg 1141919_original.jpg

I replaced this hoax balun with proper Guanella balun, and VSWR measurements improved dramatically.

As for proper combiner/splitter, I would make PCB Wilkinson bridge. Or T-joint + L/4 transformer, but Wilkinson is better in many ways:
1) more wideband
2) more tolerant to non-identical arms (tolerate phase lag and non-equal impedances)

If you have large enough metal box, it is easy to make Wilkinson using just single copper wire in the air, soldered to 3x F-connectors

I can help with either design (PCB or metal box)

Also, I don't like your cable management along Yagi boom. It doesn't look symmetrical. Should be exactly along boom.

BTW, here are my investigations on perfect Guanella 4:1 design:
(was discussed on this forum earlier)

Nowdways, PCB printing is very cheap and fast, if you don't like DIY PCB at home.
 

·
Registered
Joined
·
406 Posts
Here is more on Wilkinson.
It covers 470-692 MHz UHF TV with amazing VSWR < 1.23


To manufacture it, you need zinc-plated П-shaped ground.
Length will be ~10", width ~1.5"-2"
AWG11 copper wire need to be placed h=2.2 mm over ground to become Zo=106 Ohm transmission line.
Length of each arm is 129 mm for f=580 MHz central frequency

At one side of П-shaped box screw F-connector
At other side screw 2 x F-connectors as close as possible.
Small distance is needed for R=100 Ohm SMD resistor

For fully-covered box, gap wll be slightly different and depends on chosen box height.
 

·
Registered
Joined
·
111 Posts
Discussion Starter · #15 ·
I personally consider all this splitter stuff from shops as bullshit (from microwave point of view. It work perfect for what it is inteded - mechanically splitting ultra strong signals after LNA, in cable networks). They have no predictable S-parameters.

Also I buy, disassemble, make 3D drawings (simulate in HFSS), and measure with vector impedance dozens of antennas, including UHF TV.

Many many industrial TV antennas (including 91XG) use fake balun. It is typically lambda/2 loop for 800 MHz
First photo is 91XG
At frequencies other-than-800 MHz it neither produce 75 Ohm output, neither block common mode currents (coax outer shield work as antenna, distorting radiation pattern and introducing more mismatch)
View attachment 11936 View attachment 11937

I replaced this hoax balun with proper Guanella balun, and VSWR measurements improved dramatically.

As for proper combiner/splitter, I would make PCB Wilkinson bridge. Or T-joint + L/4 transformer, but Wilkinson is better in many ways:
1) more wideband
2) more tolerant to non-identical arms (tolerate phase lag and non-equal impedances)

If you have large enough metal box, it is easy to make Wilkinson using just single copper wire in the air, soldered to 3x F-connectors

I can help with either design (PCB or metal box)

Also, I don't like your cable management along Yagi boom. It doesn't look symmetrical. Should be exactly along boom.

BTW, here are my investigations on perfect Guanella 4:1 design:
(was discussed on this forum earlier)

Nowdways, PCB printing is very cheap and fast, if you don't like DIY PCB at home.
Yurii have to ask you what do you mean my cable management along the Boom? You don't think the coax should be fastened to the Boom?

If you're talking about the antenna cable said they were perfectly measured and I made them myself out of the same exact coax cable.

I am interested in your Wilkinson power divider. I've heard of it but I don't really have a concept of how it works. I would be interested in trying that.
 

·
Registered
Joined
·
406 Posts
In simple T-joint you add 2x50 Ohm = 25 Ohm
and then using 25*sqrt(2) = 35.355 Ohm transformer raise impedance back to 50 Ohm
In Wilkinson, you first raise impedance from 50 to 100 Ohm using 50*sqrt(2) = 70.711 Ohm transmission line. Then combine 2*100=50 Ohm

Impedance of transmission line in Wilkinson is twice of simple L/4 transformer. Usually it is easier to manufacture mid-range Zo line, compare to low-imped line.
In lossy (PCB, high-freq, thin wires) transformers, doubling voltage decrease current twice, resistive losses decrease 4x times.

In simple L/4 Divided ports are soldered together in parallel. There is no isolation between ports. Any magnitude/phase imbalance in T-joint introduce reflection back to the source (RX antenna).
In Wilkinson, resistor dissipate out-of-phase energy imbalance.

As for cable management, I don't think cables should hang around assymetrically to boom
11951
 

·
Registered
Joined
·
111 Posts
Discussion Starter · #17 · (Edited)
Very good Rabbit !

I'm taking this project in baby steps. As you know, maybe we haven't settled on the exact length of the directors yet but I did start the groundwork. Also for anybody who's reading this I am not destroying two perfectly good 91XGs. They were damaged in a storm Beyond repair. I don't even have enough X elements in plastic brackets to do what I have to do to so I'll have to steal some from a brand-new unassembled 91XG kit.


3123987



I already have my son involved. I'm teaching him life skills, how to work with his hands. He was having a little hard time sniffing the tips of the X elements so I had him filing the burrs.






In simple T-joint you add 2x50 Ohm = 25 Ohm
and then using 25*sqrt(2) = 35.355 Ohm transformer raise impedance back to 50 Ohm
In Wilkinson, you first raise impedance from 50 to 100 Ohm using 50*sqrt(2) = 70.711 Ohm transmission line. Then combine 2*100=50 Ohm

Impedance of transmission line in Wilkinson is twice of simple L/4 transformer. Usually it is easier to manufacture mid-range Zo line, compare to low-imped line.
In lossy (PCB, high-freq, thin wires) transformers, doubling voltage decrease current twice, resistive losses decrease 4x times.

In simple L/4 Divided ports are soldered together in parallel. There is no isolation between ports. Any magnitude/phase imbalance in T-joint introduce reflection back to the source (RX antenna).
In Wilkinson, resistor dissipate out-of-phase energy imbalance.

As for cable management, I don't think cables should hang around assymetrically to boom
View attachment 11951
I see what you're talking about those were early pictures before I even perfected my quarter wave Transformers stubs.

Here's a picture after I had everything complete you can see virtually no cables hanging.

I know you're not supposed to but I have used black UV resistant zip ties to keep the RG6 coax cables tight to the booms on all four antennas.

11965
 

·
Registered
Joined
·
1,279 Posts
Please excuse my ignorance. But why are you using 50 ohm coax for the stubs in a 75 ohm system?
 

·
Registered
Joined
·
111 Posts
Discussion Starter · #20 · (Edited)
Please excuse my ignorance. But why are you using 50 ohm coax for the stubs in a 75 ohm system?
Sorry, I've been working on another project. The 50 ohm coax is your Transformer. Because you can't just add two 75 ohm signals together you need to somehow transform the signal so that the 50 ohm section does that for you. It also needs to be tuned in length. UHF sections are much shorter than the VHF sections shown here:

20200525_113605.jpg


Anything more beyond that is all theory which I can't really answer. I think Yuri tried answer it but he shows some math to help explain it.

Many of us don't know all the theories or can even do the calculations we just wanted to experiment and make it work without knowing all the theory behind it. At least that's me! I just like using the technique improving that it's a good technique. Explaining why, I'll refer that to the experts here.
 
1 - 20 of 26 Posts
Top