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

 

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

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

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

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.

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.

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... :)

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.

RTL-SDR R820T Tuner 10m band

 

I wrote that I have been tinkering with RTL-SDR to enable wide band receiving capability. I have a couple of different RTL-SDR variants:

One is equipped with Elonics EL4000 tuner, and the other one is based on R820T chip. Both work very nicely in Win7 with sdrsharp software. The difference between the two tuners is that the R820T should go down to 24MHz enabling receive on 12 and 10 meters without the need of upconverter.

The following is the screenshot from 10m band during the ARRL 10m contest last weekend.

Not much happening, but band conditions were terrible. Some signals coming through. I compared the receive capability with my Flex 1500. Same signals came through on both receivers. The benefit of using TV receiver USB stick is really the 2MHz pan-adapter. Based in initial experiments a band filter before the tuner is a must. Too much interference coming through from local FM stations.

I user my ole TS-2000 (modified) by transmitting 5W CW to a dummy load below 24MHz (low spec limit of R820T) and checked how low the receiver can actually go. And specs hold true. The R820T tuned at 24MHz with the low end of receive frequency at 23MHz. The tests confirm that R820T based TV tuner can receive 12 and 28m HAM bands without the need for upconverter.

Upconverter Tests to Follow

I do have two different upconverters. I did some tests with E4000 stick and the DxPatrol.com v4.0 upconverter earlier. I will report when I have time to compare DxPatrol kit to one that nooelec.com has in sale. I got my Nooelec.com upconverter just two days ago.

First Snow of The Season

Preparing for Winter

It is that time of the year again. All the old antenna projects that are laying in the garden needs to be dismantled. Garage needs cleaning from the aluminium and copper pieces, not to mention a random pile of several hundred meters of "official Finnish antenna wire" i.e "Killu".

So basically the time between this post and the previous has been cleaning. Very little radio time.

Lots of time in the shack though.. preparing for the winter. Got my KX3 delivered by QRP Project guys. I definitely can recommend their great service. I also got the milliwatt WSPR beacon for 30m band from QRSS specialist Hans Summers http://www.hanssummers.com/ I already have constructed their QRSS CW kit, but wanted the "ultimate" version for 30m WSPR work. For the RX side I constructed and tested the SoftRock Ensemble rx ii kit. I have a couple more of the SoftRocks in pieces and during the winter the plan is to put these online to serve the ham community.

The low band antenna upgrade took a leap forward. Last weekend I were on the road and transferred the tower #2 pieces from eastern Finland to my QTH. The plan is to construct 160m 1/4 lambda vertical and 80m four square to complement the antenna setup. The 160m plan is still in flux.

Radio Time

As said earlier very little time with the radios in QSO mode, but of course some qsos in the log.

Three new IOTAs from P29 IOTA island hopping pedition P29NI. The band conditions have been reasonable and thaks to SM6CVX in the team they have been on the right bands at the right time to provide the new IOTAs to Scandinavia. Big Thank You for remembering us up here! Unfortunately, due to travelling I missed the Lihir OC-069.

Vatican

On Sunday when I got back home with the tower #2 I turned on the radio and noticed a big pileup on 15m band. That was HV0A, Francesco from Vatican. I do have Vatican in the log but not for this year so this was definitely must to work. Since it was SSB I decided to go QRO and heat up my amp. The split was running 10k wide. While the amp was heating I tried to listen the pileup. His antenna was beaming stateside and he was taking lots of Gs, DLs, and of course Ws and VEs. The Ws were coming from the back of my beam so when the amp was ready I was still with no clue where he was actually listening. He was +20 on my S-meter so my guess was that if I hit to his passband I will be at least +20 and getting through. I set the XIT to about 4 up, gave my call once and he came back "OH3 59". I repeated my call and we exchanged the reports. That was a lucky catch.

T30PY Aftermath

T30PY pedition is over. Total number of QSOs exceeds 40k according to their website..

The band conditions have been reasonably good, but the produced signal to OH has been quite disappointing. The unusually thin OH participation in the Clublog charts tells the sad story.

So I think some aftermath from the "customer" point of view is in place.

First I want to stress this is not a complaint. I had my QSO with the pedition and got T30PY to my log and my call is listed in the clublog. I was looking forward the QRP qso, but there was no opportunity to even try.

1. Band conditions

As I already wrote the band conditions were reasonable. SFI was in steady raise (120+) and K stayed low during the pedition. So the conditions were reasonable.

K indexes for the period of the pedition (source NOAA):

day16  17  17  18  20  21  22  23  24
K*  1    2    1    1    0    0    0    2    1 

*)Daily K average calculated from 3 hour average by me

The VOACAP calculation that I did earlier was for QRP assuming the optimal placement of their VDAs. I redid the calculation with the more recent knowledge about the antenna placement and the picture did not get any better. The best opportunity was on 15m between 8z-11z and propability 50%. In a pileup situation this is pretty much hopeless.

The best propagation between OH and T30 is during the normal office hours, when people are at work. Looking back the calendar there was only one weekend during the pedition (Sat 20, Sun 21). I think this had a significant impact to OH land participation.

    2. Antenna placement

The common agreement is that on island DX the VDAs are the way to go. 9M4SLL (Spratly) did outstanding job with simple multiband verticals placed very close to the ocean so expectation for T30 was al
The T30PY main antenna according to their website was:

• Big IR SteppIR 6-80m mounted near the water

There is good deal of wisdom available about what that "near water" must be for saltwater vertical to really come alive. Team Vertical has done numerous studies what is the right placement for verticals near salt water. The empirical result is 1/4 wave lenght or closer to the ocean! Very interesting reading is the study they conducted in 1998(!!) Quote from their study (appeared in CQ): "As the vertical was moved back from the water, there was little change until we came close to 1/4 wavelength from the water. At that point there was a 3 dB increase in signal level!"

So I take that the antenna selection was the right one.

The placement what I see from the pictures from their website was not optimal shoreline 1/4 lambda, but further inland, so this may have impacted the signal strength both ways.

3. Pedition focus

Cross review of band activity and propagation to OH based on cluster spots found from DX Summit archives.

As can be found from Voacap forecast the best propagation is between 6am-12 on the high bands, so lets see where they were and for QRP it had to be CW.

T30YP: 6am - 12am band activity

Oct 16 Tue
30m
17m SSB

Oct 17 Wed
40
30
20 SSB+CW
17 CW

Oct 18 Thu
80
40
30
20 SSB

Oct 19 Fri
160
80
40
20 SSB (lots of EU spots on cluster)

Oct 20 Sat
160
80
40
20 SSB

Oct 21 Sun
160
80
40
30
20 CW

Oct 22 Mon
80
40
30
20 CW + SSB
17 CW
12

Oct 23 Tue
80
40
30
20
17
12
20 CW

Oct 24 Wed
40
20 RTTY :), CW

Summary

Great job on low bands. Lots of activity on 30m, which is very good.

OH-wise the most optimal frequencies and times were not utilized by the T30PY team, which indicates that the Northern Europe was not high in their priority list.

Basically for QRP we can forget all SSB openings. The only viable option would have been Monday to try the QRP qso, but as said that was a work day and I was traveling. So no QRP this time.

I think the ops made good effort asking JA/SA/NA to QRX when working EU.