home brewing a gain master ???

[Marconi]

If we are discussing homebrewing a GM, and the complexities of the choke and what it may or may not be doing, then adding a torroidal choke would not only simplify construction, but be an effective choke to boot, I'm not questioning the Sirio version
As for modelling, I have every faith in a well designed model, anything you've got would be of interest.

35, what experience do you have in modeling, and what program do you use, if any?

 

[Needle Bender]

I plan to do a video of the Vector, showing radiating currents, and upload it on youtube for all to see. I hope to get this set-up and done soon, and it should put an end to this controversy for good.

One thing that troubles me about Shockwaves voltage projections at the Gainmaster center feed point; that part of the antenna is a high current node, not the highest current node but very high and therefore it should be a very low voltage node so I don't see how you can have voltage numbers much above a hundred volts, let alone thousands of volts?
Where I expect the Gainmaster to fail is at the high voltage point at the very bottom of the radiator where the coax leaves the top of the coil with a bend. At that point the RF voltage is highest and RF current is lowest. I would expect to see the shield arc through the dielectric to the center lead and destroy it at that point.

I want to try building a home made Gainmaster from RG213 since it has a 5kv dielectric rating instead of the much lower 1900v rating of the Red coax.


_______________________

 

[Shockwave]

Marconi, designing collinear models of the Vector just like you did with the 5/8 wave is exactly how I've come to understand how the radiation currents are distributed on the Vector and also why I say EZNEC is not the program to model this design in. I know nothing about the inner workings of these software modeling programs but I sure can spot when the model does not match the results on the test range and is why I prefer CST. Case in point, when you model a two section Vector in EZNEC it wants to see a 180 degree phase delay to drive the top section.

That sure supports the "half wave Vector thinkers" and may be where they get some ideas but EZNEC misses the boat 100% on the cone of the Vector. If we really need a 180 degree phase delay to drive another section, that means we only have 180 degrees of radiation taking place in the original radiator. In the field using a 180 degree phase delay causes a drop in gain but using a 90 degree phase delay causes an increase in gain. I know this because I currently hold patent pending status on two phasing networks to be used in this design and probably have 12 or 13 different failed attempts at this because I listened to EZNEC+.

I do my best to report what I find in testing and to avoid opinion. The only way a 90 degree phase delay line can be optimal is if we already have 270 degrees of constructive radiation taking place below the phase line. That's 3/4 wave and as was advertised decades ago "The entire antenna radiates". If the cone was simply carrying transmission mode currents, it would appear more like coax and be a half wave requiring a 180 degree phase delay to stack. For those who might want to see a video of the Vector radiation currents, it's already been done by top professionals in the field with an extensive understanding of CST and the Vector design. The model can be viewed at this link http://fmbroadcastantenna.com/images/Dominator NWE-34 in CST.gif

With respect to the breakdown voltage of the cap used in the GM, I said it's not what you would expect at a typical feedpoint. You already know the coax used in the cap is of a diameter that produces a rated breakdown voltage of 1900 and that this cap failed on original production runs at just over 600 watts. Lot more voltage for the wattage then you would see on a low impedance 50 ohm line. I think 5000 volts on a 50 ohm line is 500,000 watts so you can see this antenna is unusual to be developing these high RF voltages.

 

[northern35s]

35, what experience do you have in modeling, and what program do you use, if any?

I'm by no means a skilled modeler, I've used both EZNEC and MMANA-GAL, the freebies of course

If you have a jpeg of your Vector 4000 model I wouldn't mind having a look, I'd venture there would be to many segments for the free version of EZNEC.

 

[bob85]

nothing about the vector/sigma is in argument with accepted theory 35s, so long as you use the correct theory,

you may not get much radiation from window line or close spaced wire j-poles but its known that wider spaced j-poles have some radiation from the short leg (CEBIK)

its known that the 1/4wave sleeve of an open sleeve antenna radiates in phase with the upper 1/2wave (arrl)

its known that big sticks and gain masters radiate from the lower half in phase with the upper half,

figure out what causes the radiation from the above antennas:sneaky2:

 

[northern35s]

Hi Bob, I know you're a proponent of the Vector 4000 and have read your comments about the same before now, I'd like to get my head around why it works as you, and Marconi, say it does

 

[Marconi]

Marconi, designing collinear models of the Vector just like you did with the 5/8 wave is exactly how I've come to understand how the radiation currents are distributed on the Vector and also why I say EZNEC is not the program to model this design in. I know nothing about the inner workings of these software modeling programs but I sure can spot when the model does not match the results on the test range and is why I prefer CST. Case in point, when you model a two section Vector in EZNEC it wants to see a 180 degree phase delay to drive the top section.

That sure supports the "half wave Vector thinkers" and may be where they get some ideas but EZNEC misses the boat 100% on the cone of the Vector. If we really need a 180 degree phase delay to drive another section, that means we only have 180 degrees of radiation taking place in the original radiator. In the field using a 180 degree phase delay causes a drop in gain but using a 90 degree phase delay causes an increase in gain. I know this because I currently hold patent pending status on two phasing networks to be used in this design and probably have 12 or 13 different failed attempts at this because I listened to EZNEC+.

I do my best to report what I find in testing and to avoid opinion. The only way a 90 degree phase delay line can be optimal is if we already have 270 degrees of constructive radiation taking place below the phase line. That's 3/4 wave and as was advertised decades ago "The entire antenna radiates". If the cone was simply carrying transmission mode currents, it would appear more like coax and be a half wave requiring a 180 degree phase delay to stack. For those who might want to see a video of the Vector radiation currents, it's already been done by top professionals in the field with an extensive understanding of CST and the Vector design. The model can be viewed at this link http://fmbroadcastantenna.com/images/Dominator NWE-34 in CST.gif

With respect to the breakdown voltage of the cap used in the GM, I said it's not what you would expect at a typical feedpoint. You already know the coax used in the cap is of a diameter that produces a rated breakdown voltage of 1900 and that this cap failed on original production runs at just over 600 watts. Lot more voltage for the wattage then you would see on a low impedance 50 ohm line. I think 5000 volts on a 50 ohm line is 500,000 watts so you can see this antenna is unusual to be developing these high RF voltages.

Shockwave, your link is just a neat feature of the model that is meant to simulate the current flow. I think it does suggest some collinear type activity that may be non apparent as Cebik told Bob. I also think I see something similar with my model, as all the antenna looks to me to be in the same current phase, from top to bottom except for a small area around where the gamma match sits. My model of a similar 3/4 wave without the hoop, tends to reject the idea that it too will act like a non apparent collinear. Reason being is the radials and the radiator are not in phase. This is what I see, right or wrong. Can you show me what your antenna results produce?

Did you find some other data in my model that was way too low or way too high or the pattern skewed somehow? How does my gain, angle, and SWR or source impedance compare? Does your CST model include the gamma match? My model doesn't.

My model may be wrong, but just for grins compare it to your details for your CST model, and tell me how my gain, angle, SWR, and source compare at 36' feet if you can.

I have added my tabular current distribution chart, maybe you can compare the phasing relationship idea relative to the collinear affect too.

At the end of the PDF file I have added the Big Hair Model @ 36' feet, what I consider is possibly a more true collinear. IMO, this is a base line for how the tabular current distribution and phase all works out in this fairly simple antenna. I am not suggesting that the phasing stub design is correctly dimensioned, but I added that to indicate that it was at least working as intended and reversed the currents at the right spot in the radiator.

I also added a pattern overlay to the Sigma4 vs. the Big Hair collinear 5/8 wave vertical.

View attachment Sigma4 vs. Big Hair Collinear 36'.pdf

 

[Shockwave]

Marconi, I think the model strongly supports the new term "Non apparent collinear". I recognize the fact some will still disagree. All I'm doing is sharing the latest information available on the old Sigma design. When I say that EZNEC misses the boat 100% on the cone, that doesn't mean the EZNEC models are useless. They actually have been helpful but they have not been able to properly demonstrate the antenna mode currents on the cone that constructively add to the main radiator.

That would cause EZNEC to report lower gain, a radiation phase angle error off by 90 degrees and wider beamwidth since it sees the cone as transmission line only and the antenna more like a 1/2 wave. Through CST and EZBob we have learned that there are indeed two currents on the cone. EZNEC only shows one current but when you look at CST, you clearly see transmission mode currents canceled inside the cone and antenna mode currents combining with the main lobe on the outside of the cone. It is not closely spaced parallel conductors simply carrying transmission mode currents that have been canceled.

The antenna has been modeled in CST with a gamma. If you look close, you can actually see it facing directly forward. Again, I am not the CST expert and can only tell you what I've learned from them. The gamma is not so important in the model for matching purposes. You can adjust the model to tune to an alternate impedance. It is added to the model so you can confirm it does not alter the pattern from the desired omni.

The posted model also has a couple of feet of mast attached. Other CST models have shown the currents descending down the mast continue in a constructive phase for the first 1/4 wave before beginning to invert and decrease in intensity as energy is radiated off the mast. Therefore, in just about any foreseeable installation, the antenna will perform better when coupled directly to a conductive mast. One might tweak the performance by using a 108 inch mast and choking off the coax an electrical 1/4 wave down from the connector.

Your Sigma model without the loop behaved differently because you didn't extend the radials to compensate for the loss in electrical length once the loop was removed. Extend them some and the antenna will perform more like the loop but the pattern is less even and the mechanical strength is reduced.

 

[PhilipB]

Thanks Shockwave.

"That can screw everything up from the value of the cap to the location and length of the tuning stub on the bottom coax."

Hi Shockwave, that should include my head, this thing is a nightmare.

"The capacitor is quite low in pf value. I once mentioned in another forum that the tolerance of this cap should be held to within .25 pf and was told that was impossible. The gentleman refused to understand this was about 3% of the total value and not some impossible tolerance or value to achieve. The value is 8.7 pf, no where near 15 pf. Most air variables with aluminum plates won't tune this low. The tiny trimmer caps used to pad crystals often tune this range. Problem is the RF voltage at this point is much higher then you would be expecting. More then 2000 volts will be present above 600 watts and about 1000 volts at 150 watts."

Thanks for that, glad i didnt purchase the high voltage 15pf cap I was looking at, and I know now that I cant trust the plans either, the air variable didnt work as you say, and my chineese POS L/C meter wont read correctly that far down either another headache (luckily when I complained about how poor the quality was he refunded me in full and told me to keep it!)

"Problems with VSWR can be traced to any one of a number of issues in this homebrew design. If adjusting the value of the cap and single conductor wire above it do not make it resonant in the correct place, remember the length of coax below the cap is just as important and will vary with different cables. The velocity factor here will also change the physical location of the tuning stub."

Dont think this thing was ever meant to be copied its driving me mad! if I had any hair I would be pulling it out.

"I consider myself fairly skilled at antenna building and the GM is not one antenna I would feel confident in knowing it had been built 100% correct without an accurate comparison against the real McCoy. I don't say this to discourage anyone from trying this, just be aware there are several areas where one can go wrong and not obtain the desired performance. Most antennas only have the main radiator length and perhaps a tap on an inductor to adjust its resonant point."

Yep I totaly agree with that and have a pityfull example to prove it.

"The GM is far more complex with its cap value, length of orange coax, length of top black wire, length and positioning of the coax stub and last but not least, proper cutoff of the radiation within the RF choke. Many of these perimeters can be off and you'll still be able to get a match by compensating elsewhere. Except you'll be missing the full performance unless all elements of this antenna are tuned properly."

It is strange wonderfull anoying frustraiting and evil thing all at the same time, I made the coil to the given plans length (bigger former) placed the stub match on a T connector to help trying different coax, got the capacitor as close as I could to 8.7pf by measuring with it connected to another larger capacitor (not easy and I doupt reliable) and pruned the radiator till I had 1.2:1 swr at 27.00mHz (this was lying horizontal resting on chairs in the garden) put it upright and everthing was messed up again! lowest swr about 25.400 Mhz reading 1.8:1 and much less of a bandwith.
I am ready to admit defeat and return to a simple end fed long wire and un/un, I think Sirio must be laughing their socks off.
4000?

Phil
26WF183
M6MRP

 

[PhilipB]

HB GM

One thing I forgot to mention is that this is first antenna I have ever used that tunes up exactly as it reads in my MFJ antenna tuner manual as per the settings recomended and tunes very easily! so even though it is a pityfull effort it does have something going for it.
Phil

 

[Needle Bender]

the antenna will perform better when coupled directly to a conductive mast. One might tweak the performance by using a 108 inch mast and choking off the coax an electrical 1/4 wave down from the connector.

Since it's the coax shield which will be operating in antenna mode, it might be closer to what is wanted, to choke the coax at it's mechanical 1/4 wave, taking into account the VF of the casing shortening it (to maybe 98%?) instead of it's internal electrical 1/4 wave, don't you think?

Now, say the mast is 2" in diameter or about 6" in circumference, should that, or 1/2 of that 6" be deducted from the 108" undermast, or perhaps even a little more deducted if the [234÷27.2mhz=1/4wave] formula is used?

Which Vector type antenna should I use for my test, or what modifications should I make to a new Sirio Vector to insure the best (or at least a lot closer to the best than stock) performance?

 

[Marconi]

What's going on? If you need advice as to what to do, then tell us what you have in mind to do, so we won't just be just guessing in the dark. Asking for advice on your plan of attack around here can be risky business, and in particular if you're going to give us your final results beyond all doubt for how the Sigma4 works.

NB, I have no idea what you're about to do here, and I don't intended to suggest you expand your work unnecessarily. If it was me, and based on the limited reports that Bob85 and maybe others have given us...regarding steering the maximum angle of radiation by tuning, and that the Vector will only produce maximum results, at a distance, if tuned like Bob did his and made it longer, I would do the following.

A simple approach, might be to build a setup as close to a stock antenna as possible...just like what the manual suggests. I would suggest using a New Vector 4000, since that one is available and it has some instructions on setup.


I only thought you were just going to prove to us, beyond all doubt, that the Vector/Sigma4 didn't radiate from the bottom 1/4 wave cone element?


For me and to do that, it don't matter which antenna you use, as long as it is similar.

 

[Marconi]

NB, personally I don't think you'll conclusively be able to determine beyond all doubt what is going on in the bottom of the Vector/Sigma4. However you might try and duplicate what Bob85 experienced some years ago, and see if you can get some similar results at a distance like he and Shockwave did in their testing.

On a limited bases and among others, I've tried to duplicate Bob's results, but I could not tell any difference without using some "pixie dust" or Merlin's magic wand. I believe Bob experienced what he claims, but I think it was conditions that come and go, just like all of us mortals experience from time to time using our two way radios that are prone to be at the mercy of Mother Nature's conditions. I also don't believe a single long distance test between any two test stations, at a distance or not, makes for conclusive testing results.

Does anyone really believe that those Avanti guys or the guys at Sirio today are so inclined as to leave their best ideas for a really superior product on the table...when a simple tuning exercises, like Bob recounts, will make the difference? That just doesn't ring reliable in my thinking. I have experience similar situations with unusual long contacts, but I didn't attribute it to anything I was doing, and for sure that I was steering the antenna's angle of radiation during the tuning process.

I think my modeling experiences also has shown me that...other than changing the height of my models, I doubt anything reasonable I could do to modify one of my CB verticals, including tuning, will notably affect the maximum angle of radiation, period. Frankly, I don't think, within reason, that tuning has much affect on signal performance anyway even though theory says the best match produces the best performance and efficiency. It may be the big deal in theory...that tuning makes a difference, but how much difference does it make.

In my real world experiences, I have never found an antenna designed and sold for CB work that ever made a detectable difference in performance due to a simple modification like tuning for the best SWR. Tuning just didn't make that big of a difference as I could tell just using my radio and confirming signal reports.

NB, if you do this side of the issue in your testing, and find that you can duplicate Bob's efforts and results...then I could be convinced that I am all wrong.

 

[PhilipB]

GM antennas

Hey Marconi Ive been looking at all your GM Vs other antenna vids nice camera work, I am not going to comment which one I like the best but out of curiosity how many antennas do you own? Ive heard of antenna farms but you must be growing a greenhouse full of antenna seedlings
Phil

 

[Marconi]

Marconi, I think the model strongly supports the new term "Non apparent collinear". I recognize the fact some will still disagree. All I'm doing is sharing the latest information available on the old Sigma design. When I say that EZNEC misses the boat 100% on the cone, that doesn't mean the EZNEC models are useless. They actually have been helpful but they have not been able to properly demonstrate the antenna mode currents on the cone that constructively add to the main radiator.

That would cause EZNEC to report lower gain, a radiation phase angle error off by 90 degrees and wider beamwidth since it sees the cone as transmission line only and the antenna more like a 1/2 wave. Through CST and EZBob we have learned that there are indeed two currents on the cone. EZNEC only shows one current but when you look at CST, you clearly see transmission mode currents canceled inside the cone and antenna mode currents combining with the main lobe on the outside of the cone. It is not closely spaced parallel conductors simply carrying transmission mode currents that have been canceled.

The antenna has been modeled in CST with a gamma. If you look close, you can actually see it facing directly forward. Again, I am not the CST expert and can only tell you what I've learned from them. The gamma is not so important in the model for matching purposes. You can adjust the model to tune to an alternate impedance. It is added to the model so you can confirm it does not alter the pattern from the desired omni.

The posted model also has a couple of feet of mast attached. Other CST models have shown the currents descending down the mast continue in a constructive phase for the first 1/4 wave before beginning to invert and decrease in intensity as energy is radiated off the mast. Therefore, in just about any foreseeable installation, the antenna will perform better when coupled directly to a conductive mast. One might tweak the performance by using a 108 inch mast and choking off the coax an electrical 1/4 wave down from the connector.

Your Sigma model without the loop behaved differently because you didn't extend the radials to compensate for the loss in electrical length once the loop was removed. Extend them some and the antenna will perform more like the loop but the pattern is less even and the mechanical strength is reduced.

Shockwave, I sure wished you had some visual evidence for what you're suggesting here. I can't argue with your words, and frankly most is over my head. I can only show what these models show.

I have no preconceived ideas whether these models are right or wrong, and you have no evidence for your words either except to say what you think others have told you.

You claim Eznec is in error, and not capable of showing what is going on. That is easy to argue with just words, but what about support for your words with evidence for what you claim instead of preaching the words of others that Eznec has limitations. The fact is, you and I probably don't know the answers for sure, but I'm not willing to cast aspersions on CST, simply because I don't know the facts. If somebody could demonstrate using CST that there are differences to be noted that would convincingly show such differences, that would be great, but obviously that is not going to happen any time soon I fear.

I don't know if the errors you suggest are a big time issues that could shine the light of truth on this subject if CST was used instead of Eznec, or if the differences you imply are just little errors that only a microscopic understanding could reveal.

I don't disagree with the link you showed us, where the CST is emulating the current follows, because I think I see what you are talking about. I also think my model for the S4 shows something similar, concerning phase, but you say Eznec can't show what I think I see in the Eznec's Tabular Current Log.

So, even though I think we tend to agree, we are still likely to be considered by most readers to be in total disagreement, and that is a strange kettle of fish.

Here is a YouTube video I did to try and explain what I see, maybe that will give you a clue for my opinion on the matter and why, or else it could shead some light on what is wrong.

I left out a look at the current log for the 3/4 wave with slanted radials and no hoop, to show how the currents for that one are not in phase like the S4 model, so maybe I'll do another video and try to be more inclusive with my idea. I also had the thought to use the S4 model and remove the hoop wires in the video and then show the results, to see if trying to describe what I see will prove fruitful form that direction as well, trying to duplicate the effort if you will.

Just recently I tried to ask Henry about currents in another matter concerning a simple 2 element beam, in reference to current phase, but I didn't get to first base with him. He did not understand my point of view for the question, and told me he did not understand the question. So, see I continue to strive to learn what is right and what is wrong, but I don't get any help. I could be wrong in how I perceive these current phase matters, and thus continue down a path with no logic, but one day I will get the help I request.

Sigma4 vs. .75 wave with slanted-up radials and no hoop. - YouTube