Monday, February 25, 2013

Quad vs. Yagi (part 2)

 

In my previous post I was discussing about absolute Quad vs. Yagi performance figures. I think it is clear from the analysis that Quad does not beat a three element yagi when installed over real ground. Now.. why does quad is claimed to work so much better? Oftentimes I hear claims that Quad is more tolerant to construction errors. So lets have a look…

Quad Behavior, Construction errors

For reference the original design in free space below.

image

Impact of boom Length

Boom + 30cm Boom +60cm
image

image

Boom -30cm Boom -60cm
image image

Lengthening  the boom has very little impact, but making the boom too short in such tightly coupled design starts disrupting the front/back. The gain figures remain fairly constant and within 0,2dB range.

Element length errors.

This quad design is based on HRS model, where the radiator is fed with ladder line and tuner is required. So I keep the radiator same and modify the resonant reflectors.

Reflector Loop -25cm Reflector Loop -50cm
image image
Reflector loop +25cm Reflector Loop +50cm
image image

So the reflector loop measurements significantly more critical than the boom length. The used 25cm measurement error sound fairly large, but trust me, it is very easy to make if you do not for example use wrong velocity factor when you design the antenna.

Yagi Behavior, construction errors

Again, the original design in free space for reference.

image

Boom Length Construction Errors

Director Placement

Director +30cm Director +60cm
image image
Director +30 Reflector -30 Director +60, Reflector -60
image image

As expected when the length of the boom is increased the gain increases.

Lets next make the boom shorter and see simulate what happens.

Director -30cm Director -60cm
image image
Director –30cm, Reflector +30cm Director –60, Reflector +60
image image

I would call this a very robust design. In normal construction tolerance the placement in the boom has only few dB difference. So as long as reasonable common sense is used in construction tolerance yagi is no worse than quad.

Antenna Tolerance To Element Length Errors

Lets first start shortening the elements.

Director -20 Director -40
image image
Director –20, Reflector -20 Director –40, Reflector -40
image image

And now lets make elements a bit longer

Director +20 director +40
image image

As expected lengthening the director turns it quickly a reflector disrupting the pattern completely. It depends completely on the yagi design how critical this measurement is. In this design the tolerance is reasonably small, so you better learn to measure or at least fail on the short side.

Reflector +20 Reflector +40
image image

Unlike the director the reflector is reasonably tolerant to impact of shortening the element.

Summary of Construction Error Tolerance

The myth: “Quad is more tolerant to construction errors than Yagi”.

I would conclude based on these two antennas compared side by side that there is no truth behind the claims for quad being more tolerant to errors than yagi.

Of course by making yagi director too long it is reasonably easy to disrupt the pattern completely, and even reverse the beam, but that is an extreme case. Both designs are reasonably tolerant and if you can measure with accuracy of 5cm and you have the material velocity factors, including the material thickness taken into account when designing the antennas, the measurements are not that critical. Accuracy of few centimeters is well in the limits of tolerance and capability of average builder.

I think the main reason for people considering quad being more tolerant is the number of bad yagi recipes out there and builders using just wrong materials. Good yagi design is also tolerant to tapering and material thickness variation. I checked this yagi model with –50% and +100% material thickness with little impact to pattern.

Part 3 and final conclusions still to follow Smile

Tuesday, February 19, 2013

Quad antenna myths: Quad performance vs. Yagi (part 1)

I am not a big believer in published gain figures of antennas. Marketing superlatives claim astronomical gain figures with unlimited front/back ratios. One of the claims, which I have heard and read on multiple occasions is the performance of quad antenna versus yagi.

Earlier I was referring to my HRSQ listening performance against a yagi antenna, but let us have a view what to expect when doing on-the-air comparison tests between the three element yagi and HRS quad.

As the baseline for my HRS quad development I used groundbreaking work of dk7zb. I built the HRS quad model based on dk7zb quad, and modified the design for material selections I was using. I further improved the model to take material and feed losses into account to optimize the tower setup.

From the my previous it may have already become already clear that I am a big believer in computer modeling. Having said that, it is important to understand the drawbacks and potential problems of various modeling software.

The following models can be compared, but the real life performance due to material losses, nearby structures, etc. can be significantly different i.e. less than the model indicates.

Please note on my dB practice: All gain figures are dBi, front/back etc. absolute figures are dB. When ground reflection is included in dBi figures I try to note that clearly.

HRSQ vs. 3 element Yagi

Myth: “Two element quad beats three element yagi”.

I have described the HRSQ, there is a way to improve the performance on 20m, but for this comparison the radiator is one lambda length on 20m band. But how about the yagi? First thing we have to set is to define the yagi to be compared against HRSQ.

The most significant factor of yagi performance is the length of the boom, not the number of elements as many marketing guys want you to believe. A typical 3 element yagi for 20m band in my opinion would be about 0,4 lambda long. There are similar monobanders available from several manufacturers and they are reasonably easy to construct. In my model I used three element yagi with 5.6 meter boomlength.

Free Space Performance

Elevation

image image

Free space elevation patterns are very similar. Yagi is 0,3dB better in gain, but quad is 3dB better in front to back performance. 3dB beam angle is very close to each other.

Azimuth plot

imageimage

Since the azimuth is the different projection of the elevation discussed above the gain and f/b are the same. The horizontal 3dB beam angle is a bit better (5 degrees) on HRSQ than with yagi.

Summary

Free air performance of two antennas is very close to each other.

But.. so far I have never seen HF antenna being in use in free space. There is always ground reflection and the antenna height to be accounted for. For the next model I used Eznec standard average ground, and placed both antennas at 10 meters above the ground. For the HRSQ the 10 meter point was the center of the quad structure.

Performance With Ground

Elevation patterns

imageimage

Now we start seeing a little more differences between the antennas. The gain for yagi is 0,6dB better, the f/b is 3dB better, and the beamwidth is also slightly better, as the free space comparison would indicate.

Azimuth patterns

imageimage

Here yagi wins with 4dB f/b, 6dB f/s. The only area where quad is better is 3dB beamwidth winning that by slight 7 degree margin.

Summary of part 1

Both antennas are solid performers. There are no grounds for the claim that the two element quad would beat well designed three element yagi. Free space figures would suggest that quad would beat yagi on all of the areas (f/b, f/s, horizontal 3dB beamwidth) but when modeled above real ground the yagi is a solid performer beating quad in all of the figures except horizontal beamwidth, where quad is 7 degrees better.

But that is not the end of the comparison, part 2 will follow… with some analysis on antenna bandwidth, construction accuracy requirements, and final notes.

Monday, February 18, 2013

Quad vs. Yagi Case Noise Buildup

HRSQ number one up!

I am stacking total of three of these, that is why this is number one :)

I have heard many times that Quad structure has less rain induced static buildup than yagis. This was not the primary reason for selecting this structure, so lets call it a bonus.

The picture below was taken in the afternoon. No it is not dark, this is normal daylight up here. I do not remember when I saw sunlight last time. Must have been weeks ago..


I raised up my HRS quad (HRSQ) a couple of days ago. Building and rising quad structures solo is always interesting challenge.. The boom height is about 9 meters. I have been testing the front / back figures and comparing the measured on-air figures to antenna model. So far everything is maching the model quite nicely.

Noise buildup

Today I had a chance to compare noise levels between hrsq and yagi in first hand, or first ear.

I was checking 9U4U pedition on 12 meters band and noted a significant static buildup in my yagi when it was snowing. I tuned my hrsq to 12 and alas, no static whatsoever. The gain figures are comparable between my yagi and single cell hrsq, verified with on air performance comparison a couple of days earlier.

I tried to visualise the static with hdsdr and kx3 but the pictures were not too good. I decided to give cw skimmer a go and that was much more successful. Unfortunately, when I got cw skimmer up and running the band was already dead. The following is unaltered screenshot (except the added text) from cw skimmer. Both antennas pointing to the same direction side by side to minimise possible coupling. The hrsq tuned to swr=1.0 with S-match tuner, rig KX3, i/q out to Creative E-MU 0204. I was swithing between antennas manually while recording the picture.


Quite interesting, isn't it? The yagi noise was over two S-units measured with KX3.

Thursday, February 7, 2013

Tower Design Criteria


 

"Someone still stupid enough to build quads".


First of all, It is not a quad. It is a copycat of professional grade broadcast HRS antenna. Well, ok.. yes it looks like a quad, mechanical challenges are very similar to quads, and behaves almost like a quad with some interesting additional properties.

I wrote earlier that I am planning to prove that quads can be built to survive the Finnish winter. I was discussing with a fellow HAM yesterday and he was really suprised to find that somebody is still planning to build quads after so many documented failures. Well.. it is not a quad :)

So I think my plan calls for an explanation.

Background

I started planning this tower project in fall 2011.

I constructed a 2 element beam for 40m band to my current tower. Since the tower is only 15 meters high the performance was clearly sub-par, but I wanted to try it out. The antenna was very nice for short skip contacts, but for DX it was a miserable failure. Further, the 40 beam was way too close to my 20m beam causing distruction in the 20 pattern so I had to make the hard choise to take the 40m antenna down. Never regretted that decision.

I started a lengthy study and computer modelling excercise before the decision was made regarding the tower setup.



I have done literally hundreds of hours of EZNEC modelling of all kinds of beams during past 14 moths. I have consulted people who I consider to be the best in the field in OH-land (people behind the radio Arcala monsters). I have read thousands of pages of material from reference books and from various Internet sources.

Further, I have visited fellow hams and discussed with elmers behind dozens of successful tower setups. They have x-mas trees full of 5 el beams, stacks of 6 el monobanders, logperiodic stacks, phased wires, etc. I put all of those through the EZNEC in order to seach the best possible combination for my criteria. I have made many new friends during this project and I have to say I have met amazing people with incredible antenna systems and stations.

This was an interesting process from idea to setting the criteria, testing the ideas with computer models, finally implementing the prototypes, and test running the antenna setups (for example worked 7O6T, E40VB, 6O0CW with the prototype one cell HRS antenna). For prototyping I took the approach to fail early and fail cheap. Several antennas were shattered in pieces after gusty winds, but that was to be expected. I learned mechanical construction in hard way and the places that are suspectible for cracking.

But I think I am jumping ahead to the conclusions. Lets discuss about the thinking behind the decisions.

Decision Making Criteria for Tower Setup

My primary purpose for this tower is DX, IOTA, and SOTA. The QRP contesting is secondary use, although a very important one.

Must haves:

  • > 20dB front/back on all bands
  • beams for 40-10
  • 1.5kW QRO support on all bands
  • low takeoff angle with beam takeoff control for short skip EU contacts
  • no compromise WARC bands
  • mechanically rugged design (gusty winds, frost, heavy wet snow)
  • wide bandwidth to sustain frost, water, snow
  • beam control via station automation
  • Optional and nice to have for this tower:

  • 80/160 bands
  • ability to construct and maintain antennas without heavy lifting cranes on site
  • instant 180 degree direction reversal
  • multi-directional beam i.e. ring rotors or top rotor for one beam to beam multiple directions with one or two tx
  • power split control between beams on single band
  • simultaneous use of same multiband antenna via triplexer on multible bands
  • SO2R support
  • Other factors to consider

  • Winter, winter, winter
  • Small hill on the North side
  • Space of the lot, other antennas
  • Cost of the setup
  • Reference Point

    Quite a few DX'ers are using very successfully the 4 el SteppIR antennas.

    I decided to put the performance figures of 4 el yagi with same dimensions with SteppIR 4 el with 11m boom as the reference point as well.

    So thats it. Not much to ask... :)

    Tuesday, February 5, 2013

    Tower project


    Quad antennas and winter in Finland is thought to be a bad combination.

    My plan is to prove them wrong. A set of quad antennas, made from first class materials will survive the brutal attacks of snow, ice and wind.

    To support the quad array is assembled a 30m (100ft) tower.

    So here it goes. First picture and an empty tower waiting for aluminum, fiberglass, and wires.