Francis Amazer Patent info.

[unit_399]

I had two of them mounted in a co-linear array, 3" apart on an aluminum plate atop a heavy duty spring and used nylon spacers to keep the antennas exactly 3" apart from top to bottom. Worked very well on the '89 Toyota pick up I had at the time.

I tried exactly the same setup on my '87 Toyota 4x4. I had the antenna mounted atop the roll bar. I was using a reinforced "gumdrop" mount, and the wind resistance caused the mount to crack. However, I couldn't tell one bit of difference between the pair and a single Amazer in either xmit or recv. I wish they still made them . Best mobile antenna I ever ran.

- 399
BTW - my Amazers were gray.

 

[Marconi]

At some point in the past shortly after Francis sold to Barjan I read an interview in a Trade magazine where the inventor RJ Francis reported the sell on his company. The only comments I recall was Francis reporting that Barjan would not be purchasing the tools and construction rigs use by Francis. He also said he warned Barjan that may be a mistake. No further discussion was published as to any consequences of that warning. However, later we heard reports that Barjan stopped production of the line due to high shipping cost of the long product.

At this late date I think we probably have some idea of what the consequences really were, because we later heard the product did not hold up in use.

I don't know who earlier made this comment here on WWDX that Barjan had production problems, but I agree with whoever posted the idea...the Barjan story was just BS. The likely truth of the matter was that Barjan had production problems and was making faulty products and the report I noted above supports such an idea possible.

I just unsuccessfully tried to model the idea for the 3 wire version that was offset to one side from the middle wire. This version used a creative and effective taper grinding technique toward the tip and was used to create the different wire lengths the design called for.

See image below. I marked approximately where I think the exposed wires show up due to the tapered grinding, but if you want to check the PDF file closer you can zoom in using the % feature at the top of the attachment. Also note the image is not to scale as noted near the base.

BTW, I also checked my Wheeler Dealer and I can physically see the two shorter wire ends exposed in the side of the fiberglass and they are approximately where I see the ends in the image from the patent.

My model failed because Eznec has a limitation for very close wires, but I was just guessing at the possible diameters. I could not even physically see the very small wires I was creating in the program as required by the design. Again this was just a guesstimate.

I might try the idea at a much lower frequency and then re-scale the model up to 11 meters and see if that works without geometry errors. I would like to check the idea and see if modeling can really detect the gain and bandwidth claimed.

My guess is the idea might work as claimed, but as usual I will probably find the differences are so small as to be insignificant in real world operations. I always found the two Francis antennas I own to work very well on the bumper of my P/U trucks.

 

[The DB]

When I read through the patent again after this thread popped up for the first time in years, what they were doing came across awfully similar to a fan dipole in concept. In essence, Francis was using both length and width to control the electrical length of a given wire, where a fan dipole would only control length.

It is another object of this invention to provide an antenna structure having a multiplicity of elongated, electrically conductive elements grouped in parallel relationship and having a desired characteristic impedance for a particular design.

It is a further object of this invention to provide an antenna structure for a particular design frequency band and having a multiplicity of electrically conductive elements of different diameters to provide a relatively wide bandwidth while maintaining a relatively high response and radiation characteristic for the entire frequency band.

It is also an object of this invention to provide an antenna structure for a particular design frequency band and having a multiplicity of electrically conductive elements of different lengths to provide a wide bandwidth while maintaining a relative high response and radiation characteristic for the entire band.

Similar text is also mirrored later in the patent...

However, if all they were doing was creating a fan monopole with individual bands that are close to each other to try and create an antenna with significantly increased bandwidth, as it seems by said patent, then their concept in this case is fundamentally flawed.

The key to this flaw is how subsequent bands in such a multiband desin act...

Quarter wave antennas are desirable because when endfed they approach resonance. Resonance is a state where inductance and capacitance reactances are equal and as a result the total impedance is its direct current resistance.

I modeled a multiband fan type dipole made for the 10, 15 and 20 meter ham bands. The SWR curve for the 20 meter ham band (the lowest frequency band) was normal. However, the SWR curves for 15 and 10 meters were very small, especially for the 10 meter ham band where the bandwidth available wasn't even enough to fully cover the voice portion of the band with 2:1 or lower SWR. This mirrored real world experience as I know of multiple people with fan dipoles who need to use an antenna tuner with their antenna to cover said portion of the 10 meter ham band.

Further, the area in between said ham bands the reactance (X variable) peaked high (over 1000), then low (over -2000) in both of the between band areas, and impedance (R variable) is well over 1000 in this same region.

I then attempted to model a 10/11/12 meter fan dipole, and I literally couldn't make it work. Even with the elements at 0.1 meters apart (a hair under four inches) I could not get all of the elements to tune properly. The antenna was simply not tuning the 11 meter element resonance to the frequency I wanted it to. The further apart the elements are the closer I could get that element to tune properly, however, even at the near four inches between elements I could not get its resonant length to below 27.5 MHz (was shooting for 27.19 MHz) before the longer 12 meter element became dominant and rendered this element useless.

I feel this distance between elements is critical as in the patent the elements are much closer together than the near four inches that I had trouble with. In the patent, the furthest apart any of the individual elements are are within half the diameter of the fiberglass portion of the antenna, while none of the figures give an exact width, I don't see it being more than half an inch in diameter. Any estimate based on said figures will put said elements well below 1/4 inch apart from each other at the largest area of separation. That is about 1/16 the distance that I still had serious issues in trying to model said type of antenna.

If I put such elements all within a half an inch of each other, which is still at least twice the distance apart as the elements in the patent's pictures, the antenna simply only uses its longest (12 meter) element. It completely ignores the 11 meter element, and 10 meter element creates a small, very narrow, anomaly in the 10 meter ham band range that causes SWR to drop very slightly over a very narrow range of frequencies (less than 0.02 MHz, and it doesn't even get below 2:1 SWR). Based on my experience with modeling, especially my recent experience modeling my ford explorer, in a mobile environment you wouldn't even be able to tell such a minor anomaly exists.

All this being said, I simply don't see this antenna working as claimed. This isn't to say that this isn't a great antenna, this is only saying that the patent is, in and of itself, flawed.


The DB

 

[Marconi]

Attached below are the models I did to try and duplicate the Francis Wheeler Dealer, with three wires offset from the center wire to one side.

I used a model on file (#1 .25 slanted radial RES 33") to to build the Francis, to have something to compare my SWR, match, bandwidth curve, gain, and angle if I was successful in building the Francis. My first attempt failed because I could not see the small wires or spaces I was building.

I change the frequency of the model to 1.80 mhz and clicked on scale the model. I did this in order to get the wires physically larger, so I could work on them more easily. My plan was, if successful, I would reverse the scale back to 27.205 mhz and hopefully compare results with the default model noted above.

I included this model at 1.80 mhz as a reference, so we can see the difference in wire size going in at .125" inches up to 1.89" inches on scaling the model down in frequency.

Note: I have not yet tried to modify the wire length differences as noted in the patent, but the results suggest I need to shorten something in this three wire radiator. You will note that even so, the claims made in the patent look to do what is claimed, more bandwidth, and more gain. I posted earlier I did not expect this much difference and this confirms for me that I was wrong.

I will continue to work on the model...so these results might change.

The 1st model is the 1/4 wave default model at 33" inches above Earth.
2nd model is after scaling to 1.80 mhz.
The last model after I scaled it back to 27.205 mhz.

I also added SWR curves so we can see the bandwidth results. I added notes. The last page of each model shows the feed point match predicted.

 

[Marconi]

All this being said, I simply don't see this antenna working as claimed. This isn't to say that this isn't a great antenna, this is only saying that the patent is, in and of itself, flawed.

DB, I might have to eat my words after I spend more time with this Francis model, but I think you are wrong here.

This is not the first time that my models have supported reported claims in patents that were way back in time (<>1970) and long before we have modern technology to help those old antenna gurus design and actually build these detailed and sophisticated designs with little more than a slide rule and some understanding of old physics which goes even further back it time...before we had electricity.

I will share my model if you like. Two heads are sometimes better than one.

 

[Marconi]

However, if all they were doing was creating a fan monopole with individual bands that are close to each other to try and create an antenna with significantly increased bandwidth, as it seems by said patent, then their concept in this case is fundamentally flawed.

The key to this flaw is how subsequent bands in such a multiband desin act...

DB, when I posted that I thought you were wrong in your comments above I should have explained better what I meant.

Your info about a fan dipole might be correct, but what I meant was...I don't think this was what Francis was considering in their idea.

IMO the difference is...the fan dipole does use various parallel wire lengths in its design, but these wires are cut for totally different bands. This design is intended to provide for some multi-band use.

The Francis also uses various parallel wire lengths and wire sizes in another version, but they are cut at different frequencies within the same band...and for me that is a totally different issue. This design was intended for improved mono-band use.

If you will notice in the SWR curve view for my new Francis model there are three distinct dips in the high frequency side of the curve...and this is similar to what I've seen and read about in other articles on broad-banding vertical antennas using modeling. I saw similar dips in the curve with the original Francis model that had the elements all set at the same length, but the advantage in shortening two of the radiators within the CB band add a noticeable broadening of the bandwidth even more. IMO this proves the design idea was right and not flawed.

Just to be sure my "original Francis" model I posted was not corrupted in the process of scaling and re-scaling the model, I removed the added wires to return it to a single element 1/4 wave again just like I started with. I can report that model lost the nice gain and some of the bandwidth advantage the Francis model predicts.

1. The first model below is the original Francis model and is the one I got after re-scaling back to 27.205 mhz. I added an antenna view showing the top of the radiator so we can see the three radiator wires were all the same length at this point.

2. I took this model and played with the wire lengths for wire #2 and #3, to try and duplicate what Francis did. I made about 6 modifications in various lengths until I saw a peak and then a subsequent decline in results. This model works exactly like Francis predicts. I know the theories for this were around, but I'm amazed that they were able understand and do this back in 1970's, before modern tools were available like we have today.

Click on the PDF image file below to see these two models.

 

[The DB]

IMO the difference is...the fan dipole does use various parallel wire lengths in its design, but these wires are cut for totally different bands. This design is intended to provide for some multi-band use.

Most of what I said was discussing what would happen from starting with a multiband fan dipole and adjusting the elements to be closer together length wise, and that fell apart when the lengths were in neighboring bands (10, 11, and 12 meters).

What I didn't talk about as much was if I take a multielement antenna such as the Francis and started all elements at the same length and gradually separate them. What I saw was the result of the models could easily be explained as the continuation of what was happening as I moved the lengths of the different elements of the fan dipole closer and closer together.

I actually did a test with four closely spaced vertical elements as well. One thing I noticed...

Just to be sure my "original Francis" model I posted was not corrupted in the process of scaling and re-scaling the model, I removed the added wires to return it to a single element 1/4 wave again just like I started with. I can report that model lost the nice gain and some of the bandwidth advantage the Francis model predicts.

There is a bigger difference in the SWR curve of my models going from a single element to four elements of the same length than the difference in a model with four elements of the same length evenly spaced compared to a model with four elements evenly spaced over about 4.5 inches of separation...

I suggest making a model with the elements all being the same length and use that as your SWR base comparison. You will see that changing the different element electrical lengths doesn't actually change the general SWR curve very much, it only adds very narrow banded areas of low SWR, and, at least in my models, there are high SWR spikes in between those areas.

In essence, what you see as two different concepts is simply two extremes of the same concept. The differences are predictable, and show as extensions of each other. They are, in essence, directly related.

From what I can see, the biggest cause of the difference in SWR bandwidth over a single antenna, as I mentioned above, is the separation of said elements. The fact that you have multiple elements means that they will take up a larger area, and this, in and of itself, is known to increase SWR bandwidth...


The DB

 

[Marconi]

Most of what I said was discussing what would happen from starting with a multiband fan dipole and adjusting the elements to be closer together length wise, and that fell apart when the lengths were in neighboring bands (10, 11, and 12 meters).

Most of what I said was discussing what would happen from starting with a multiband fan dipole and adjusting the elements to be closer together length wise, and that fell apart when the lengths were in neighboring bands (10, 11, and 12 meters).

What I didn't talk about as much was if I take a multielement antenna such as the Francis and started all elements at the same length and gradually separate them. What I saw was the result of the models could easily be explained as the continuation of what was happening as I moved the lengths of the different elements of the fan dipole closer and closer together.

DB, you comment on Fan Dipoles in the three postings noted above, but I still fail to see any connection to be made to the Francis ideas. If you get a FD's wires real close it would not surprise me that it would fail to act as intended for all of the bands it was cut for.

I actually did a test with four closely spaced vertical elements as well. One thing I noticed...

There is a bigger difference in the SWR curve of my models going from a single element to four elements of the same length than the difference in a model with four elements of the same length evenly spaced compared to a model with four elements evenly spaced over about 4.5 inches of separation...

Are you suggesting if I make my space to 4.5" of separation between my wires on my model and make the wires all back to the same length...I will see even more difference in bandwidth compared to the 1/4 wave mono-pole I started with. I'm just guessing here, but if that would be 4 whips at 4.5" inches apart that would be about 13" inches. That might look similar to a setup guys use to run on their trucks, with 2 x 102" whip antennas on a short bar, if I'm right?

If so, then I would not be surprised they might see some difference and it could be an advantage. All I know about such setups was the guys all raved about how well they worked compared to a single 102" whip. One neighbor I had use to run a 2000 watt AC Pennypincher amplifier into such a setup, and he too raved about how well it worked. He never mentioned it showed a good bandwidth, but like you say...such an idea of multiple radiators closely spaced would seem to make good sense regarding bandwidth.

I suggest making a model with the elements all being the same length and use that as your SWR base comparison. You will see that changing the different element electrical lengths doesn't actually change the general SWR curve very much, it only adds very narrow banded areas of low SWR, and, at least in my models, there are high SWR spikes in between those areas.

This is exactly what my buddies use to complain about with their vertical Fan Dipoles. The antenna would show the dips for the bands, but they were too narrow for some bands.

In essence, what you see as two different concepts is simply two extremes of the same concept. The differences are predictable, and show as extensions of each other. They are, in essence, directly related.

You might be correct. You seem to prefer to always check out the extreme edges with your antenna models, and I tend to shoot right down the middle with mine. IMO, this is just a difference in approach.

From what I can see, the biggest cause of the difference in SWR bandwidth over a single antenna, as I mentioned above, is the separation of said elements. The fact that you have multiple elements means that they will take up a larger area, and this, in and of itself, is known to increase SWR bandwidth...

I agree DB, but what do you think would happen with the gain and bandwidth if we just made the mono pole radiator diameter as big as the overall space a Francis occupies...using very thin wires instead?

 

[gamegetter]

something interesting i noted while glancing through the Francis "Multielement radio-frequency antenna structure having linearly arranged elements" patent....

In the second image, known as sheet 2 of sheet 3 on the patent filing, there is a figure 3 which appears to me to show a cross section of the antenna with possibly 5 wires.

in the description, this is described as :

"FIG. 3 is a transverse sectional view similar to FIG. 2 of an antenna structure of modified construction"

Could it be the amazer may have more than 3 wires(possibly5)and the wheeler dealer could be the 3 wire antenna?

I don't suppose anyone wants to tear into either of these antennas if they have one to find out. I know i would not want to even though i have a couple of the wheeler dealers!

 

[Marconi]

Gamegetter, I think you're right that Fig 3, is likely the Amazer. However it only has 5 wires. Good observation anyway.

Check me out. On column 4 lines #7 thru #10 you will see that notation that #25 is the core, and is made of a similar material to the FG support structure for the Francis, and it is not a wire. It looks to me like a separate dowel in the middle of the antenna, and maybe it is used to provide the proper space for the 5 wires during construction.

I see the asymmetry of some of these Francis wire designs working similar to the asymmetry we see in a typical J-Pole but with less effect. I think that is because the wires are much closer together. I don't think any of this asymmetry noted in the models is detectable in operations using our radio however.

 

[gamegetter]

yup i hit a typo!

5 versus 3 wires though, this is the first i have heard of that, i previously only knew of the 3 wires.

 

[bob85]

Eddie,
I think it was freecell that talked about barjan having issues with the amazer, the ferrules were not glued well enough to the Fiberglas,

I also think somebody claimed the amazer gave a match closer to 50ohm @ resonance in a typical install than a stainless whip,

could it be that they used very thin wires to raise the end impedance when compared to a single fat conductor?

If you have a single thin wire operating a little above its resonant frequency so its inductive & a very closely spaced shorter wire operating below its resonant frequency so its capacitive,
can the combined effect cancel X or do you end up with some reactance on any frequency you test on?,

if you can cancel reactance that way maybe by using appropriate diameters and lengths can give low reactance over a wider bandwidth than a single wire,

we know that fatter conductors give a wider bandwidth & I think a lower end impedance,
and that multiple wires in a cage act similar to a single fatter conductor,,

maybe the amazer is not resonant but gives a match closer to 50ohm on a typical install than a stainless whip,

cbers back then did not have analysers to measure reactance, I would think they would be happy if their vswr meter showed a good match.

 

[Marconi]

Eddie,
I think it was freecell that talked about barjan having issues with the amazer, the ferrules were not glued well enough to the Fiberglas,

I can't be sure of the comments freecell made in this regard, but I do seem to recall discussing the Francis long ago. I've posted my recollections about the sale of the company before. Maybe freecell also recounted a similar story about Barjan having construction issues.

I also think somebody claimed the amazer gave a match closer to 50ohm @ resonance in a typical install than a stainless whip,

In the 60's -70's I ran both antennas mobile at some point, and I don't recall having problems with either...even though the Francis was shorter. I liked it simply because it was much stiffer and didn't bend over like the whip.

I have old antenna notes however that claimed I saw the Francis producing resonance much lower in frequency than the 102" whip, and I did not understand that at the time. I wanted to raise the frequency, so I cut a little off the tip of a red Wheeler Dealer I still have. If the WD does have 3 wires of different lengths, I didn't realize it at the time and I could have changed the mix of three wire design ratios inside.

At the time I didn't even know there were three wires of different lengths inside, but I do remember seeing a very small wire in the tip...and it was very hard to cut...it was like steel.

could it be that they used very thin wires to raise the end impedance when compared to a single fat conductor?

If you have a single thin wire operating a little above its resonant frequency so its inductive & a very closely spaced shorter wire operating below its resonant frequency so its capacitive,
can the combined effect cancel X or do you end up with some reactance on any frequency you test on?,

I don't think I made my Eznec model to precisely follow your idea noted above, but IMO you are correct in the order of the wires. Without destroying the antenna, I had to guess at all the wire dimensions and spacing I used, so my model could be a little in error. The results seemed to fit the design idea in the patent...so I stopped.

if you can cancel reactance that way maybe by using appropriate diameters and lengths can give low reactance over a wider bandwidth than a single wire,

we know that fatter conductors give a wider bandwidth & I think a lower end impedance,
and that multiple wires in a cage act similar to a single fatter conductor,,

Well Bob, what you say here might be right on point. There is a lot more understanding that a more accurate model might show us. The model has very small dimensions and was not easy to make. I had to scale the idea into broadcast frequency range to build it, and then scale it back to 11 meters. I ended up with a model with Geometry errors, which Eznec shows us in an error message. The error was considered to be a minor error however, and I could still scan the model, so I didn't talk about the errors. Later I fiddled with the model and adjusted the model where it showed no such errors...but I decided not to discuss this either. The results on fixing were to reduce some of the gain difference my posted model showed compared to a 102" whip mount at the same height above Earth. The important difference, the bandwidth, did not change much however, so I didn't comment on this either.

maybe the amazer is not resonant but gives a match closer to 50ohm on a typical install than a stainless whip,

I did a model of what I think is the Amazer design also, and it was much more symmetrical. The added bandwidth sill remained. The thread didn't get much response...so I didn't post this model either.

In my personal experience with the Francis designs, I had, all worked fine. Again I like the stiffness of the FG over the ss whip, but I don't think I could tell a bit of difference in performance. Back then, I did not consider the bandwidth differences if there were any.

cbers back then did not have analysers to measure reactance, I would think they would be happy if their vswr meter showed a good match.

You are right, but why would a Francis necessarily show a better match than a 102" whip?

The match was the question I use to ask myself all the time, "...why do many guys report that the 102" whip does not show a good match." This nagging question is what got me into the past time of trying to better understanding antennas, and gave me the idea for testing...using the Marconi X antenna instead of a variable mobile vehicle body.

 

[bob85]

I don't know they do give a better match @ resonance Eddie, they were never available over here,
its just something I remember somebody claiming, I recall me imagining some neat trick with different diameter & length wires maybe doing that but never had the chance to look into it,

Francis claim that taper grinding the whip results in the wires been of different lengths and diameters as the outer wires been the first to be exposed are ground shorter than the inner wires,
I would think just cutting the antenna tip will upset the balance of wire lengths if their patent claims are true,

we know a typical 1/4wave whip install on a car roof when tuned for lowest vswr is not tuned for resonance and when tuned for resonance will have a lower than 50ohms impedance, something in the mid 30ohms range when mounted over a large ground-plane,
same deal with 1/4wave groundplane antennas, you can angle the radials downwards to raise the impedance,

I have a Shakespeare wondershaft a fatter avanti hippo and a 1" diameter 1/4wave copper tube,
none of them show lowest vswr @ resonance on my car roof, the fat copper tube indicates the lowest impedance at resonance and the broadest bandwidth,

I don't know if my past observations fit with your models of different diameter 1/4waves over the same groundplane,

do you remember the tri-quad antenna freecell showed use using 4 amazers spaced around a 4" or so circle,
he claimed very wide bandwidth for that setup.

 

[Marconi]

I would think just cutting the antenna tip will upset the balance of wire lengths if their patent claims are true,

Yes I agree Bob, but in my fiddling with my model I would guess if one cut less than a 1/4" there might not be much difference one could tell.

I don't remember what happened as a result either, but if it had been dramatic...I think I would have recalled that. Beyond that, all I recall is the surprise at how hard the thin wire was to cut.

do you remember the tri-quad antenna freecell showed use using 4 amazers spaced around a 4" or so circle,
he claimed very wide bandwidth for that setup.

I have an image of 4 Francis set in a square in mind, but I don't remember any details. I have an old buddy that use to brag about his 12" bar with a feed point in the middle and two 102" inch whips on each end that might have been similar.