Mar 04, 2011, 02:19 AM
Registered User

# Stacking Two PinWheel Antennas

Stacking two PinWheel antennas provides an excellent omni directional circular polarization system for an FPV RX base. This post will explain how to do it.

First, Mille will be happy that I have modified my model for a single source feedpoint:

http://www.af9y.com/PINWHEEL62m.nec

The model uses the following variables:

After a lot of optimization runs, I found quite an improvement in performance for both axial ratio and gain if the arc length is made significantly shorter than what has been used up to this point. A ratio of arc length to leg length = 1.6x instead of the normally used 2.0x will produce the following:

Gain = 1.7 dBic

RHCP to LHCP ratio = 13.4 dB

For a 1280 Mhz antenna, the variables were set as follows:

Leg Length = 60 mm
Feed Length = 7 mm
Arc Length = 100 mm
Arc Angle = 90 degrees
Pitch Angle = 50 degrees

(memory wire with a 0.6 mm diameter used for elements)

To achieve more gain, a stack of two PinWheels are mounted with the ground leg sides facing one another and separated by aprox 1/2 wavelength (120 mm for 1280 Mhz). They are fed with a special set of coax lines to achieve 180 degrees phase between the two antennas which results in the following pattern:

Gain = 5.5 dBic

RHCP to LHCP ratio = 14.9 dB

That is almost as good as the two stacked CPODs I'm currently using which produces 6 dBic.

Interesting that in this case, 1 + 1 does not equal 2. It equals 2.4!

Here is how the two coax line lengths were calculated to achieve both the 180 degree phase shift and the combiner function:

The added effects of the coax connectors and T connector did not produce good results for the calculated coax lengths. It took quite a few iterations of coax assemblies and testing with a spectrum analyzer to find the correct lengths. Here is a picture showing the final lengths:

I used the above coax combiner approach for the three sets of dual CPOD antennas currently on my tower. (Results have been most excellent.):

OMM
Last edited by Old Man Mike; Mar 06, 2011 at 12:48 AM.
 Mar 04, 2011, 07:33 AM Engineer for Christ Mike, why are you using RG179 coaxial cable? If the pin wheels are set for 100 Ohms (to make a balanced 50 ohm load) then why not use RG-62A/U? It's 93 ohms and thus would be a closer match, right? I think the reason for different coaxial cable lengths is the impedance mis match. I could be wrong, though. In any case, this is a great find! Looks like I'll be building your modified pinwheel for my RX antenna. -Alex
 Mar 04, 2011, 08:07 AM Registered User Great job Mike and I'm so glad to see someone actually taking into consideration the Velocity Factor of the relevant coax in use Nigel.
Mar 04, 2011, 09:45 AM
Registered User
Quote:
 Feed Length = 7 mm
I see you have a feed length on the X axis
That's the PCB you use to solder the legs?

-Hugo
Mar 04, 2011, 04:32 PM
Registered User
Quote:
 Originally Posted by IBCrazy Mike, why are you using RG179 coaxial cable? If the pin wheels are set for 100 Ohms (to make a balanced 50 ohm load) then why not use RG-62A/U? It's 93 ohms and thus would be a closer match, right? I think the reason for different coaxial cable lengths is the impedance mis match. I could be wrong, though. In any case, this is a great find! Looks like I'll be building your modified pinwheel for my RX antenna. -Alex
The PinWheels are 50 ohm antennas (Zin). I am using a multiple of quarter wavelength 75 Ohm cable to convert that to aprox 100 ohms (Zout) via the formula:

Zout = Zcoax * Zcoax / Zin

Zout = 75 * 75 / 50 = 112.5 ohms

When the two 112.5 ohms are connected in parallel at the T connector, the result is 56.3 ohms

(If you had used 93 ohms the result would have been 86 ohms)

As long as either coax is a mulitple of 1/4 wave, you will get the Z transformation above. In addition, if you make one 3/4 wave and the other 5/4 wave in length, the difference is 1/2 wave which produces a 180 degree phase shift. (did you really think that I was just using random lengths of coax to combine and phase offset feed these antennas?)

By the way, you can use the above technique to stack any two antennas but the optimum spacing between the two antennas will vary depending on gain.

Quote:
 Originally Posted by Hugeone I see you have a feed length on the X axis That's the PCB you use to solder the legs? -Hugo
I did that in the model so that the two legs would be equal lengths. It makes it simpler to cut and form a symmetrical loop. And yes, I had thought the 7 mm on the ground leg could be a small PWB 14 mm diameter ground plane to which the four gound legs and the coax connector would be soldered. The 7 mm vertical feed would be a 7 mm long thick wire soldered to the coax center connnector to which the four loop memory wire elements would be soldered.

OMM
Last edited by Old Man Mike; Mar 04, 2011 at 05:24 PM.
Mar 04, 2011, 07:11 PM
...just an earth bound misfit
Quote:
 Originally Posted by Old Man Mike There really is no reason to mount a single PinWheel in any position other then the normal position (four ground legs parallel to ground). I thought you were worried about the small 1 dB gain at -15 degrees elevation compared to +15 degrees elevation. If you mounted the antenna with the four ground legs up rather than down, you could make the 1 dB gain be at +15 degrees rather than -15 degrees. However, in practice this would be very difficult to do since it would force the coax to run thru the middle of the four loops and that would greatly distort the pattern. For a RX antenna, a stacked set of two PinWheel antennas would be an excellent choice for higher gain while retaining omni performance to cover a full 360 degree Az. (my next post will show how to do this). OMM
I wouldn't worry over 1 dB, but I do worry about the toroidal radiation pattern's Z-axis null since I frequently loiter high overhead. The null in your optimization video appears to have a -5 dB half-angle of 22.5 degrees, a -10 dB half-angle of 12.5 degrees, and -15 dB half-angle of 7.5 degrees.

So flying at 500' AGL there will be a 380' diameter circle overhead where I'm more than 5 dB down, a 200' diameter circle where I'm more than 10 dB down, and a 130' diameter circle where I'm more than 15 dB down.

Thanks for the comments Mike, does my concern make more sense now?
Mar 04, 2011, 07:48 PM
Registered User
Quote:
 Originally Posted by Dirt-Torpedo I wouldn't worry over 1 dB, but I do worry about the toroidal radiation pattern's Z-axis null since I frequently loiter high overhead. The null in your optimization video appears to have a -5 dB half-angle of 22.5 degrees, a -10 dB half-angle of 12.5 degrees, and -15 dB half-angle of 7.5 degrees. So flying at 500' AGL there will be a 380' diameter circle overhead where I'm more than 5 dB down, a 200' diameter circle where I'm more than 10 dB down, and a 130' diameter circle where I'm more than 15 dB down. Thanks for the comments Mike, does my concern make more sense now?
Well, let's do the math:

For your high flying, we'll take the distance from plane to ground as 600' and an attenuation of 15 dB compared to a distance flight in the main lobe of the antenna. Of course these are worse case numbers since it is actually a shorter distance for your 15 db number.

You know that every time you double the distance you lose 6 dB so 600' at 15 dB reduction is the same thing as flying 4800 feet away in the main lobe of the antenna. I been able to do that with 10 mw so I'm sure your example will be no problem.

https://www.rcgroups.com/forums/show...&postcount=147

OMM
Last edited by Old Man Mike; Mar 04, 2011 at 10:48 PM.
 Mar 04, 2011, 07:59 PM Tri-Quad-Hexa-Octo-copters!! Hi Mike, I recently came across this 12v step up buck inverter http://shop.ebay.com/i.html?rt=nc&LH...=p3286.c0.m301 thread by PatrickLupo. From the way he describes his success with using it, it might be another thing for folks to consider to clean up the video: https://www.rcgroups.com/forums/show....php?t=1361685 Your thoughts appreciated as always Cheers, Jim
 Mar 04, 2011, 08:02 PM ...just an earth bound misfit I'll be driving one of Hugeone's 5.8 GHz SP antennas with Range Video's 200 mW transmitter module. I wouldn't have been expecting to run that far out on omnis, but I guess it's time to throw it up and see what they'll do. Thanks Mike.
Mar 04, 2011, 09:01 PM
Registered User
Quote:
 Originally Posted by jesolins Hi Mike, I recently came across this 12v step up buck inverter http://shop.ebay.com/i.html?rt=nc&LH...=p3286.c0.m301 thread by PatrickLupo. From the way he describes his success with using it, it might be another thing for folks to consider to clean up the video: https://www.rcgroups.com/forums/show....php?t=1361685 Your thoughts appreciated as always Cheers, Jim
as an Electronics Engineer, I can only agree with the comments made in the thread: step up inverters are based on high frequency switching and thereofore are normally very good noise generators! Also lack of "serious" capacitors on both inputs and outputs can only make it worse and while this can easily be fixed afterwards, this suggests that RFI has not really been taken care of during product design and especially PCB layout which is critical and can not be fixed afterwards.
Sure this is the theory and goes against the finding that the thread found his video quality much improved by adding this inverter.
I did not go through the whole thread but checking RC range would be really urgent I would say.

In one word, I would not put such step up inverter on my RC plane / helicopter, at least not until I can really get access to professional equipment to confirm noise / RFI performances.

only my \$0.02, especially considering the level of expertise in RF in this thread with antennas I will never feel confident to build by myself, not to mention designing them

Vincent
Mar 05, 2011, 12:05 AM
Engineer for Christ
Quote:
 Originally Posted by Old Man Mike The PinWheels are 50 ohm antennas (Zin). I am using a multiple of quarter wavelength 75 Ohm cable to convert that to aprox 100 ohms (Zout) via the formula: Zout = Zcoax * Zcoax / Zin Zout = 75 * 75 / 50 = 112.5 ohms When the two 112.5 ohms are connected in parallel at the T connector, the result is 56.3 ohms (If you had used 93 ohms the result would have been 86 ohms) As long as either coax is a mulitple of 1/4 wave, you will get the Z transformation above. In addition, if you make one 3/4 wave and the other 5/4 wave in length, the difference is 1/2 wave which produces a 180 degree phase shift. (did you really think that I was just using random lengths of coax to combine and phase offset feed these antennas?) By the way, you can use the above technique to stack any two antennas but the optimum spacing between the two antennas will vary depending on gain. OMM
Oh I got it! I forgot the transformer could be used to increase impedance. I guessed that there was some mutual coupling between the antennas the transformed them to ~100 Ohms. In my empirical calculations, the Pin Wheel is about 60-65 Ohms, so this would explain the shortened coaxial cable. Looking in the ARRL antenna book and running the calcs based on a 60 Ohm impedance and RG179, your lengths hit within 5% of the calculation.

-Alex
Mar 05, 2011, 01:53 PM
Registered User
Quote:
 Originally Posted by Old Man Mike First, Mille will be happy that I have modified my model for a single source feedpoint: http://www.af9y.com/PINWHEEL62m.nec OMM
Mar 11, 2011, 04:26 AM
Registered User
Quote:
 Originally Posted by Old Man Mike Stacking two PinWheel antennas provides an excellent omni directional circular polarization system for an FPV RX base. This post will explain how to do it. First, Mille will be happy that I have modified my model for a single source feedpoint: http://www.af9y.com/PINWHEEL62m.nec
Old man Mike that ist one hell of a .nec file, my hat is off to you!

Really a beautiful piece of work, I had a lot of fun playing with it, like making the number of lobes a symbol :
http://chello.peterrenner.net/FPV/ne...62m_980_6a.nec
changing it to 6 lobes (adjusting initial angle to 60) and let the optimizer play with it Really a lot of fun

I took me a while to find the switch for circular ploarization but a couple of optimizing questions remain:
• how can I tell the optimizer to make the ratio of RHCP-gain/LHCP-gain as big as possible ?
I recall you or IBCrazy to have said that the "multipath rejection capability" of an RHCP antenna is proportional to the amount the RHCP signal is stronger than the LHCP signal, ideally not radiating in the undesired LHCP direction at all. Besides that the antenna wouldn't waste power for a polarization I'm not interested. Hence I'd like the optimizer to go for that goal, too.
• Some basic is missing here: Optimizing SWR towards 1 means alle energie goes into the antenna nothing is reflected back, correct?
What is "F/B" front-to-back ratio, "F/R" front-to-rear ratio (sound like a synonym to me) and "Radiated Power%", what effect do they indicate ?
Isn't an SWR of 1 identical to Rad.Pwr = 100%
• where in the NEC file can I specify the min max borders to prevent the optimizer from trying variable values that would ruin the geometry?
If not permanently in the nec file how can I limit the variables values before starting an optimization run ?
Mar 11, 2011, 05:44 AM
Registered User
Quote:
 Originally Posted by reely340 Old man Mike that ist one hell of a .nec file, my hat is off to you!
Thanks!

Quote:
 Originally Posted by reely340 .....I took me a while to find the switch for circular ploarization but a couple of optimizing questions remain: how can I tell the optimizer to make the ratio of RHCP-gain/LHCP-gain as big as possible ? I recall you .... to have said that the "multipath rejection capability" of an RHCP antenna is proportional to the amount the RHCP signal is stronger than the LHCP signal, ideally not radiating in the undesired LHCP direction at all. Besides that the antenna wouldn't waste power for a polarization I'm not interested. Hence I'd like the optimizer to go for that goal, too.
Yes, I've been advocating circular polarization for years because of the multipath rejection capability. I've noticed that it is finally catching on.

As for making the optimizer maximize for RH to LH ratio - not currently possible. I've asked the author of the program (even offered some money) for that feature but he has not yet done it. This is why designing circular polarized antennas for maximum performance is so difficult. Of course the program will accurately tell you the ratio but getting a design to that point is mostly a trial and error game. After enough experiments you get a feel for what works.

Quote:
 Originally Posted by reely340 [*]Some basic is missing here: Optimizing SWR towards 1 means alle energie goes into the antenna nothing is reflected back, correct?
Yes, that is basically it.

Quote:
 Originally Posted by reely340 What is "F/B" front-to-back ratio, "F/R" front-to-rear ratio (sound like a synonym to me)
That is for optimizing directional antennas. Normally you want the beam to be as concentrated in one direction as possible. The higher the ratios, the better the directivity which is especially important for the RX side to reduce noise.

Quote:
 Originally Posted by reely340 and "Radiated Power%", what effect do they indicate ? Isn't an SWR of 1 identical to Rad.Pwr = 100%
This relates to TX efficiency. An antenna can have good SWR and directivity for RX but not be a good antenna if the efficiency is low.

Quote:
 Originally Posted by reely340 [*] where in the NEC file can I specify the min max borders to prevent the optimizer from trying variable values that would ruin the geometry? If not permanently in the nec file how can I limit the variables values before starting an optimization run ?[/list]
Use the "Evolve" selection in the optimization window.

OMM
Mar 14, 2011, 03:11 PM
Registered User
OMM,

I scaled your "optimized" pin-wheel for 2.4ghz, have a look: