|
Cobra UltraLite Multi-band Antenna?
Opinions please. :-)
http://www.k1jek.com/ It got pretty rave reviews on eHam. http://www.eham.net/reviews/detail/3...9a7885c56da2cc -- Vy 73 de Bert WA2SI FISTS #9384 QRP ARCI #11782 |
On Sun, 21 Nov 2004 15:39:31 GMT, "Bert Craig"
wrote: Opinions please. Hi Bert, When I read a testimonial like: In direct comparisons with a G5RV at the same height, I am constantly receiving 3 S-Units better when using the Cobra. I have to wonder about the source. $90 for zipcord? Well, opinion aside, I modeled this (each arm is a single wire, folded back on itself twice) and when I tried to run a SWR sweep in EZNEC from 2MHz to 30MHz in 0.1MHz steps, EZNEC stalled and gave up. I suppose this may have been due to warnings of wires being too close. As it was, I pulled the zip wires apart 1.2" (0.1'). EZNEC wanted a minimum 6". I tried again with 1MHz steps, and the full sweep drew a line across the top of the chart (max SWR). Hoping for more resolution, I tried with 0.2MHz steps, and EZNEC stalled again. Third time is charm (so the saying goes), and with 0.25MHz steps - it stalled again. -sigh- 73's Richard Clark, KB7QHC |
Hi Bert,
When I read a testimonial like: In direct comparisons with a G5RV at the same height, I am constantly receiving 3 S-Units better when using the Cobra. I have to wonder about the source. $90 for zipcord? Well, opinion aside, I modeled this (each arm is a single wire, folded back on itself twice) and when I tried to run a SWR sweep in EZNEC from 2MHz to 30MHz in 0.1MHz steps, EZNEC stalled and gave up. I suppose this may have been due to warnings of wires being too close. As it was, I pulled the zip wires apart 1.2" (0.1'). EZNEC wanted a minimum 6". I tried again with 1MHz steps, and the full sweep drew a line across the top of the chart (max SWR). Hoping for more resolution, I tried with 0.2MHz steps, and EZNEC stalled again. Third time is charm (so the saying goes), and with 0.25MHz steps - it stalled again. -sigh- 73's Richard Clark, KB7QHC I agree with your comments about the testimonials. A piece of wire is a piece of wire. Concerning NEC modeling; the segment lengths should be equal to the wire spacing. The wire diameter is also a factor to be considered. Anyway, to establish the validity of the segmentation, I model a parallel wire transmission line, and see if it is close to the expected impedance. The antenna appeared to use something similar to 300 Ohm ribbon, so half inch spacing is probably closer to the correct dimensions. 73, Frank |
On Sun, 21 Nov 2004 23:24:45 GMT, "Frank"
wrote: so half inch spacing is probably closer to the correct dimensions. Hi Frank, Well, my wires were separated by nearly three times that, it was suggested by a warning window that it should have been 6" and all this is immaterial to the design being Zip cord. OK, so we proceed with an antenna that is "like" the topic, but is spread out to attend to the modeler. I took all wires and separated them by those 6" and did a SWR sweep from 2 to 30MHz by 0.2MHz and EZNEC plunged on through them all without clutching up. Results are 12 +j550 Ohms at 2MHz -11dBi @ 10° resonates below 80M -6dBi @ 10° 1975 -j160 Ohms at 60M -4dBi @ 10° 570 -j890 Ohms at 40M -0.7dBi @ 10° 104 -j680 Ohms at 30M 1.1dBi @ 10° 440 -j610 Ohms at 20M 5.5dBi @ 10° 210 +j1300 Ohms at 17M 7.7dBi @ 10° 420 -j520 Ohms at 15M 8.2dBi @ 10° 173 +j404 Ohms at 12M 10dBi @ 10° 330 -j360 Ohms at 10M 10dBi @ 10° 73's Richard Clark, KB7QHC |
"Richard Clark" wrote in message ... On Sun, 21 Nov 2004 23:24:45 GMT, "Frank" wrote: so half inch spacing is probably closer to the correct dimensions. Hi Frank, Well, my wires were separated by nearly three times that, it was suggested by a warning window that it should have been 6" and all this is immaterial to the design being Zip cord. OK, so we proceed with an antenna that is "like" the topic, but is spread out to attend to the modeler. I took all wires and separated them by those 6" and did a SWR sweep from 2 to 30MHz by 0.2MHz and EZNEC plunged on through them all without clutching up. Results are 12 +j550 Ohms at 2MHz -11dBi @ 10° resonates below 80M -6dBi @ 10° 1975 -j160 Ohms at 60M -4dBi @ 10° 570 -j890 Ohms at 40M -0.7dBi @ 10° 104 -j680 Ohms at 30M 1.1dBi @ 10° 440 -j610 Ohms at 20M 5.5dBi @ 10° 210 +j1300 Ohms at 17M 7.7dBi @ 10° 420 -j520 Ohms at 15M 8.2dBi @ 10° 173 +j404 Ohms at 12M 10dBi @ 10° 330 -j360 Ohms at 10M 10dBi @ 10° 73's Richard Clark, KB7QHC Hi Richard: I have modeled wires separated by 0.5" as transmission lines, and gotten good agreement with the predicted impedance. Only problem, with an antenna of this type, is you exceed the 2000 maximum number of allowable segments. I think I can go to 2" spacing, with 2" segmentation, and will give it a try. I suspect I will get results very similar to yours. I also suspect -- but have yet to verify -- that these antennas will behave pretty much the same as a conventional dipole of the same overall length. Note that a 140 ft dipole resonates at 3.34 MHz. 73, Frank |
On Mon, 22 Nov 2004 01:52:00 GMT, "Frank"
wrote: Note that a 140 ft dipole resonates at 3.34 MHz. Hi Frank, I also note that the "Cobra" doubles the number of resonances across the 2 to 30 MHz spectrum with the resonances coming in adjacent pairs. Our sometime correspondent, Dr. Slick, thinks he invented this characteristic with his mystery, garden variety dipole; but it is a simple matter of putting a bend in the wire. This one manipulation was also the foundation of the "fractal" antenna, but "scientists" gushed on with baroque explanations and hyperbolic claims to disguise a rather mundane oddity. 73's Richard Clark, KB7QHC |
"Richard Clark" wrote in message ... On Mon, 22 Nov 2004 01:52:00 GMT, "Frank" wrote: Note that a 140 ft dipole resonates at 3.34 MHz. Hi Frank, I also note that the "Cobra" doubles the number of resonances across the 2 to 30 MHz spectrum with the resonances coming in adjacent pairs. Our sometime correspondent, Dr. Slick, thinks he invented this characteristic with his mystery, garden variety dipole; but it is a simple matter of putting a bend in the wire. This one manipulation was also the foundation of the "fractal" antenna, but "scientists" gushed on with baroque explanations and hyperbolic claims to disguise a rather mundane oddity. 73's Richard Clark, KB7QHC Hi Richard: Here, below, are the data from a 140 ft "Cobra" type antenna compared with a 140 ft dipole. I modeled #24 AWG, spaced 1", with 1" segments. My previous comment about 2000 segments maximum was incorrect, and NEC can go to 10,000 segments. As it is, the model contained over 5,000 segments, and took almost 2 hours to run. The NEC output file contained about 250,000 lines of data in a 14.5 MB text file. While the 1" segmentation is well below the minimum lambda/1000 recommended by the NEC manual, a transmission line with similar dimensions gave results very close to the expected value of 550 Ohms. These results could therefore be considered as a reasonable validation of the antenna model. Not sure why EZNEC cannot deal with such a model. From an operational point of view there seems to be very little difference between a 140 ft "Cobra" antenna, and a 140 ft dipole. On some frequencies it is noted that the Cobra exhibits a lower efficiency. The data are noticeably different from your results (note that at 2 MHz both our results show an imaginary part of 0), but probably due to the large difference in dimensions. I was not able to verify the number of resonances due to the very long run times. If I had attempted 0.2 MHz steps the program would have run for 8 to 10 hours and generated over a million lines of output data. I think what I need is 64 bit dual processor system to tackle such a problem. Most of the claims I have read on e-ham.net can be considered as unverifiable nonsense. I certainly agree with your comments on the design. As for "fractal" antennas, I do not really know anything about them, so cannot comment. One thing that does come out of this modeling is that an antenna does not have to be very much longer than a 1/4 wave long to provide reasonable input impedances. Incidentally my model was run in free space, so have no data on take-off angle. I doubt it would be very different. 73, Frank Cobra 140 ft Dipole Frequency Real(Z) Imag(Z) Efficiency Real(Z) Imag(Z) Efficiency (MHz) (Ohms) (Ohms) (%) (Ohms) (Ohms) ( %) 2 71.216 201.878 36 18 -737.554 98 4 140.456 810.61 66 125.32 272.626 99 5 511.886 -23.295 92 359.372 872.604 99.6 7 1889.15 -1965.46 98 2670.88 -2355.48 99.6 10 137.527 -758.652 46 90.313 -165.43 99 14 1465.88 -1284.19 97 2158.99 -1485.13 99.6 18 233.57 679.892 65 180.16 224.359 99.2 21 1272.82 -1009.83 97 1926.04 -1113.17 99.5 24 217.887 -864.421 36 120.373 -159.68 98.9 28 1155.79 -852.805 97 1783.27 -886.08 99.5 |
On Mon, 22 Nov 2004 23:08:32 GMT, "Frank"
wrote: Hi Richard: Here, below, are the data from a 140 ft "Cobra" type antenna compared with a 140 ft dipole. I modeled #24 AWG, spaced 1", with 1" segments. My previous comment about 2000 segments maximum was incorrect, and NEC can go to 10,000 segments. As it is, the model contained over 5,000 segments, and took almost 2 hours to run. The NEC output file contained about 250,000 lines of data in a 14.5 MB text file. While the 1" segmentation is well below the minimum lambda/1000 recommended by the NEC manual, a transmission line with similar dimensions gave results very close to the expected value of 550 Ohms. These results could therefore be considered as a reasonable validation of the antenna model. Not sure why EZNEC cannot deal with such a model. Hi Frank, Well, your segmentation at one inch intervals over 420 feet of wire seems rather over-the-top, but I suppose you arrived at that through your transmission line modeling. Perhaps I should attempt that myself in my own TL models. EZNEC, itself, offers segmentation and geometry tests and where there are issues it will remedy them itself. For this turn in the road I am simply the loose nut behind the steering wheel. The first pass I offered here (that brought EZNEC to its knees) consisted of only 263 segments. I then spread the wires out to accommodate those issues EZNEC reported but would not correct of its own volition. Were your wires as close as conventional 3-wire zip cord? From an operational point of view there seems to be very little difference between a 140 ft "Cobra" antenna, and a 140 ft dipole. On some frequencies it is noted that the Cobra exhibits a lower efficiency. No doubt from one proximal wire bucking the other. The data are noticeably different from your results (note that at 2 MHz both our results show an imaginary part of 0), but probably due to the large difference in dimensions. I would offer more likely due to nearby embrace of earth. I was not able to verify the number of resonances due to the very long run times. If I had attempted 0.2 MHz steps the program would have run for 8 to 10 hours and generated over a million lines of output data. I think what I need is 64 bit dual processor system to tackle such a problem. Most of the claims I have read on e-ham.net can be considered as unverifiable nonsense. I certainly agree with your comments on the design. As for "fractal" antennas, I do not really know anything about them, so cannot comment. Consult: http://www.qsl.net/kb7qhc/antenna/fr...och/review.htm for one small sample from many hundred pages of results. One thing that does come out of this modeling is that an antenna does not have to be very much longer than a 1/4 wave long to provide reasonable input impedances. Incidentally my model was run in free space, so have no data on take-off angle. I doubt it would be very different. 73, Frank Cobra 140 ft Dipole Frequency Real(Z) Imag(Z) Efficiency Real(Z) Imag(Z) Efficiency (MHz) (Ohms) (Ohms) (%) (Ohms) (Ohms) ( %) 2 71.216 201.878 36 18 -737.554 98 mine 12 +j550 Ohms 14 1465.88 -1284.19 97 2158.99 -1485.13 99.6 mine 440 -j610 21 1272.82 -1009.83 97 1926.04 -1113.17 99.5 mine 420 -j520 It appears that your reader, Outlook Express (a piece of software crap by the way) is line length limited. Comparisons were difficult at best so I confined them to those above. I was going to do a side-by-side comparison with the same 140' dipole, but it was unremarkable when earth dominates the drive point Z. Another explanation for variations may be simply found in the high degree of Z variation across frequency, we do agree on the imaginary sign and the values do track in the higher bands. 73's Richard Clark, KB7QHC |
Hi Frank,
Well, your segmentation at one inch intervals over 420 feet of wire seems rather over-the-top, but I suppose you arrived at that through your transmission line modeling. Perhaps I should attempt that myself in my own TL models. You are correct Richard -- It appears my segmentation is seriously "over the top". The segmentation was driven by not wanting a segment to be longer than the spacing. Some texts on NEC do recomend such constraints. So far I have gone to 16" segments, with 1" spacing. For the transmission line model it shows even better convergence. #24 AWG spaced by 1" has a characterestic impedance of 550 Ohm. My transmission line model is 100 ft long with 75 segments in each side. It would be interesting to know if EZNEC would produce the same result. EZNEC, itself, offers segmentation and geometry tests and where there are issues it will remedy them itself. For this turn in the road I am simply the loose nut behind the steering wheel. The first pass I offered here (that brought EZNEC to its knees) consisted of only 263 segments. I then spread the wires out to accommodate those issues EZNEC reported but would not correct of its own volition. Were your wires as close as conventional 3-wire zip cord? No, as explained above they are #24 spaced by 1". Possibly I missed something there. Is the antenna made from simple 3-wire zip cord? Unfortunately NEC cannot model insulated wire, so not sure how that would effect the model. Is the VF very significant? From an operational point of view there seems to be very little difference between a 140 ft "Cobra" antenna, and a 140 ft dipole. On some frequencies it is noted that the Cobra exhibits a lower efficiency. No doubt from one proximal wire bucking the other. That's what I figured. The data are noticeably different from your results (note that at 2 MHz both our results show an imaginary part of 0), but probably due to the large difference in dimensions. I would offer more likely due to nearby embrace of earth. Very likely. How high did you place your model? Did you use the Sommerfeld/Norton grouind model? Probably not really a factor at 30 ft, or more, above ground. Consult: http://www.qsl.net/kb7qhc/antenna/fr...och/review.htm for one small sample from many hundred pages of results. One thing that does come out of this modeling is that an antenna does not have to be very much longer than a 1/4 wave long to provide reasonable input impedances. Incidentally my model was run in free space, so have no data on take-off angle. I doubt it would be very different. 73, Frank Cobra 140 ft Dipole Frequency Real(Z) Imag(Z) Efficiency Real(Z) Imag(Z) Efficiency (MHz) (Ohms) (Ohms) (%) (Ohms) (Ohms) ( %) 2 71.216 201.878 36 18 -737.554 98 mine 12 +j550 Ohms 14 1465.88 -1284.19 97 2158.99 -1485.13 99.6 mine 440 -j610 21 1272.82 -1009.83 97 1926.04 -1113.17 99.5 mine 420 -j520 It appears that your reader, Outlook Express (a piece of software crap by the way) is line length limited. It sure is! Comparisons were difficult at best so I confined them to those above. I was going to do a side-by-side comparison with the same 140' dipole, but it was unremarkable when earth dominates the drive point Z. Another explanation for variations may be simply found in the high degree of Z variation across frequency, we do agree on the imaginary sign and the values do track in the higher bands. 73's Richard Clark, KB7QHC Now I have gotten out of my segmentation fixation I may be able to run some more realistic models. 73, Frank |
On Tue, 23 Nov 2004 05:14:14 GMT, "Frank"
wrote: My transmission line model is 100 ft long with 75 segments in each side. It would be interesting to know if EZNEC would produce the same result. Hi Frank, I will try to fill that request. If not soon, eventually. Were your wires as close as conventional 3-wire zip cord? No, as explained above they are #24 spaced by 1". Possibly I missed something there. Is the antenna made from simple 3-wire zip cord? Yes, by inference from close-up photography at the web site. Unfortunately NEC cannot model insulated wire, so not sure how that would effect the model. Is the VF very significant? EZNEC can accommodate insulation, but I rarely fine tune to that degree when I am looking at the panorama of SWR sweeps. Really, this will only nudge things in the single digit percentage range and such things are swamped from one Ham's situation to another's. I would offer more likely due to nearby embrace of earth. Very likely. How high did you place your model? Did you use the Sommerfeld/Norton grouind model? Probably not really a factor at 30 ft, or more, above ground. I placed it at the reported height, 35 feet, that raised the signal levels 2 S-Units over a similarly mounted G5RV. ;-) I choose one of two ground models when I work on these things: Perfect, or average, high accuracy lossy. That height seems to be a commonly reported value with those who put up wire antennas and have no tower to work from, so it works for me as an "Everyman's" best. 73's Richard Clark, KB7QHC |
On Tue, 23 Nov 2004 06:47:32 GMT, Richard Clark
wrote: On Tue, 23 Nov 2004 05:14:14 GMT, "Frank" wrote: My transmission line model is 100 ft long with 75 segments in each side. It would be interesting to know if EZNEC would produce the same result. Hi Frank, I will try to fill that request. If not soon, eventually. Hi Frank, I gave this several passes with various segmentations. The best, flattest response appeared at 400 (total) segments over a 10M length of line of #24 separated by 1". This resulted in a characteristic 575 Ohm line with a 1.045:1 SWR ripple over the 2 MHz to 30 MHz band. When I doubled and then tripled the segments, the low end got a little whacked out (getting worse as the segments went up). I was working from many segments down, so I did not lower the count to your longer segments - maybe tomorrow, I was paying attention to your (and EZNEC's) advice about segment length equaling line separation. 73's Richard Clark, KB7QHC |
Hi Frank,
I gave this several passes with various segmentations. The best, flattest response appeared at 400 (total) segments over a 10M length of line of #24 separated by 1". This resulted in a characteristic 575 Ohm line with a 1.045:1 SWR ripple over the 2 MHz to 30 MHz band. When I doubled and then tripled the segments, the low end got a little whacked out (getting worse as the segments went up). I was working from many segments down, so I did not lower the count to your longer segments - maybe tomorrow, I was paying attention to your (and EZNEC's) advice about segment length equaling line separation. 73's Richard Clark, KB7QHC Hi Richard: Found some three conductor zipcord and measured its dimensions: wire, stranded, #14 AWG, center to center spacing 3.5mm (0.138"). Experimented with a 100ft NEC transmission line model, with various segmentations from 4" to 16". Characteristic impedance agreed with :- 276*log(2*S/d) at 171 Ohms, where S is the center to center wire spacing, and d is the wire diameter. From your comments it seems EZNEC is in close agreement with my NEC program. Applied the above dimension to a 140 ft "Cobra" type antenna at 35 ft above an average ground of: sigma 5 mS/m, and Er = 13. The antenna was segmented at 6". Hope the formatting of the following results is not totally destroyed by Microsoft's Outlook Express. Cobra Antenna: Freq Re Im S Eff. (MHz) (450) (%) 2 29.8 93.8 15.7 66.3 3.8 109 637 12.6 83.4 5 251 58 1.8 95.4 7.2 999 -1381 6.8 99.0 10 516 1850 16.7 68 14.2 1287 -989 4.1 98.9 18 315 861 7.2 81.3 21.2 1096 -915 3.3 98.6 25 301 672 5.3 82.7 28.6 669 -712 3.6 98.2 140 ft Dipole compared Freq Re Im S Eff. (MHz) (450) (%) 2 8.8 -856 235 66.3 3.8 79 238 7.3 83.3 5 361 1127 9.8 95.4 7.2 1959 -2379 13 99 10 111 -195 4.9 68 14.2 2581 -2020 9.3 98.9 18 181 251 3.4 81.3 21.2 2225 -1871 8.5 98.6 25 188 233 3.1 82.7 28.6 1252 -1577 7.4 98.2 If the NEC models are correct there does not seem to be a lot of difference between the "Cobra" and a 140 ft dipole. Interesting to note that the Imaginary part of Zin, at 2 MHz is still 0. With 100 ft of 450 Ohm line on 2 MHz the loss is only 0.12 dB compared with 1.5 dB on the regular dipole. |
On Wed, 24 Nov 2004 03:49:20 GMT, "Frank"
wrote: If the NEC models are correct there does not seem to be a lot of difference between the "Cobra" and a 140 ft dipole. Hi Frank, So it would seem. Hi Bert, You asked for comments, but you've been quite silent yourself. Would you think that the extra wire was worth the $90? Would you think you could trust testimonials, especially those claiming 2 S-Units gain over a G5RV? 73's Richard Clark, KB7QHC |
I guess we scared him away Richard.
73, Frank "Richard Clark" wrote in message ... On Wed, 24 Nov 2004 03:49:20 GMT, "Frank" wrote: If the NEC models are correct there does not seem to be a lot of difference between the "Cobra" and a 140 ft dipole. Hi Frank, So it would seem. Hi Bert, You asked for comments, but you've been quite silent yourself. Would you think that the extra wire was worth the $90? Would you think you could trust testimonials, especially those claiming 2 S-Units gain over a G5RV? 73's Richard Clark, KB7QHC |
Look at it closely. It is three wires (Radio Shack rotor wire) joined at
opposite ends to make one long wire. The 70 foot model actually has a "run" of 210 feet - 105 feet on each side. Funny, Radio Shack rotor wire comes in 70 foot rolls. Attach ladder line to a middle insulator. Add up the cost and you'll save a lot by doing it yourself. I have seen a Cobra and it looks very well made but I have never used one. The principal is called linear loading, and it is a very effective way of getting maximum benefit out of short antennas. It is far more efficient than coils. ON4UN's book talks a lot about it but I couldn't find much discussion on the effect of having the wires so close together. He does advise, on page 8-13 of the Third Edition, "make sure the separation between the element and the folded linear-loading device is large enough, and that you use high-quality insulators to prevent arc-over and insulator damage." He doesn't define "enough" and the phrasing of the sentence makes it sound like the only concern with having them too close is arc-over. On the same page, he warns modeling is "very tricky." From the discussion that follows on this thread, I would say that is an understatement. -- Radio K4ia Craig "Buck" Fredericksburg, VA USA FISTS 6702 cc 788 Diamond 64 "Bert Craig" wrote in message . net... Opinions please. :-) http://www.k1jek.com/ It got pretty rave reviews on eHam. http://www.eham.net/reviews/detail/3...9a7885c56da2cc -- Vy 73 de Bert WA2SI FISTS #9384 QRP ARCI #11782 |
On Thu, 25 Nov 2004 15:46:40 GMT, "Frank"
wrote: I guess we scared him away Richard. Hi Frank, Well on to other speculations.... Sometime later this morning, I will release a study in transmission lines that your discussion aided me with. It won't be so much about the lines themselves, but about a compendium of topics revolving around the Zc of the source, and what is called Mismatch Uncertainty. I've done the grunt work, it just needs some introduction. However, as is typical with this topic, the thread will no doubt run to half a dozen comments, if that much. :-) 73's Richard Clark, KB7QHC |
Glad something I did helped Richard. Will look forward to your posting. I
enjoy reading these huge threads, though don't always feel I can contribute much. 73, Frank "Richard Clark" wrote in message ... On Thu, 25 Nov 2004 15:46:40 GMT, "Frank" wrote: I guess we scared him away Richard. Hi Frank, Well on to other speculations.... Sometime later this morning, I will release a study in transmission lines that your discussion aided me with. It won't be so much about the lines themselves, but about a compendium of topics revolving around the Zc of the source, and what is called Mismatch Uncertainty. I've done the grunt work, it just needs some introduction. However, as is typical with this topic, the thread will no doubt run to half a dozen comments, if that much. :-) 73's Richard Clark, KB7QHC |
| All times are GMT +1. The time now is 03:29 PM. |
Powered by vBulletin® Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
RadioBanter.com