Posted by
donovanf on
Nov 22, 2017; 5:41pm
URL: http://elecraft.85.s1.nabble.com/Re-KX3-Field-Ant-for-80-40-30-tp7636045p7636173.html
Hi Eric,
Don't give up on WSPRlite, with experience you'll learn that its an
exceptionally powerful antenna evaluation tool.
The "simultaneous spots" comparison tool on DXplorer.net allows you
to select the maximum distance for the WSPR reports being evaluated.
It also computes the mean and standard deviations in dB between the
two WSPRlite transmitters (and antennas) being compared. There's
no magic here.
Because the ionosphere and your antennas are the media connecting
your WSPRlite transmitters to WSPR receivers all over the world,
its very important that you take steps to reduce interaction
among the antennas being compared and to reduce the ionospheric
variability affecting usefulness the reported data. The important
necessary steps are:
1. The antennas should transmit the same polarization.
2. The antennas should be as close as possible but not so close that
they interact with each other (EZNEC helps you reduce interactions)
This discussion needs to move off of the Elecraft reflector, it has little
to do with the purpose of this reflector or Elecraft products The
WSPRlite page on Facebook is a better choice.
73
Frank
W3LPL
----- Original Message -----
From: "Erik Basilier" <
[hidden email]>
To:
[hidden email],
[hidden email],
[hidden email]
Sent: Wednesday, November 22, 2017 8:43:57 AM
Subject: RE: [Elecraft] KX3 Field Ant. for 80/40/30
Don, please see my comments below:
>About all you can conclude from the WSPR readings is the relative difference between antennas. If you want to compare two horizontal antennas, place them end to end. >The radiation at the end is at a minimum and the antenna will not interact.
Years ago I was under the naïve impression that if you buy a perfectly symmetrical GP, put it in a vertical position high up in the air, with no nearby metal objects, and run the coax perfectly vertically to a transmitter where the feeder hits the ground, you will get a nice pattern that is well predicted by a model. After all, the antenna manufacturer probably showed us that pattern in their catalog, and we have seen that pattern before in our antenna book. Well, not too long ago QST published an article that said (paraphrasing): "Not so fast. Sometimes you will get a very different pattern, and it may be terrible at the low elevation angles that you want." The reason: the feedline has current on the outside. How much will depend on its length. Putting a common mode choke at the feedpoint won't save you if the feedline is long; you may have to break up the current by several chokes along the line." The author supported this by theory as well as measurement. And why not? The feedline is a nearby metallic object, so why should it not affect the performance of the antenna? With that in mind, I have been expecting interaction also between two horizontal dipoles installed in line. I just went to EZNEC to check it. For the GP with a vertical, non-connected halfwave wire hanging under it, the center current in that wire came to about 1/4 of the GP drive current. The pattern was changed but not too badly. Then I modelled a horizontal dipole with another one mounted close off the end, in line. The center current in the non-driven dipole came to about 38% of the drive current in the driven dipole. The azimuth pattern was now a 4-clover similar to that of a full-wave dipole. In comparison tests where patterns are important, that is a lot of interaction between the two dipoles arranged end-to-end. I also modelled two horizontal dipoles mounted at right angles (one driven dipole end close to one end of non-driven dipole). Now the current in the undriven dipole got as high as about 58%. The azimuth pattern was now bidirectional with no difference front to back, but with a wider lobe on one of the sides. It seems that interaction is very significant regardless of orientation as long as the wires are close.
>As for comparing a vertical with a horizontal antenna, or two verticals, the only thing one can say is at that particular time and distance for propagation, one antenna is better >than the other. That may not be true for other propagation conditions, so be careful when generalizing.
>As far as two horizonal antennas oriented in different directions, you would expect greater signal strength in directions broadside to the antenna. That directivity may be >useful in actual use, but is not a valid comparison between the two antennas.
Granted, but since I drank the WSPRLite antenna comparison coolaid, I view the ralated capability through a certain mental filter, which says, on the one hand, that it is a significant advancement in the art to using WSPR for antenna testing (the two WSPR transmitters combined with some special software proprietary to SOTABeams is more than just a combination of the transmitters), and on the other hand that the technique is based on certain assumptions that need to be examined if one tries to determine the limits of what is doable and what is not.
For those readers that have not looked closely at what this new approach involves, even if they are familiar with WSPR by iteslf, I will summarize my understanding here. If you run a WSPR transmitter (or two) you can go to the free website and get a list of stations that heard you and see the s/n ratio for each. You quickly get overwhelmed by all the data, The distance to the receiving stations that heard you is probably a positive to you, and so is the number of stations that heard you. Maybe you like a contest that gives you more points for greater distane, so you like to see dx stations in the list of stations that heard you. Maybe you do Field Day, and you get the same points regardless of distance. In either case, trying to evaluate the data presented to you as an antenna comparison is time-consuming and confusing. Enter SOTABeams. They set out to create software that takes data off the WSPR website, for each of the repeating WSPR transmissions, and combines it into one performance number for each transmission, for each of the two transmitters. The algorithm is proprietaty, so you don't know if a high number is high mostly because of good distances or because of many stations hearing you. I think we can assume that both distance and number of receiving stations play a role. The numbers we get from this software certainly doesn't correspond to something that we would really like to see, but it does reduce the confusing messes of data to single numbers that we can use to compare the two antennas. We log into a website (not the WSPR website that has all the s/n numbers) but one that shows two graphs overlaid, one for each transmitter/antenna. While we don't know what iach curve really represents, the people who created the algorithm seem to have struck on some reasonable choices. With only a few minutes between transmissions, we see the graphs suggest how one or the other antenna is doing better as a function of time, for a given day. We may see how one antenna does better than the other in late afternoon, consistent with the band "going long", and we may see the same thing repeat on multiple days. Of course we cannot translate the graphed numbers into anything like signal strength differences in db or relative gain. So, if we don't really know what the results represent, is this just a toy with no real usefulness? First of all, you could use two of your existing radios with digital mode setups, and just pay a modest amount for the software subscription (price and availability unknown), or you could buy the two WSPRLite transmitters at toy prices (about $75 each) and get a year of the software for free. To me the latter approach was a no-brainer. Once you program each transmitter (requires a computer) the WSPRLite transmitter (size is like a fat matchbox) doesn't need a separate computer to do its job (if you change call sign or grid square you do need to re-program.) So, in my view it doesn't cost a lot for an entry ticket to this capability to see how two antennas do better or worse relative to each other in some sense that is undefined, yet designed to be meaningful to the pursuits hams are typically involved in. While new data points appear on the two graphs every few minutes, selective fading means that we cannot even trust that Antenna 1 is doing better than Antenna 2 at this moment just because Graph 1 is on top of Graph 2 for the most recent transmission cycle. Over time we should get a clearer picture though. The WSPRLite transmitters have a start button that allows us to synchronize the transmission cycles; this helps make the results comparable, but other factors can hurt comparability: Distance between the two antennas, and difference in frequencies used (very very small). Frank suggested that antennas should be located no more than one wavelength from each other. I might want to model to see if interactions are necligible at that that distance. Obviously much depends on how long we are willing to test before we draw conclusions.
Now that I have explained what the comparison system does, how can it exactly be useful to me to justify my investment ot $150? How "good" an antenna is cannot be stated without first defining what you want to use it for.I like to experiment with alternative antennas for Field Day. This means all contacts give me the same points, regardless of location. Located in AZ I expect to work the West Coast easily, while the East coast is more of a challenge. The first tool I use is EZNEC. I don't look for the highest maximum gain, but the gain at the elevation angle that is expected for the East Coast. I compare to a standard dipole, which has considerable gain over an isotropic, due to both its azimuth directivity and to ground reflection gain. The proposed new antenna may have a gain of, say 11dBi, but the dipole may be at say, 7dbi, which means the new antenna is expected to provide a 4 db improvement. I don't want too much gain/directivity anyway, since I still want to work the West Coast. But, does the actual antenna work as expected from the model? I haven't yet used the WSPRLite method for this actual comparison, but it seems to me that if I erect the new, horizontally polarized antenna close to, but not too close to, the standard dipole, both oriented toward the East Coast, and if I allow sufficient time for the comparison run, I have a good chance to see the new antenna dominate in the graphs. This isn't proof that the new antenna is better, but seeing it would give me increased confidence that the new antenna works as designed. My other use case is optimizing a long wire for quick setup portable operation using a single 24ft support, as previously mentioned on this thread. Here the two antennas would be very similar, which should improve the validity of comparison, but, I would again need to find a compromise distance between antennas that minimizes interaction while minimizes selective fading differences. With a horizontal wire mounted so low I expect that most qso's will be within the Southwest, and within that area the population of WSPR receiving stations may be homogeneous enough that antenna orientation doesn't matter much, but I don't see any reason not to to point both antennas in the same direction. It seems like I will need to use considerable feedline lengths in order to get the separation while keeping the transmitters close toghether so I can use both hands to start both at the same time. That probably means that I need more space than my back yard provides.
I can't help thinking about how the comparison software might take into account the distance to receiving stations versus number of stations. I speculate that different formulas might be optimal for use cases involving differet types of contests. Maybe SOTABeams might be persuaded in the future to let users configuse the software for such different needs. Maybe the next step after that would be to let the formulas be tailored further to fit the beamwidth of a beam, and to compensate for different parts of the worlkd having different densities of WSPR receiving stations.
73,
Erik K7TV
>73,
>Don W3FPR
On 11/21/2017 8:40 PM, Erik Basilier wrote:
> I hope my interest in WSPRLite antenna comparisons doesn't lead this thread too far off topic, but I have further thoughts on how to orient the two antennas being compared.
>
> Frank, who is much more experienced with this comparison system, suggested that two horizontally polarized antennas should be oriented end-to end, due to parasitic interacton between the antennas. I questioned whether the end-to-end configuration would be free from interactions. Be that as it may, but while thinking about configuration choices, I came up with another reason why end-to-end would be the right thing to do.
>
> Suppose we are comparing two omnidirectional antennas such as verticals. Even if the receiving stations are unevenly distributed in different directions, the comparison based on received reports should be fair. If instead we are comparing two horizontal dipoles, that are not pointing in the same direction, and receiving stations are not distributed evenly in all directions, the antenna with fewer receivers in the main lobes would likely be at a disadvantage. If the feed system is the part that is different between the two ontennas, one could compensate, as I suggested, by swapping antennas for each feed system, but the time taken allows the conditions to change, so one would probably have to go back and forth a number of times to gain confidence in any observed difference in performance. Close to the coast receiving stations would be largely missing in roughtly half of possible compass directions, and unidirectional antennas would be affected more than a dipole with its bidirectional pattern. Much seems to depend on the proprietary algorithm used to composite a single performance number for from the WSPR received s/n rations at multiple receiving stations. What is the balance between the number of good reception reports vs. the distance for each one? When we talk about difficulty in comparing one vertical and one horizontal antenna, I suspect that similar considerations may account for result being inconsistent or difficult to interpret.
>
______________________________________________________________
Elecraft mailing list
Home:
http://mailman.qth.net/mailman/listinfo/elecraftHelp:
http://mailman.qth.net/mmfaq.htmPost: mailto:
[hidden email]
This list hosted by:
http://www.qsl.netPlease help support this email list:
http://www.qsl.net/donate.htmlMessage delivered to
[hidden email]