How to build a UHF antenna... - Page 16

I like that simple bowtie and the simple loop. They both did well when I tried them out. Atmospherics are distorting that list a bit though.

Net gain on the SBGH was calculated taking VSWR into effect. I take it that that particular station is behind some large obstruction, like a mountain ?


I always liked the simple bowtie much better than a loop. For one thing, you can attach it to a dowel and more twin-lead and move it around a room to find the hotspot.
Put it behind/in front of some cheap Pier 1 art deco metal sculpture or some of that decorative Home Depot aluminum sheet metal used for hiding radiators, and you have a piece of functional sculpture that probably would move up a couple of notches on your performance scale.

Yeah, I've got mountains/hills in the way My DTX9900 will read signal from them, up to about 6% on it's scale. Any suggestions on what to use for a reflector for a SBGH? How far behind the elements should I mount it?

Modeling has shown 95mm to 115mm reflector spacing to be the optimal range with 100mm being about the best.

For indoor use, because you dont have wind issues, any flat metal or screen or mesh 30 X 40 inches will work.
For outdoor use, a 30 X 40 inch grid screen with up to say a 3 X 6 inch mesh could be used. The smaller the mesh, the more the wind will catch it, but the better the forward gain.
Or you could build the colinear pair reflector. Link here : http://www.digitalhome.ca/ota/superantenna/design.htm
The yellow lines in the drawing are the reflector pairs.

If the VSWR losses are more than 10dB less than the difference between the receiver noise and the ambient noise, the VSWR loss may not have a significant effect.
VSWR degrades the ambient noise (celestial and man made) as much as it does the signal so unless you are receiver noise figure limited (which might be the case at UHF in distant rural areas) the receiver will often just bump up the gain and the impedence mismatch will have less effect than you would expect. Thus, in many situations raw gain is more important than net gain.

HDTVprimer writes that radio shack preamps have a really bad noise figure and can undo the effect of going to a larger antenna.

I still think a good quality preamp on the antenna is a good idea. It reduces noise picked up in the cable, a good preamp may have lower noise figure than a cheap receiver (though cheap receivers often do
better than you would think), it can reduce the impact of pass throughs and splitters, and some receivers may be gain limited. And receiver noise, amplified by higher gain, does add some noise, but not simple addition: sqrt(a^2+b^2) (linear). Thus if the two noise voltages are equal, you get a contribution of 3dB but if one is 20dB below the other, you get a contribution of 0.04dB. 3dB (power) down gives 1.76dB noise contribution, 6d down gives you 1dB of noise contribution and 12dB down gives you a 0.27dB.

noise contribution = 20*log(sqrt(1^2+(10^(x/20))^2))
where x is how many db lower than ambient noise the (amplified) receiver noise is.


At 600Mhz (compared to kTB at 290K)
Receiver noise: -3dB
Atmospheric noise: negligible above 50 mhz.
Galactic noise: <-10dB
Suburban man made noise: +6dB
Urban man made noise: +22dB
At 60Mhz (low VHF):
Receiver noise: -5dB
Atmospheric noise <<-10dB
Galactic noise: +10dB
Suburban: +28dB
Urban: +45dB

Galactic noise intersects receiver noise at 300Mhz. Receiver noise is 0dB at around 2.2Ghz. Atmospheric noise can exceed galactic noise below 20Mhz. Source Fig 29-3, Antenna engineering handbook, 3rd Ed, page 29-10.

If the difference between the receiver noise ambient noise is more 6dB above the impedance mismatch loss, VSWR doesn't matter.
Thus, VSWR mismatch, less than 20dB (power), is probably not an issue in an urban environment, could have as much as 3dB in suburban, and has a substantial effect in rural environments at UHF and doesn't amount to squat for low VHF in suburban/urban enviroments but can have a slight effect on rural ones. In an urban or suburban environment, an antenna with 10dBi VHF raw gain but -20dBi net gain could still work pretty well.

Yeah, and I would think the velocity factor of the coax limits the performance on that colinear antenna. An air dielectric would be better. Otherwise, it could be scaled considerably. Also, since phase errors are frequency dependent you are likely to end up with a narrow band.

I can attest to that. I have that cheapie Radio Shack 10db preamp, and even when it was working right, you really couldnt tell any amplification was going on at all. My new low noise Channel Master 0264 preamp (I needed 300ohm inputs and 75ohm out, otherwise I would have gotten the Titan series) is a whole different breed of animal in terms of performance, and only about $10 more than the cheapie RS at RS's inflated mall prices.


Yeah, thats the only reason to use a preamp, to compensate for the losses from the antenna to the TV. Also besides the .5-1.5db balun loss and minimum 3db splitter loss, RG6 losses are about 1db every 18ft per hdprimer.

Cool. The model I posted wasn't intended to be a close model of the DB2 for performance comparions so much as a model of what a do it yourself one along the same dimensions would be with a flat reflector, regular mesh, and no metal pole.

Preamp ALSO compensates for the tuner's 7-10 dB Noise Figure.

Hmm air, thats an interesting unique concept that I dont think Ive seen before.(or maybe I did) Basically it may be an improvement on a simple dipole. It would basically be : a wire (or tube) across for 1/2 wavelength, up for a small distance, then across for another 1/2 wavelength, then down for a small distance, then across again for another 1/2 wavelength. Then repeat multiple times for more gain. Ill have to try modeling that when I find the time. A four or eight bay pair of such dipoles would make for a very large uhf antenna, but that may be a big positive for people in tree enclosed areas.

I figured that if someone was looking for the dimensions maybe posting the DB2 output file may save some time.

Edit I was confused.. Hmm so you mean four dipoles like in a square?
sort of like the bowties eh? stacking and phasing. :P

Yeah, kind of like the Winegard four bay straight dipole bowtie type, but with the fancier longer elements. Im not sure it would even work, but in a lot of ways it seems theoretically reasonable.

Thanks for that, comparative plots like this are very useful. Unfortunately, antenna software often doesn't make it easy to do this.

Be careful about the 4 channel spacing. I have seen significant structure in UHF that can easily be missed at that spacing. But, I certainly undestand the motivation. If you look at HDTVprimer's comparison net gain plot, you can really see the wide spacing screwing things up on the VHF. Some antennas actually look ok on VHF on the full 2-80 plot but when you look at the higher resolution VHF plots you see they are pretty wildly fluctuating in VHF.

I find it surprising that 4" worked better than 4-1/2" spacing when, in theory, 5-1/4" spacing would be best. Antennas have a way of making their own rules, though. The issue is confused by the fact that you have bent forward elements and curved reflectors on the 4" models. A 4" initial spacing element bent forward 2" has an average spacing of 5" which could indicate that higher spacing is better even when the plots seem to be indicating the opposite. But maybe there is a phase shift going on that causes the performance to be best at a shorter spacing, though it does look on your plots like the reflector resonance may be too high in frequency. The tendency to drop off more sharply above the resonant frequency also can mean you want a resonance a little higher than the center of the band. Still most antennas are weakest at low UHF. Yours are looking better at low UHF than the commercial ones, including probably the PR-8800 (which is the best at low frequencies). And it looks like your 4-bays are outperforming the 8 bay equivalent DBGH in overall net gain over the post transition UHF channels. The move to longer whiskers probably has a lot to do with this.

Were those total gains or horizontally polarized ones? Shouldn't be a huge difference with these designs.

4-bays that are outperforming other 8-bays has some very practical application. 4-bays have lower wind load, can be rotated in confined spaces, and are easier to build. Also, 8-bay's are more susceptible to variations in field strength.

Good Work!

I usually do 2 channel spacing or less on UHF but since I've modeled alot of these antennas and know they are pretty stable on UHF, I went to 4 channel spacing to speed things up for that chart. On VHF-HI the UHF 4 bays are very unstable so I usually do 5 mhz max steps there.

I would have thought that 5"+ spacing would be better too!

There is a comparision between the 4" and 4 1/2" spaced flat reflectors and it acts as you would think 4" works better on the higher freqs. and 4 1/2" lower. I have ran models with larger spacings 5" + and it does enhance the lower channels but rolls off much quicker on the higher ones.

At 15 to 16" spacing with the curved reflector there is a range of frequencies around 600 - 700 mhz where the gain jumps and the beam width narrows. Must be getting a focused dish like effect, but unfortunately it's only over a small range of channels. I didn't put that plot on the chart because I didn't want to get too far away from the original purpose which was to compare the effects of different screens (wire fence)

I've modeled some with reflectors from 48" down to 30" wide in 2 " increments and the UHF gain just gets incrementally smaller across the band as the size goes down. I know that 30" is still well above what is needed for a 4 bay on UHF and maybe there is a sweet spot at a smaller size.

The main reason for the large reflector was for VHF-HI performance but using the curved extra large reflector appears to enhance the UHF too.

The gains plotted were the horizontal polarized ones at 0 degrees elevation free space.

Also thanks for some very informative posts sometimes I have to read them 3 times just to digest it all.

Very true. Putting that large screen on the four bay bowtie, pictured above, dramatically increased my reception on channel 17.


@mclapp,
Any testing done on your 6 bay model yet ?

And the feeder lines to the additional 2 bays, are they straight or crossed ?

Ok, I see it now, the cyan and dark green. Not a lot of difference but it does take gain from where there is a surplus and put it where there is a deficit. Sine wave addition is pretty tolerant of small phase errors then goes downhill rapidly for large ones. Phase errors are especially bad when they accumulate, such as when you have a crossed feedline and put more than 4 elements on it. If you were to build a vertical 8-bay with straight feedline between the middle elements and crossed feedlines, with 3" extra length, on the others you end up with a 180 degree phase shift in which 4 of the elements completely cancel each other out. I thought
I would calculate
Interestingly, if you pick the middle frequency (584Mhz) it is not the same as picking the middle wavelength (21.002" = 562Mhz) of the band. If we had picked the middle wavelength, error would have been higher at the higher frequency end of the band.

Now, the second time the signal hits the elements, it is weaker which reduces the effect some. And I am thinking that to explain the results, there must be a phase shift in the plane wave after the signal hits the first element.

The narrow frequency response is easy to explain. At one 1/4 wavelength spacing, you have 1/2 wave + a 180 degree phase shift (equivalent to 1/2 wave). At 3/4 wavelength spacing, you have 1.5 wavelengths + 1/2 wavelength from the reflection = 2 wavelengths. Thus a 1" change in wavelength is divided by 2 wavelengths instead of 1 and thus the phase error is double. This means it takes half the wavelength change to produce a 180 degree error and total cancellation or half as much to get whatever limit you set: -3dB, -6dB, etc. So, you expect the bandwidth to be about half. If you want to have fun, set the reflector spacing to 10 wavelengths and you should get a nice comb pattern, though weak and sinusoidal.

I used to work on Fabry-Perot etalons, very high resolution tunable interferometric filters in which light bounces back and forth between two almost perfectly parallel highly reflective mirrors. We used mirrors for infrared and meshes etched on silicon wafers for longer wavelength stuff.



Yeah, that is about what I would expect. I am using 32" wide right now because I figured that was wide enough that the reflector size wasn't substantially influencing results. A scientific wild a** guess (SWAG). My elements, and the feedline spacing, are a little wider than yours.

I have been avoiding the curved reflectors and swept forward elements for the time being. Harder to visualize what is going on and it would take quite a while to try the various permutations of element angle and reflector shape. I would think, though, that curving the reflector about the other (horizontal axis) would be better. Make a parabola with the whiskers at the focal point or a 90 degree corner reflector surrounding each element. As a bonus, you get structural strength. Corner reflectors with bowties have been around for a while. A single bowtie with a 25" wide and top and bottom (along the hypotenuse) corner reflector with 40 degree flare angle gets 9dBd at 500Mhz and almost 12dBd at 700Mhz and 12dBd at 900Mhz. However, you get a lot less gain when you reduce the vertical size of the reflectors: 3.5dBd at 500Mhz and 8dBd at 8 inch diagonal measurement which is about what you would have with the elements spaced 10" apart. Lower frequencies much more sensitive to size of the reflector. Those gains are relative to a resonant dipole. A single bowtie with corner reflector was 8dBd at channel 14 and 12.75dBd at channel 78 (add 2db for dBi). (Antenna Engineering Handbook); dimensions were a little vague but that was probably a 25"x50" reflector screen bent into a corner reflector 38 inches high. A 2-bay to 4-bay achieves much the same effect, and higher gain, by filling a flat reflector with multiple elements. But I do wonder what you would get by using four 10" high parabolic reflectors (or larger, with feedline adjustments). Parabolic antennas scale well.

Thanks for all the stimulating discussion.

Yeah, I checked the DTT7685x tuner (used in many CECBs) datasheet and it has a noise figure of 6dB typ, 8dB max. FM rejectin on CH06 was only 2-5dB. hdtvprimer
gives some noisefigures for different preamps at channel 30:
Channel master 7775 (discontinued) 2dB, winegard AP-2880 2.6dB, Radioshack 15-1109 3.6dB, 15-1170 6.0dB, 15-1108 7.0dB. Stark Electronic
gives noise figures of CM-7776 (disccontinued) 2.8dB, CM-7777 2.8dB, and CM-7778 3.0dB at an unspecified frequency. Denny's Antenna Service lists the winegard AP8700 as 2.8dB VHF/UHF, HDP 269 <3dB, AP8275 (UHF 2.8dB VHF 2.9dB, and AP 4800 UHF 2.7dB.
Todd Emslie lists the Wineguard AC-4990 as 2dB, RDX UA-900 (discontinued?) at 1dB, wineguard AP-4800 at 2dB, Televes MRD 2dB, Research communications 98030 0.5dB (expensive), Alcad BR-105 1.5dB.
Research Communications 9620 (UHF) 0.4dB and 9254 (UHF/VHF) 0.4dB ($320 with outdoor housing with waterproof cable glands).
And a google search for something like "TV preamplifier noise figure" turned up some homebrew preamps using a transistor with 0.5dB noise figure, though actual performance was not as good as 0.5dB.

1.5dB
and 0.6dB at 1296MHz but extends down to UHF TV and maybe VHF (50 ohm design). In sensitive applications, like radio astronomy, preamps are sometimes cryogenically cooled. The noise produced in electronic components is roughly proportional to temperature. Running your preamp at liquid helium temperatures (hard to do because of self heating, not to mention equipment that costs tens of thousands of dollars) gets you 37dB, and at liquid nitrogen temperatures 11.6dB.
Many components, however, do not function properly at cryogenic temperatures in addition to other practical issues, though Peltier coolers with 68 celsius degree temperature drop could give you 2.3dB (20*log10(225/293)) improvement, or 5.4dB with double stacking (but operating well below temperature rating for most components), while consuming a lot of power and lightening your wallet, provided you are still receiver noise limited. A 0.5dB noise figure amplifier is already acting as if it was at 40K equivalent temperature.

So, if you live in a rural environment where receiver noise might be the limiting factor then a good preamp might give you a 3dB improvement and a premium amp might give you 6dB improvement and a crappy amp no improvement. If you have a digital TV with a poor tuner, you might get another 3dB. Impractical draconian measures might possibly give more performance but once you exceed receiver noise figure (6-10dB) or ambient noise, you gain nothing.

First, use the best antenna you can. A preamp can't improve on the signal to noise ratio from the antenna, it can only prevent it from being further degraded.

I had also wondered why Bill gates had been able to do what IBM never did with email and the answer to me was that he invented something and applied it, such as the GUI interface and event driven programming enabling a full harnessing of the power of the computer in a relatively simple user friendly way.
I guess it would be imbossible to invent something with antenna engineering at least for me inventing the wheel was unesesarry but Ive had good luck with the bowtie and guess the 2 bay Hoverman can do very good above channell 14.
Certainly havent tried making a foil parabola, but since I like indoor stuff and shun lightning flashes on my appliances, maybe a foil refelctor with some sort of dipole in the center would have some gain, since it seems the more we increase the bowties in 8 bay configurations the more the loss.

Hey everybody, what do you recommend for a quad bowtie regarding whisker length and bowtie spacing center to center. My channels are between real 34 to 50 UHF.

Add to the list my Channel Master 0264DSB figures:
* Dual 300 Ohm VHF/UHF inputs and a single 75 Ohm output
Gain: 16 dB VHF / 23 dB UHF
Noise Figure: 3.0 dB VHF / 2.2 dB UHF

The only transistor I saw in it has the following markings:
(first line) PH68
(second line) BF G198


I googled, but couldnt find the specs for it.

I was looking at the reflector and for sure the full mesh style blocks signals from the back while the horizontal 2" reflector lets signals in. The problem is that you have to always rotate the antenna with a reflector on, and others on here for sure have just used it with no reflector 30 feet in the air which is OK. Indoors you can have the reflector separate for those hours of viewing when a target station is desirable and then easily removed.
So I'm designing this refelctor which is made from the 3/8" rods, preferably old VHF antenna aluminum pipe but in this case 26" curtain rods.
It slips over the pipe holes I origionally had on the refelctor support stubs. It can easily be slipped on or off as desired and is 14" behind the phase lines as suggested, but sometimes 16" back is nice too and more clear. At 5" back its just about as good. The rear piece of wood takes 12 drill holes 3" apart to to insert the rods.

That should work fine and the 14"-16" spacing or 4" - 5 1/2" spacing should both work well. The 14 -16" spacing enhances the VHF-HI.
In either case the smaller spacing (14" or 4") will enhance the higher UHF channels and larger spacings enhances the lower channels.

The design I posted earlier in this thread should work well in that range

http://www.avsforum.com/avs-vb/showpost.php?p=13920610

If your going to use a reflector attach it 4" behind the phase lines of the bowties.

If your looking for VHF-HI reception use a 30" wide reflector minimum to be safe

Since that style of curtain rod moves in and out kind of like rabbit ears, you may be able to use them to fine tune the reflector for that particular station you are watching at the time.

mclapp,
In those pictures in that link, what did you use for the element standoffs ? And any tips to building them ?

I think Im gonna build 9" whiskers and 9" bowtie spacing - center to center.

http://www.nxp.com/pip/BFG198_3.html

SOT223 8GHz (unity gain) NPN 1W 20V 100ma Hfe 40min 90typ (<30Mhz), power gain at 500Mhz 18dB typ, 15dB typ at 800Mhz. No noise specs. Circa 1995.

Datasheet includes schematic and PCB layout for an amplifier (not injector powered).

If the whiskers are fanned out forward, would that help it with tree leaf wind scatter? Or is it important to keep them in a flat plane for better performance?


And Ill be using copper wire as thick as I can find it. The thicker the better right, becasue the the increased surface area will help performance...as the the signals travel on the surface...correct?

Up to a point, about 1/20 to 1/10 wavelength max IIRC. 4 gauge solid is the biggest Lowes or HomeDepot has, 6 gauge is easier to work with. Use DogTs method of straightening the wire. Clamp strongly one end in vice and the other end in vice grips. Pull straight and hit vice grips with hammer hard multiple times.


Yep, thats the package. Heh, theyre overstating the specs. I knew it. But so far so good, its working well.

TVFools new 6/27/2008 update gave me the answer. All my digital stations are projecting Feb KW usage much higher than the 3/08/2008 update did, except for one. Some have more than quadrupled projected KW power usage. Do you think theyre listening to us.

That gives me a minimum 3 to 9 db plus more gain.