Hugues and a few other folk have been looking at using dipoles on 433MHz, I presume this to be the telemetry or RCS radio frequencies used by them. 

I have been playing with some concepts here to try come up with a repeatable implementation with reasonable performance. The main aim is to obtain a good match to the feedline and RX/TX equipment, and to ensure that the dipole radiation pattern is preserved as best possible. The idea is to use a vertically orientated sleeve dipole on the ground as this could be housed in a plastic tube with the radio connector at the base of the tube. On the aircraft such an antenna is rather unwieldy since it is around 400 to 500mm tall, so a conventional dipole coaxed into a V form is suggested. This can be cut into the fuselage and vertical stabiliser foam, or simply taped to the foam on the outside for tests. It could also be taped onto the horizontal stabiliser, one element on each, with the coax down the center of the fuselage, although the signal polarisation to the ground vertical is not optimum this way. If this were the prefered installation, then a similar 'V' could be used on ground.

I have not indicated all the final construction details, such as the tube into which the sleeve dipole could fit, or a substrate onto which the V antenna could be fitted for robustness when used for ground antenna - I can help with further ideas and suggestions in this regard if need be - it is just a little difficult as I do not know what materials may be available to each in his part of the world...

Onto the antennae:

The sleeve dipole resembles a section of coax cable with a 1/4wave section of the braid folded back onto itself, over the insulation jacket. This leaves a 1/4wave section of the center conductor at the top, and then below the 1/4 wave braid.

This forms the dipole and the braid section forms both the one half of the radiator, and a balun to suppress unwanted current from flowing down the outside of the coax shield and degrading the radiation pattern of the antenna.

However, the performance and matching of such a 'folded braid' sleeve dipole is not optimum.  A more optimum sleeve is where the inside diameter of the sleeve is 2 to 4 times the outer diameter of the coax cable, and when around 4 times, the sleeve should be around 0.21 wavelengths long. With this in mind, an antenna was constructed and tuned to see if this would be easily reproducible. The first used the 'folded braid' method, and then a further three employed brass tubes as the sleeve, of 7mm, 9.4mm and 11mm internal diameter. In each case the sleeve length started out at 0.26 wavelength.

In each case the antenna was tuned by trimming the length of the top vertical element till resonant. Then an RF current probe with a spectrum analyser was used to measure the currents flowing in the braid of the coax , a 400mm long section below the sleeve on the antenna. Then the sleeve was trimmed in length by 2mm, and the top element re-trimmed for resonance, and the currents measured again. This process was repeated on for each diameter sleeve tube.

The results indicate that for the thinner diameter tubes ( folded braid being the 'thinnest') the SWR would remain above 1.5:1, and the common mode current on the coax outer shield was still high. As the tube diameter increased, the tube would need shortening, the top element lengthening, and the SWR would improve. Simultaneously, the coax currents began to reduce dramatically. The 9.4mm diameter tube showed excellent results, with SWR of 1.15:1 at 434MHz, a tube length of 145mm,( around 0.21 to 0.22 wavelength,)  a top element length of 182mm, and the coax current was less than 7%  of that measured on the folded braid version. The 11mm tube showed no worthwhile improvement, so the 9.4mm tube is chosen as the optimum.

Dimensions are indicated in the drawings below: 

The construction is as follows - the tube has a small brass nipple , or disc, with a hole in it dimensioned to pas the coax braid. This nipple is soldered into the top of the tube. Inside the tube are three plastic spacers, through which the coax passes, keeping the coax centered in the tube. The end of the coax is tripped of the insulation for about 2mm for the center conductor, and then the braid cut back to expose around 2mm on the insulation of the center conductor. The outer braid insulation is cut back about 6mm and the braid then enters the nipple, with the exposed center conductor protruding. The braid is then soldered to the nipple, and the top element soldered to the protruding coax center conductor.

different sized tubes with top nipples                                                  Nipple and coax prepared

                                

            

Nipple soldered to coax and braid                                                       Exploded view

Plastic spacers fitted

The following images show SWR and Smith chart data for the final antenna. Note that this antenna is not fitted into any housing or tube. I fitted it into a length of 20mm diameter PVC conduit tubing to measure the effect, and it is quite dramatic. Therefore, if anyone wished to package is so, please let me know what the tubing that you wish to use is - I will try to obtain something similar, and retrim the elements to compensate for the tube shortening effect.

SWR is 1.15:1 with no plastic overtube.

SWR is 1.04:1 at 425.35MHz - a big change with the plastic over tube.

SWR at 434MHz is now 1.8:1 with the overtube.

The V Dipole is made with a 1.4 wave section of the same coax serving as a balun to suppress the common mode coax currents. It is seen as the parallel section in the photo at the beginning of this blog.

here the antenna is taped to a tall block of polystyrene, which does not affect the antenna characteristics, while taking measurements. The balun is clearly visible.

 

   his shows how the balun is terminated at the top -

the main coax shield connects to the left element.

The main coax center conductor connects to the shield of the balun and to the right hand element. The other shield end of the balun section ( photo above) connects to the main coax braid at that point.

The length of the balun is the same as one half dipole element, close to a 1.4 wave. Measurements were taken with the current probe with and without the balun - with the balun current levels were almost 22dB less, a significant amount.

This shows the balun shorted to the main coax at the balun bottom end.

Here are SWR Plots:

 

SWR is 1.01:1 at 434MHz

In this case the bare copper wire dipole elements was replaced with 'servo lead wire' - wire covered in plastic insulation - the resonant frequency has moved down considerably.

SWR is now 1.07:1 at 422MHz.

You cannot just put any plastic over the wires without re-tuning.

Also, 2mm removed from the wire ends shifts the frequency up by 1MHz - it is sensitive to adjustment! 

This Smiths Chart plot shows the excellent match of this antenna - no reactance and a good 50 ohm match.

The 50 Ohm match is achieved by the V shaped elements - bending them into the V form lowers the feedpoint impedance in this antenna. The element show very good match for inter element angles between 100 and 115 degrees, typical of V antennae.

The elements can be bent upwards or downwards ( away from or towards the coax feeder), as required by the installation.

Tapping the antenna elements to an EP or polystyrene aircraft frame or wings will have no effect on the element length or SWR. However, placing the elements against any fibreglass or plastic ( PVC, etc) surfaces will affect the tuning detrimentally. The shrink iron-on cover materials used on model planes will have no effect either.

I hope this will be of some use to all - if anyones ends up building any of these ideas, let me know if I can help with re-tuning for you choice of materials and mounting methods - I will try!

Joe

The Nampilot.

Views: 21356

Comment by mP1 on May 29, 2014 at 6:22pm

A 900mhz antenna should be 1/2 or 1/4 (sorry i dont know the exact scaling) of a 433 and yet in mobiles the antenna is much much less and yet gives significant performance. Even supposed 1.2 antennas for FPV are much larger than their phone equivalent and yet dont see to have any problems. How is this possible ?

Comment by Quadzimodo on May 29, 2014 at 8:12pm
Gisele and Joe - Thank you so much for yet another wonderful explanation on antenna construction. I will put this tutorial to good use.

100KM
Comment by Hein du Plessis on May 29, 2014 at 10:43pm
Thanks Joe, very in depth! How do you think this will compare with the horizontal "turnstile "antenna by ibcrazy? It's orientation allows it to be buried in the wing.

Moderator
Comment by Vladimir "Lazy" Khudyakov on May 30, 2014 at 12:27am

Many thanks!!!

I'm waiting for this post, thanks again!

Comment by Gisela & Joe Noci on May 30, 2014 at 12:42am

Hello Hein,

I have raised some of my concerns with that turnstile in numerous past posts, but that gets lost in the quagmire of posts so quickly...

Assuming the Turnstile works as intended/believed, the comparison does not really apply since that turnstile is supposed to be circularly polarised, and, also if working as intended , it will have some gain over a dipole, not much, but a little..

So, you should really use antenna with similar polarization at both ends to get the best from them. A properly functioning horizontally placed turnstile also radiates equally around its vertical axis, whereas a dipole shows the characteristic doughnut shape broadside to itself, so orientation is important. Dipoles are also linear polarisation, so a working pair should have the same orientation as far as possible.  A properly working turnstile is superior to a V antenna or dipole on the aircraft, and if optimized turnstiles are used at both ends, good results will be achieved. However, a 'properly' functioning Turnstile is very difficult to achieve on a small plane, for reasons I will repeat here in summary, hoping not to annoy those who have already read it in previous posts...

The turnstile has to abide by a few rules.

The basic turnstile consists of 2 co-planar dipoles at 90deg to each other, fed at the center, in phase quadrature. The signal feeding one dipole must reach the other dipole with a 90deg phase shift ( lead or lag..). This results in circularly polarised radiation. If this antenna is laid flat ( horizontal) , the radiation pattern is omnidirectional around its vertical axis ( with some small deviations) . However, the shape of the pattern, and the quality of circularity of the polarisation is highly dependant on the mechanical construction as well. Symmetry is important - the dipoles should be perfectly straight and at 90deg to each other. The feed coax must leave the feed point at 90deg to the dipoles, ie, if the dipole lie horizontal, the coax must exit upwards or downwards for at least a 1/2 wavelength. If the coax exits in the same plane as the dipoles, at some angle between the elements, it becomes part of the radiating structure - the antenna match will deteriorate, the radiation pattern will be totally distorted with unpredictable nulls, and the circularity no longer present.

Lastly, even if the coax did exit at right angles to the antenna plane, the feed mechanism is flawed since there is an unbalanced feed line connected directly to a balanced antenna, with no Balun. Antenna currents will flow down on the coax shield exterior, causing feed matching problems, and the coax will radiate, linearly, and further destroy the radiation pattern and circularity.  

So, a balun needs to be fitted, the coax needs to be properly dressed as it exits the turnstile assembly, and that is a big problem on small planes - there is no space for a long length  of coax  at right angles to the antenna buried in the wing. 

Incidentally, the attempted  90deg phase shift for the requirement of quadrature feed, for the IBcrazy turnstile, is achieved by having made one dipole element longer that the other - one is 'capacitive' and one 'inductive', but reactively equal  and opposite amounts, giving the required phase shift. I am sure a few km range will be achieved with a pair of these antenna, but many more would be achievable with optimal implementations. 

I use a version of the vertical sleeve dipole on 868MHz, on ground for my datalink, at 300milliwats, with the dual-ring Slot radiator on the aircraft. I have not gone beyond 28km range yet, but the signal is perfect at that range. This from two vertically polarised dipole style antenna.

So, although there are far better antenna types for this application, the size and space available results in compromises that could reduce their effectives to less than that of the the humble dipole or V antenna. Also, the V antenna radiation pattern tends to be more omnidirectional than the dipole, so it is more forgiving on the aircraft than a dipole.

The silver circle in my datalink/Video TX/antenna module below coprises two rings of copper, spaced about 6mm and fed with coax .

The Skew planar wheel is for 2.4GHz

Left is Top view with xBee datalink showing

Right is side view of the two rings 

Video TX is at bottom.

The skew planar wheel if constructed properly is a a good solution for these applications, but becomes huge below 1.2GHz so not practical. I used to use one on 'Guppy' ( see some of my other blogs) on 900MHz - it is about 220 diameter and 110mm tall. On the skew planar wheel, the coax also needs to exit at 90deg to the antenna, for at least a half wavelength for all the same reasons.

Joe

Comment by Martin Poller on May 30, 2014 at 12:52am

Many thanks for taking time to post this, and also for your previous articles.

The detail of the build instructions and explanations are exactly what we need.

Despite the negative comments by some "naysayers" the majority of us are here to learn and the input from yourself and other contributors is very much appreciated.

Comment by Jesus A on May 30, 2014 at 1:20am

Very good and technical article Joe

28 km with 300mW and omni antennas is a lot more than what I thought could be achieved! 

Did you do any link budget calculations? What Telemetry modules are you using?


100KM
Comment by Hein du Plessis on May 30, 2014 at 1:27am

Thanks for the repeat of the turnstile explanation. I think in the end, the turnstile will not beat a good dipole in terms of range, but it's much easier to fit to a flying wing, for example, and if the same range is to be got, it has an application. Can you suggest an affordable SWR meter? Thanks again for your generous sharing.

Comment by Gisela & Joe Noci on May 30, 2014 at 2:42am

Martin,

Thanks. As I keep repeating,my difficulty in these posts is to provide a recipe easily duplicated. The UAV field is filled with so many disciplines - flight principles,  electronics, aircraft building, electrics, soldering, antenna, RF transmission and reception, etc, and not everyone has mastered them all, or has tools and equipment to easily do something in each of these fields. Someone may be an excellent craftsman, building beautiful aircraft, but useless at soldering, etc. So how to provide a recipe that most folk could implement is not easy - I made the nipples for the sleeve dipole on my lathe - how does a young DIY hobbyist without such facilities do this?? Ditto the plastic spacers in the tube. The spacers could be done using wooden dowel rod , but if not the correct diameter, he may have to sand it to size, etc. The hole through for the coax needs to be centered within reason, maybe 0.5mm.., how does he drill that without a lathe? I could fit the sleeve dipole into a piece of electrical conduit tube as I tested, and trim the dipole up nicely, but is the tube available to everyone else the same plastic, or will the characteristics be quite different? So the best I can do is provide an idea, with the pitfalls to look out for, let the individual find his own way of doing it, and hopefully complain to me when it does not work, and then we try together to make it work! 

Regarding 'naysayers' , lets not be harsh here..people have different styles and approaches, and we must make room fo them. If the individual knows not the subject at all and derides others, then he cannot be forgiven, but, for example, with Jonathon, I get the feeling he IS very knowledgeable and knows his field well -  It would be good for all if he could come to the party I believe..

Joe

Comment by Gisela & Joe Noci on May 30, 2014 at 2:49am

Jesus A,

Thanks as well..Yes, a pure link budget calc was done as follows:

RX sensitivity is speced at -110dBM - I settled for -105dBM.

TX power is 26dBM

Cable losses total plus one connecter is measured at 1dB (only 1 connector in the loop others are directly soldered to the relevant point).

The dipole was set at 2.5dBi

The Ring/slot radiator was also set at 2.5dBi ( both of these were set slightly lower than reality, but good enough)

A 10dB fade margin was allowed.

This gives a theoretical range of around 50km @ Rx sensitivy of -105dBM.

The module is the xBee -Pro - available here in 300milliwatt adjustable output. It is an excellent module.

Joe

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

Groups

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service